Search: 
View
  By section
  As outline
  Fully expanded

FAQ sections
  Getting started
  Effective use
  General questions
  Authoring & Posting
  Forums and FAQs
  Rocket science
  Info for admins
  Mail admin
  Infrastructure
  Calendars
  Intellectual Property
Gigascale Systems Research Center

FAQ

Welcome to the gsrc FAQ.

Getting started on the web-site
Information on how to get familiar with the site and start using its features.

What is this site?
This site is the on-line home of the Gigascale Silicon Research Center. For information on the GSRC as a whole, see the About page.

This web site has two goals. First, to present information about the GSRC and its research goals and accomplishments to the larger research and industrial communities. Second, the site provides a means for establishing a community and network of researchers who share the common vision of the GSRC.

The section of the site you are looking at is focused on this latter goal. The GSRC counts amongst its investigators many of the leading researchers in the country. This team is geographically distributed, and the web-site provides a means of overcoming this barrier to effective collaboration.

The unit of collaboration on this web-site is called a "work-group", which is simplye a fairly loose term for a group of people that are working together on some shared problem. The website provides mechanisms for providing people with accounts, and membership in various work-groups.

Each workgroup also has real-estate on the site, as you can see by browsing some of the links on the workspaces page. The site provides a collection of off-the-shelf features that can be added to each workspace -- according to the type of interaction and collaboration taking place within that group. One of the people in each group is given full adminstrative privileges over that group, thus allowing effective units of collaboration and interaction to form without (as much as possible) the intervention of some centralized "webmaster."

How do I get a login?
There are two main scenarios by which you can obtain a login on this site.
Initiated by you
In this scenario, you request an account on the site. Your request is emailed to the group admins, who approve your request and create your account. You will then be sent mail with your initial password and some instructions on logging in.

To request a group:

  1. Go to the request page.
  2. Fill in the form. When you choose which group to join, you have several options:
    1. If you are directly affiliated with the GSRC (faculty, student working on GSRC projects, or involved industry representative), choose the Member option of the gsrc group.
    2. If you are more loosely associated with the GSRC as a whole but are not really a GSRC member (interested industry representative, for example) choose the Guest option of the gsrc group. Guest account also permit one to register to attend workshops.
    3. If you are collaborating closely with GSRC researchers who have an existing workgroup, choose the Member option of that group.
    4. If you are working only loosely with GSRC researchers who have an existing workgroup, choose the Guest option of that group.
  3. Press the Request account button.

Initiated by us
A group administrator can "invite" you to create an account on this site. In this scenario, you will receive an email message containing the invitation and instructions on how to go to the web-site to create and initialize your account.

If you have received an invitation to join the site, you are not obliged to respond. If you choose to ignore the invitation, it will expire within a week. You can always request an account later.

I forgot my password, what do I do?
If you forgot your password, please visit the password resetting page.

A new random string will be emailed to you. When you get the new random password, use it to log in, and then follow the instructions in the email message to change your password.


What if I can't remember my username?

To find your username you may either search the Directory or you may perfom a search for your name in the Search section.

Who can get a login?
GSRC website accounts are available to members of the following groups:
  • Employees of GSRC Sponsor Companies
  • Students, faculty and staff at Universities that receive GSRC Funding.
  • Individuals who do not fall in the above categories, but who are collaborating with a specific GSRC workgroup may request an account with that workgroup.
  • Members of sponsor companies and sponsored universities can choose to be either a guest or full members of the GSRC workgroup. If you are planning on coming to the quarterly workshops, then you probably want to be a member, not a guest. The primary difference between GSRC workgroup guests and members is that members get more email about workshops. Guests may register to attend workshops.

    Individuals who are not members of sponsor companies or sponsored universities but are collaborating with a workgroup are usually given access only to the workgroup where they are collaborating, they are not given guest or full membership in the GSRC workgroup.

    Note that you might not need an account at all. Most papers are all available for downloading to non-members. Some of the groups are wide open, and all of the material in these groups is available to non-members. However, to access the presentation archives, one must be a GSRC workgroup guest or full member.

    To request a login, see the How do I get a login? question.

    OK, I got a login. Now what?
    Congratulations! The first thing you should do is go to the Your profile page and make sure that your personal information is correct and complete.

    Second, we recommend that you spend a little time browsing the site and familiarizing yourself with its features. From the point of view of using the web site in an interactive fashion (as opposed to just reading content), you should definitely take a look at:

    You will probably want to join one or more SIGs (instructions). We strongly advise that you only join SIGs in which you will be actively working -- each SIG is likely to generate a lot of mail and discussion and if you are not an active participant you would probably be better off just subscribing to that group's "interest" mailing list (see below). Browse the groups listed on the Workspaces page before joining any groups.

    You may wish to create your own group. Creating your own group allows you to create a shared identity on this site, together with the researchers that are working with. See the group administrator's FAQ for instructions on how to do this.

    Thanks, and have fun!

    How do I post files to this server?
    There are several places on this server where you can post files, depending on what kind of file it is and what kind of visibility you want the file to have:
    • The Publications database. Use the publications database for publications that you wish to be citable. You will be able to choose whether your file will be visible to the whole world or only to GSRC members. (More info.)
    • The GSRC Forum. Use the GSRC Forum for files that you wish to share with the GSRC community as a whole, but which are not really formal publications. By default, these files will be world-readable. (More info on using the forum.)
    • The Forum of a work-group. If you are a member of a work-group, use that group's forum for posting files that are part of the on-going collaboration with other people in that group. These files may or not be world-readable, depending on the group configuration. (More info on using the forum.)
    • The Private Forum of a workgroup. Use a group's Private Forum for posting files that you wish to be visible only other people that are also members of that group.

    How do I add a publication to the publications database?
    The publications database allows anyone with a Member account on this site to add publications to the database and upload electronic versions of those publications.

    To add a publication, first make sure you are logged in. Then go to the publications database (“publications” in the “technical” pull-down menu on the home page) and click on the link at the left of the page labelled Add publication.

    The first screen you see allows you to choose the publication type -- for example, journal article, conference paper, and so on. On this page, you can also choose to restrict access to your publication to GSRC members and guests (people with a login on this site). Finally, you can associate the publication with a work-group on this site, thereby allowing searches to be made on publications associated with a group.

    Press the Continue button to go to the next screen. (Any of the information on this page can be changed later.)

    The next screen shows the fields that can be filled in for the publication type you have chosen. Fill in as many fields as you can, noting that fields marked with an asterisk (*) are required and fields not so marked are optional.

    Press the Add new publication button to continue. (Again, you will be able to change any of this information later.)

    You will now see a screen that allows you to upload electronic versions of the publication. The preferred format for all electronic publications is PDF.

    To upload a file, click on the Browse... button and select the file. If you wish, add some additional text in the Additional comment field - this text will appear in parentheses next to the download link for the file you are uploading. Press Upload file to upload the file.

    When you have no more files to upload, press the Return to listing button to view your publication's page!

    (For instructions on how to modify a publication record, see this question.)

    How do I modify a publication?
    If you are logged in to the site, use the publications search page to find the publication you wish to edit.

    If you have permissions to modify the publication, you will see several links on the left-hand margin:

    • Edit publication allows you to change the fields associated with the publication -- the author, title, journal, and so on.
    • Change type allows you to change the publication type. You can also use this page to restrict access to GSRC members or associate the publication with a group.
    • Upload files allows you to upload additional electronic version of this publication.
    (For more detailed instructions, see the question on adding publications -- the screens are essentially the same.)

    You will be able to modify a publication record and upload files if either

    1. You are the person who originally created the publication record, or
    2. The publication is associated with a GSRC work-group or SIG, and you are a member of that work-group or SIG.
    If neither of these is the case, go to the bottom of the page for the publication, where you will see something like "Posted by Fred Nerfs on September 6th, 1999." Then send email to Fred Nerfs.

    The design of the publications facility does not permit regular users to delete publications. This is a design choice that was implemented because once something is published, it should remain in the record of publications.

    If you would like a publication deleted because it was accidentally entered twice, please send email to www@gigascale.org

    How do I join a group?

    To join a group, go to your memberships page and scroll down to the section labelled "Request membership in a group". Use the form, following the links on that page to additional help information is needed. You can become a Member of the group, which will put you on the group's mailing lists and give you write access to the group's workspace. You can also become a Guest of the group, which will generally (but not always -- it depends on the group configuration) allow you to read restricted pages belonging to the group, but you will not be on the mailing lists or be able to modify those pages.

    If you merely wish to keep updated on the goings on in a group but are not actively participating, you should not join the group, but instead subscribe to that group's public "interest" mailing list. Go to the home page for that group (from the Workspaces page) and click on the Mail below the GSRC banner. If the group has a public interest list, you will see a link called, for example, diva-interest -- follow the link to get to a page where you can subscribe to the mailing list.

    How do I alter my group memberships?

    To change your memberships options in the groups that you are already in, go to your memberships page and use the section labelled "Change your group memberships". Use the radio buttons to change your membership to what you want, and then press the Change memberships button.

    You can downgrade your membership without requiring approval from the group administrator. Specifically, this means that you can:

    • Change your membership type from Member to Guest. This will remove you from the group's private mailing lists, which may be appropriate for you if you are no longer actively participating in this group.
    • Remove your Administrator privilege.
    • Remove you Developer privilege. This will remove you from the CVS and developer mailing lists of the group (if it has these lists turned on.) This may be appropriate for you if you are no longer an active developer in this group.

    To "upgrade" your membership, administrator approval is required. Specifically, if you do any of the following, email will be sent to the group administrators containing a URL that they can use to approve your membership change. This will happen if you:

    • Change your memberhip type from Guest to Member.
    • Select the Administrator privilege.
    • Select the Developer privilege.

    Note that none of the above applies if you have enabled the override option

    How do I remove myself from a group?

    To leave a group, go to your memberships page and scroll down to the section labelled "Delete your membership in a group". Choose the group that you want to remove yourself from, and press the Delete membership. You will be removed from the group and unsubscribed from its private mailing lists. No email will be sent to the group administrators.

    To rejoin the group later, you should to follow the normal procedure for joining a group.

    How do I join the "foo" mailing list?
    There are two kinds of mailing list on this server: private ones and public ones.

    Private lists are always associated with a group, and you join the list by becoming a member of that group. For example, if you were a member of the foo group, you would automatically be on the mailing list foo@gigascale.eecs.berkeley.edu. You can't be a group member and not on the mailing list, or be on the group's private mailing list and not in the group.

    Public lists are usually associated with a group, and have names like foo-interest and foo-announce. To join these lists, go to the home page for that group, and select the "Mailing lists" link from the left-hand margin. Then choose the link to the mailing list that you want to join. If you are logged into this site, then subscribing to one of these lists requires only that you press the appropriate "Subscribe" button.

    There's no answer to my question here? Who'mIgonnacall?
    Send mail to webmaster at gigascale dot org (Note: be sure not to send mail to webmaster@eecs -- that's not us.)

    Sending mail to the webmaster list is much better in general than sending mail to an individual.

    Email attachments are not archived.
    Mail sent to most of the workgroup mail lists is archived for future reference. (For example, the archive for the gsrc mail list can be viewed at http:/www.gigascale.org/gsrc/listinfo/gsrc/, or by using the navigation links at the top of the page.)

    However, attachments to email messages are not archived. This is so that large attachments, such as Powerpoint presentations and MS Word documents, can be shared among list members via email, without incurring the expense of storing them on disk each time they are sent.

    While disk space is cheap, and getting cheaper, the cost of backing up disk content is expensive. For this reason, the size of messages that are archived is limited to 40 kilobytes. Dropping the attachments before archiving allows users to share large files, via attachments, without encountering this limitation.

    Why can't I use my yahoo (gmail, aol, ...) email address?
    The sponsors of GSRC require that our membership be limited to those who are associated with sponsor companies or partner institutions.

    The only way we have of being sure that we meet this requirement is to require each of our members to use an email address from the sponsor or partner organization to which they belong.

    Masquerading -- Acting on behalf of someone else

    Masquerading -- Acting on behalf of someone else

    You may be an administrative assistant for a GSRC member, and need to register them for GSRC workshops, or take some other action on the website on their behalf.

    One tempting way to do this is to change the name and contact information in your own account profile to look like them. This only leads to confusion, because it results in two accounts with the same name, one of which does not really belong to that person.

    The correct way is to have them designate you as their "proxy". Ask them to visit their profile page (my account -> profile, or this link: http://www.gigascale.org/options/account/profile/). At the bottom of the page, they can designate you as their proxy.

    Once that is done, you will find an item under your my account menu saying, "masquerade as ... xxx", where xxx is the username of the person who has made you their proxy.

    As soon as you click on their username, you will immediately be logged in as them, without any need for their password. You can then register them for a GSRC workshop, modify their profile, or do anything else you like under their account. When you are finished, simply click logout and you will be back to your own login.

    If you need any assistance with this, please send email to webmaster@gigascale.org.

    How to effectively use this site
    This section of the FAQ contains questions and answers on how to make effective use of the site, what its features are, and what the site has been designed the way it has.

    How should I use this site?
    If you are involved with the GSRC, you should be taking advantage of the features provided by the workspace support built into this site. The concept and implementation of a workspace gives you a number of things, depending on how you choose to look at it:
    • An identity within and association with the GSRC web-site and therefore with the GSRC.
    • A web-based workspace in which to collaborate with other researchers.
    • A tier of interaction-vs-formality trade-offs for sharing your work and your results with other collaborators and researchers.
    • A set of mechanism to enable you to communicate and collaborate with researchers outside of your institution more easily.
    • A set of useful features that you would have had to set up to build a useful research-oriented site anyway:
      1. Mailing lists
      2. Shared files
      3. A search engine

    Anyway, to answer the question :-). To make effective use of this site, you should be actively participating in a workspace. Typically, you would be a member of a small group of people working on a clearly-identifiable research area or software project. This group of people owns an area of this site, called a workspace, which is listed on the workspaces page. (All workspaces are also listed on the front page of this site.) They are also members of a corresponding permissions group.

    Now, the key to making (and encouraging) effective use of the site is understanding the different levels of interaction and formality available to you. There is no set formula enforced by the site, just a collection of mechanisms that allow you to choose and evolve the most effective way of working with them. Briefly:

    • Email lists are the most informal mechanism. Each workspace can have a number of mailing lists and automatic email archiving.
    • Forums and FAQs are the next less mechanism. These interactive web pages allow you to post information to this server to (selectively) share with other researchers.
    • The workgroup's "home page" is the next less formal mechanism. The home page is hand-authored (as opposed to automatically-generated) HTML, thus allowing you to present a polished and complete view of your research to the world at large.
    • Software releases and journal and conference publications are the most formal mechanism. You can post your publications to this site's publication database.

    How can I use the site to hold meetings better?
    The following was prompted by a real meeting that could have run a lot more smoothly. Problems such as everybody not having had access to the right materials prior to the meeting, problems connecting laptops to projectors, and similar things are very easy to avoid if you use the relevant forum. The following suggestions will be particularly helpful for scheduled meeting, especially if they involve preparation of materials for review or presentation of slides.

    Prior to the meeting, the meeting organize should create a new topic in the forum of the relevant group. (We'll assume that meeting attendees are member of the same workgroup or have appropriate access levels.) He or she should post initial material, such as an agenda and material to be discussed, and email the group so they know that the material is there.

    In the leadup to the meeting, attendees post any additional material to be discussed at the meeting in that topic, and send email. This way the other attendees will be able to review materials before the meeting.

    Anyone that is planning to present slides at the meeting should post their slides -- in powerpoint and perhaps PDF -- prior to the meeting. This ensures that (assuming that there is at least one computer that works with the projector) the slides will be easily brought up when needed, without needing to mess around with laptops and projectors.

    After the meeting, minutes, additional materials, and discussions can all be posted to the same forum topic. You will have a complete and invaluable record of the meeting and ensuing followup. Please give it a try.

    General questions about the web-site
    General information about the web-site.

    Who do I contact if I have problems with the site?
    The first port of call is this FAQ. If that doesn't answer your questions, then send mail to webmaster at gigaascale dot org.

    Is the site address gigascale.eecs or gigascale.org?
    As of September 13, 1999, the canonical name of the website is www.gigascale.org. You can also use the address gigascale.org, which will be redirected to the canonical name www.gigascale.org.

    Not that the old name, gigascale.eecs.berkeley.edu is also being redirected to www.gigascale.org. However, this may cease to function sometime in the future. Please update any links or bookmarks to this site to reflect the new name.

    The canonical server name for all mailing lists is now gigascale.org. For example, mail to report problems with the website should be sent to webmaster at gigascale dot org The server name gigascale.eecs.berkeley.edu will continue for function for the mail lists, at least for the time being.

    Note: As of September 15th 1999, the return addresses on mail from the server are still set to gigascale.eecs.berkeley.edu. This is a configuration problem with sendmail, which we are working on. When you reply to a message from a mailing list, just ignore the fact that the return address is not the new canonical address.

    What is "https"?
    The https:// string at the start of a URL indicates that the connection is being made through SSL, the secured sockets layer. In other words, it is encrypted.

    Originally, gigascale.eecs used encryption for all pages once a user is logged into the site. However, this causes (we discovered...) problems with some browsers and is also substantially slower. So SSL is only used for login pages, administrative pages, and for workgroups that are set by their administrator to use SSL for security/privacy reasons.

    A side-effect of this mode of operation is that you may get warnings about entering and leaving secure pages. If you prefer, you can disable these warnings by changing your security preferences on your browser. Unfortunately, there is no way to disable them for gigascale.eecs but leave them enabled for other sites.

    Who developed this site?
    Present

    The web site is managed and maintained by the GSRC Infrastructure Team. See the infrax home page on this site for current information on the members of the infrastructure team and other GSRC infrastructure information.

    History

    This site was initially built, from scratch, by John Reekie in the spring and summer of 1999. The database and the code that drives the whole site was designed and written by John Reekie.

    The site's graphics were designed by Aaron Walburg. The look-and-feel was designed by Aaron Walburg and John Reekie.

    The web-based video presentations were initially set up by Christopher Hylands and John Reekie. Thanks to Dan Fays and Jim Balazic of Microsoft Research for showing us how they do their web-based presentations, this saved us a lot of trouble and confusion.

    The content of the site is, as much as possible (and in accordance with the philosophy of the site), provided by the members of the GSRC who have accounts on this site! Other content was authored by Professor Richard Newton, Christopher Hylands, and Lorie Brofferio.

    The servers running this site, a pair of Solaris Ultra/60s running Apache and CVS, and a Dell 6300 running IIS and the Microsoft NetShow server, were built by Christopher Hylands and John Reekie. They are now secured and managed by Christopher Hylands and Marvin Motley.

    How big is the website?
    Obviously, the website is growing every day.

    As of 7/12/02, we had 1182 gsrc members, and the software written at GSRC to implement the website contains about 41,000 lines of PHP code, 36,000 lines of Java code, 7,000 lines of JSP code.

    The exact size of the website is a little tricky, since the website consists of email, papers, forum questions, software trees and video presentations.

    Menus are missing/Javascript is disabled?
    There are two known issues affecting the display of the dropdown menus at the top of the page.

    First, this website uses javascript to provide drop down menus.

    Unfortunately, there is a security hole in some versions of Internet Explorer that can be exploited by having javascript enabled.

    Microsoft suggests setting the security level to high and then adding specific sites to the list of trusted servers:

    1. In Internet Explorer, select Tools -> Internet Options -> Security
    2. Click on Trusted Sites -> Sites
    3. "require server verification" should not be checked
    4. Add the following hosts:
      http://chess.eecs.berkeley.edu
      http://embedded.eecs.berkeley.edu
      http://www.gigascale.org
      

    Secondly, security software like Symantec Client Firewall may remove some content on the assumption that it is unwanted advertising. Here is how to configure Symantec Client Firewall to stop it from removing our dropdown menus:

    Bring up Symentec Client Firewall
    Double-click "Ad Bocking".
    Click "Advanced".
    Click "Add Site".
    Enter this website's URL, click "OK".
    Scroll down to this website's URL in the left window, and select it.
    Click "Add..."
        Click the "Permit" radio button.
        Enter "banner" in the text area.
        Click "OK".
    
    Click "Add..."
        Click the "Permit" radio button.
        Enter "sponsor" in the text area.
        Click "OK".
    

    What if I change companies?
    What happens to the GSRC account if one shifts jobs from one of the sponsoring company to a non-sponsoring company? What if one wants to hold the account even after such a move?

    Our sponsors are concerned that products of GSRC research should be of benefit to themselves, and not to their non-sponsoring competitors. For that reason, if you change jobs to a non-sponsor company, we must disable your account. There may be exceptions to this, at the determination of our directorate.

    About authoring and posting publications
    This section of the FAQ is for information about authoring publications and presentations, and about posting authored works to the website.

    How do I handle prepublication of papers?
    The GSRC contract with MARCO states:
    • "Any technical paper, article, publication or announcement of technical advances generated in connection with work done under this contract" must credit support from Marco. See GSRC Word Templates, which includes the appropriate acknowledgement that must be included.
    • A copy of each paper that describes subject matter that is funded by MARCO should be submitted to http://fcrp.src.org/member/act/FormPublication.asp 60 days in prior to the date of proposed publication or dissemination.
    • MARCO has the right to request changes surrounding intellectual property or confidentiality issues.

    The GSRC contract with MARCO states:

    "5. Submisson and Dissemination of Works of Authorship.

    5.1 Publication
    Contractor shall assure that Contract Performers, including subawardees, when publishing papers subject to and in accordance with the provisions in Section 5 or otherwise disseminating information regarding technical developments or research findings made by Contractor or subawardees pursuant to the Focus Center Award shall acknowldege, as appropriate, financial and program support of MARCO on behalf of Participants. Contractor shall assure that all reasonable efforts are undertaken to identify and where appropriate, procure appropriate intellectual property protection relating to MARCO-funded subject matter prior to any actual publication of dissemination of such subject matter pursuant to the Focus Center Award. A copy of each paper or resume which embodies the information proposed to be disseminated shall be submitted via the URL address specified by MARCO, http://fcrp.src.org/member/act/FormPublication.asp by the Center Director to the account provided by MARCO, to MARCO for review and comment, not less than sixty (60) days prior to the date of the proposed publication or dissemination. MARCO may request that reasonable changes or deletions be made in any proposed publication or dissemination to protect any intellectual property disclosed therein or to protect confidential information of a third party. If MARCO believe that any subject matter to be published or disseminated warrants intellectual property protection, MARCO will identify such subject matter and notify Contractor pursuant to Section 5 ("Submission and Dissemination of Works of Authorship"). Publication or dissemination of such subject matter may be delayed an additional sixty (60) days assuming reasonable cooperation from Contractor Inventors so that the Contractor or MARCO may secure appropriate intellectual property protection for the identified subject matter, include but not limited to the filing of a United States patent application, copyright registration or mask work registration, in accordance with Section 10 ("Intellectual Property") and 11 ("License Grant by Contractor") of this contract."

    See also the short slide set describing FCRP requirements and procedures for posting publications, FCRP Technical Paper and Conference Presentation Submission Requirements.

    Section 5.2 is reproduced in How do I release software?

    About Powerpoint Slides

    About GSRC Slide Presentations

    Templates

    A standard GSRC Powerpoint template is available at http://www.gigascale.org/pubs/gsrc-downloads/templates/ppt/index.html, or by following links from the website menu bar:
    technical -> publications -> GSRC Powerpoint Templates
    You should find the template to be configured to show the date and title of the next GSRC workshop. If you need to change the template, see this FAQ.

    Logos

    Current logos are available for inserting into your slides at http://www.gigascale.org/pubs/logos/, or
    technical -> publications -> Logos
    Especially, please be aware that the old GSRC logo has been replaced by a new one, and the old red MARCO logo has been replaced by the new FCRP logo.

    The Powerpoint template has the wrong date.

    To change the date and footer in GSRC PowerPoint template:

    The standard GSRC Powerpoint template file, found at
    technical -> publications -> GSRC Powerpoint Templates
    should be found to have the title and date of the next GSRC workshop already placed in it. If you find that not to be the case, or if you need to change the contents of the footer and date for your own purposes, you can follow these instructions on your own downloaded copy of the .ppt file:
    • Open the .ppt file in PowerPoint.
    • View -> Header and Footer
    • Make the changes in the dialog provided, click "Apply to All".
    • View -> Master -> Handout Master
    • Text in the upper-left-hand box (Header Area) should read, "Gigascale Systems Research Center". If not, select text and change it. (Font should be Arial Narrow, 12 pt., Bold)
    • Text in the upper-right-hand box (Date Area) should be the title of the event (e.g., "GSRC Quarterly Workshop", or "GSRC Annual Symposium"), and the date(s) of the event on the second line. If not, select text and change it. (Font should be Arial Narrow, 12 pt., Bold)
    • View -> Normal
    • Save.

    Questions about the forums and FAQs
    Questions and answers about forums and FAQs: what they are and how to use them.

    How do I add an article to a forum?

    Some definitions: A workgroup has a public forum and a private forum. Access to the two forums can be controlled, see How do I control access? A forum contains topics. A topic contains articles and tasks. An article can have uploads. One way to think of the structure is that each workgroup is a directory, and each forum topic is a subdirectory, and each article is another subdirectory that can contain multiple uploads.

    If you have permission to add an article in a forum, you will see a link labeled "Add article" in the left-hand of any page in that forum. Click on that link.

    If you were at the top level of the forum, you will see a page with the heading "Choose the topic to add this article to." Beneath the heading is a list of titles of topics in the forum. Choose one of these topics by licking on the appropriate link, and you will be taken to a page that you can use to create the new article (see below).

    If none of the topics listed are suitable, there is a form further down the page labelled "Or create a new topic." To create a new topic and then continue to a page that you can use to add an article, fill in the topic's title and (if you wish) a short abstract, and press the Add topic button.

    If you were already viewing a topic or article when you clicked the "Add article" link, you will already be at the page for adding an article.

    In either case, you are now at a page labelled "Add a new article." There are a number of fields that you can fill in to add your article. Note that you will be able to change any of these fields after adding the article, so there is no need to get them exactly right if you are unsure.

    • Full title. The title of this article, which will show up in the Forum contents listing and when people visit your posted files.
    • Short title. The short title is what shows up in the left-hand margin of all forum pages. Type in a word or two here, or leave it blank and the first two words of the title will be used.
    • Attached files. Any article in the forum can have additional files attached to it: PDF files, GIFs, Powerpoint slides, and so on. If you wish to allow files to be uploaded and attached to this article, choose the Allowed option (the default). Otherwise, choose the Not allowed option.
      Note: you will be able to attach files after creating the article. To attach a file, view the forum article and use the tools that appear at the bottom of the page.
      For more information, see How do I attach or upload a file to an article?
    • Article text. If you want to upload an existing file that contains HTML or plain text that explains something about the files you are uploading, click on the Browse button and locate the file.
      Note that this upload is only for the text of the article that is displayed when the article is visited, it is not where one would upload larger attachments. This upload should only be plain text or HTML.
      If the article is created with attachments enabled, then larger attachments are added after the article is created. These attachments can be of any type.
    • (or enter text here) If you want to enter text that explains something about the files you are uploading, type HTML or plain text in here. (This text is used only if the Article text field is blank.)
    • Text format. In general, the HTML option works better. Use the plain text option only when pasting existing plaintext. (Note: this option applies only to the text entered in the above field. When uploading a file, the text form is automatically set from the file type.)
    To create the article, click on the "Add article" button.

    Important: don't forget to click on the "Add article" button or nothing will happen on the server.

    How do I attach or upload a file to an article?
    There are two types of attachments:
    1. When an article is created, one can upload plain text or html to be used as the body of the article.
    2. After the article is created, one can upload one or more named files of any type, for example, PowerPoint.

    When an article is added one can select whether file uploads are allowed.

    If you upload a file with the same name as an already existing file, then the contents of the file is replaced with the contents of the upload.

    Note that there is currently a 5Mb (actually 5,242,880 bytes) limit to the size of the attachment that can be uploaded. (The size is set in /usr/local/lib/php3.ini) If you would like to share larger files, then you may want to set up your workgroup to use CVS authoring. For details, see: How do I edit pages in a group with the "CVS Authoring" option?
    Don't forget that if you are checking in binary files, you should use cvs add -kb filename so that CVS knows that the file is a binary file.

    For information about sharing files via email, see: Why doesn't the GSRC Mailman system handle large attachments?

    How do I edit an article in the forum?
    You can edit any article for which "you have Modify permission, or for which you are the original author."

    What that actually means is that, if you are looking at an article in a forum, and you have permission to edit it, you will see a link named "Edit article" in the left-hand margin". If this link is there, click on it to go to a page that will allow you to edit the article.

    The fields on this page are the same as for the Add article page that you used to create the article. Edit any of the fields, such as the title, text, text format, or attached files buttons, and press Update article to change the article.

    There is one additional item on this page, which allows you to use a better editor to edit the text of the article. This is the "Download text" link to the right of the Browse button. If you click on this link, you will be able to download the file onto your own machine, where you will be able to use emacs or <insert favorite editor here> to edit the file. When you are done, click on the Browse button to select the file and upload it again.

    (Note: If a file is selected with the Browse button, the text in the text field labeled "(or edit this text)" will be ignored.

    Important: don't forget to click on the "Update article" button or nothing will happen on the server.

    How do I delete an article or task from a forum?
    To delete an article or task, go to Reorder Contents in the left hand column under Editing Command and follow the instructions:
    To delete an item, set its Section to -1. It won't actually be deleted, but it will disappear from all views of the Forum. Topics cannot be deleted in this way; instead, just recycle them by changing the title and text.

    Why are the URLs of forum articles numbers, and not something more meaningful?
    If you used the forums or FAQs prior to August 30th, 1999, you may have become used to seeing URLs in the forum like this:
        http://.../forum/topic-string/article-string.html
    
    The problem with this URL scheme was that the URLs would change: firstly, the id strings would change, and secondly, when we implemented support for moving items around in the forum and faq the URLs would definitely change.

    Se we decided to numerically number all items in the forums and faqs. For example:

        http://.../forum/24.html
    

    These new numeric URLs will never change, regardless of how much you modify or move around your forum or FAQ. So you can now "link with confidence" :-)

    (Some people do prefer to non-numeric URLs, but we surmised thta most people don't really need to look at the URLs anyway. For example, amazon.com has a very large and usable site, and their URLs are very long and obscure.)

    I don't have time to check the forum every day! How can I get the server to notify me when a new article is posted?
    One of the standard mailing lists provided for each group is a notification mailing list. This list contains exactly the members of the corresponding group. If the notification list for, say, group foo is turned on, then all posts and changes to the forum and FAQ of foo will send out mail to the members of foo.

    To turn on the notification mailing list, the group administrator needs to go the Admin page for that group, click on the Configure link, select the "Member notification list" checkbox, and press the "Change group configuration" button.

    Tip for group admins: if you have a notification mailing list and a public interest list, you may wish to subscribe the notification list to the interest list, so that all mail to, say, foo-interest, goes also to foo-notify.

    As of August 1999, the notification mail does not include the contents of the articles or FAQs. We will be supporting this ability at a later time.

    How do I create a "task"?
    Tasks are located in forums, together with regular articles and posted files. If this is a brand-new task, adding a task is essentially the same procedure as for articles (except that the link in the left-hand margin that starts the procedure is labeled "Add task," of course!)

    The form for creating a task is similar to that for creating an article, except for two items:

    1. There is no Attached files radiobutton. This is because uploaded files cannot be attached to tasks.
    2. There are two additional fields at the top labelled "Status" and "Assigned to." Use the menus to choose a status and to assign the new task to someone in the group that owns this forum. If the menu entries do not contain what you want, simply type a value in the text field underneath the menu.
    Press the Add task button to create the new task.

    How do I update a task?
    Tasks can be updated directly on the task view page. As long as you have permission to modify a task, your view of that task will include menus showing the task status and who it is assigned to. To change the status or assign it to someone else, change the menus (or the text fields underneath) and press the Change task status button.

    Each task has a history, which is displayed underneath the description of the task. When you update a task, you can also enter text in the field labeled "Your comments" -- this text wil be added to the history of the task.

    To change the description of a task, view the task and press the "Edit this task" link in the left-hand margin. On the task editing page, you will be able to change the task status and assignee, change the task's title and description, and add an entry to the task history. Press the Update task button to make these changes take effect.

    Hey, these tasks are pretty lame! What's the point?
    Well, this is a simple feature that we provided because we think it's useful. It's not supposed to compete with Project or whatever, and here's why:

    The goal of this site is to support collaboration between geographically dispersed researchers. All of the facilities on this site are there because they help this goal! If you are able to coordinate and manage your collaborative work using some other tool, great! However, sometimes it's just easier to use a simple web-based tool than something more elaborate, so we provided one.

    We have found that tasks are useful for simple to-do lists, and for recording future tasks in a place where it's hard to lose them (as it is with email, for example).

    If you try it, let us know if it is (or isn't) useful.

    How do I add to my group's FAQ?
    A FAQ (Frequently Asked Questions) list is a common and effective format for providing a lot of detailed and targeted information. If you have a group on this website, the group administrator can turn on a web-based FAQ interface for that group. This is likely to be particularly useful for software projects, as most software products have a FAQ as a key part of that software's documentation.

    FAQs are divided into sections, where each section contains questions (and hopefully, answers!). To add a question to a FAQ, the procedure is essentially the same as that for articles, except that the link in the left-hand margin is labeled "Add question," and other labels and buttons change accordingly. The procedure for editing a question in the FAQ is also essentially the same as editing an article. (The main difference is the absence of the Attached files option -- it is not possible to upload files to attach to a FAQ.)

    If you are a GSRC member, you can also add questions to the GSRC FAQ.

    How do I reorganize my FAQ?
    To reorganize a FAQ, click on the "Reorder contents" link on the left-hand margin. You will see a complete listing of all sections and questions in the FAQ, with editable fields that enable you to change the parent section (aka topic) of a question and to change the ordering of the topics and questions in a FAQ. Follow the instructions on that page.

    Note the following:

    • It is not possible to delete a topic. Instead, just "recycle" it.
    • There is no connection between the "order id" and each item's regular id.
    • The order ids do not have to be continguous. For navigation to work properly, however, they do need to be unique within each topic.
    • Forums can be reorganized in exactly the same way (although it's less likely that you will need to do so).

    Can I get a printable version of a FAQ?
    Yes, you can. Go to the FAQ in which you are interested and click on the View -> Fully expanded link in the margin on the left-hand side of the page. You will get a single HTML page containing the entire contents of that FAQ, which you can then print out.

    If you only want to print the contents of a single section of a FAQ, click on the View -> By section link on the left-hand margin. Location the section you are interested in and click on its title. On the right-hand side of the page, towards the top, you will see a small link labelled "Expand." Click on it, and you will get a single page containing the whole contents of that section of the FAQ.

    (Note: you can follow exactly the same instructions to get printable versions of a forum.)

    See also Fully Expanded visibility

    When I use "Fully Expanded", not all of the text is visible
    Sometimes, if you select Fully Expanded, not all the text of the faq or forum will be visible, you will have to use the horizontal scrollbar to see the text on the right side. This can be a problem if you are trying to print out the faq or forum

    This will occur if someone added a question or forum article or task in Plain Text instead of in HTML.

    The fix is to send email to www@gigascale.org and ask that they change the mimeType for that article in the history table for your workgroup from text/plain to text/html

    Note to admins:

    1. Go to http://www.gigascale.org/mysql
    2. Click on the workgroup that needs adjusting
    3. Click on the history link
    4. Browse to the article in question
    5. Change the mimeType field from text/plain to text/html

    It would be nice if we could default to just using html, but people tend to type in text with blank lines as paragraph separators, and if we display this text in an html viewer, it will connect all the paragraphs together. This is a bug in the current system which we hope to fix some day, so in the short term, you are better off if you use the HTML mode.

    How can I share a FAQ item between websites?
    It is possible to share FAQ items between the GSRC-related websites, GSRC, Embedded and Chess.

    To add a FAQ item to your workgroup that has the same content as an existing FAQ on one of the other websites, your website account must have permission to read the existing FAQ on the other website, and permission to create one in its new location.

    Simply go to the FAQ section where you want to add the new FAQ, and click the "Add question" link in the left margin. On the resulting page, read the brief instructions, and enter the URL of the existing FAQ item in the box provided for that purpose.

    Since the text of the FAQ is shared between two websites, the wording of the FAQ should be general enough to apply to both sites.

    Editing group pages, and other forms of rocket science
    Information on how to set up and edit group home pages, and other advanced features of the site.

    Some of my group's pages don't show up properly. Why?
    Because the gigascale server takes pages that you authored and "wraps" them in its own HTML to generate the GSRC header, footer, and on on, authored pages have some (reasonable) restrictions on what they can contain.
    • Frames don't work. In theory, frames could be supported but it hasn't been a high enough priority yet. Please just write simple flat HTML pages and let the server add the headers and other decoration. See the customization FAQ if you would like to have the server add your own margins or footers to your pages.
    • Pages need to be "well-formed." Because we are processing your HTML and embedding it within the GSRC pages, it's a lot easier for the server to extract what it needs from your pages if it contains correctly-formed HTML. The overall structure of the page should be like this:
      <html>
      <head>
      <title>My Title</title>
      </head>
      <body>
      ...
      </body>
      </html>
      

      Other things that can prevent your page from displaying are:

      • No <body> tag.
      • Too many <body> and </body> tags.
      • Malformed tables. Each <table>, <tr>, and <td> tag must have a matching end tag.
      • Igor Markov points out that another trick is to read a page into Netscape Composer and then write it out again. This technique is especially useful in fixing table errors.

    Links

    When you are authoring your own pages, you are better off using relative links whereever you can. Relative links allow a copy of your pages to be viewed from anywhere.

    If you must use an absolute link, consider using

    < a href="/workgroup/myfile.html">myfile.html</a>

    instead of < a href="http://www.gigascale.org/workgroup/myfile.html">myfile.html</a>

    HTML links from forum pages to the main workgroup pages should take into account that the main workgroup pages are can be found in /workgroup>/, whereas the forum is in /workgroup>/forum.

    Detection

    The GSRC Website uses the Htdig search engine. The search engine is run each night, and a log file is generated. We run a short script over that log file and send email about broken links to webmaster. The webmasters then either fix the links themselves, or send email to the authors.

    If your pages are publicly readable, you can use one of the services from the Yahoo HTML Validation and Checkers Page Most of the services cost money, and most of them can only read pages that are accessible to the public, but some services provide a free sample run. (The GSRC website has been specially modified to allow out htdig process to be able to index non-public pages).

    If you have a remote subsite, see also the remote subsite faq.

    My remote subsite doesn't work! Why?
    Remote subsites are a little tricky, because we are making the contents of your site appear to be on the GSRC website. Because of this, there are some restrictions on what you can do in a remote subsite. In two categories:

    How to structure your page

    • The remote URL that you give in your group configuration page must be a directory on your server, not a single HTML file. For example,
        http://myserver/mypath/myfile.html
      
      will not work. It needs to be
        http://myserver/mypath/
      
    • Links within your subsite must be relative. For example, the index.html file that is served by the above path should have link such as these:
        <a href="anotherfile.html">A subsection<:/a>
        <a href="subdir/">Another subsection</a>
        <a href="subdir/somefile.html">Link into a subsection</a>
        <img src="images/foo.gif"> An embedded image
      
    • Generally, it is preferable to place all the contents of your subsite underneath the root of your HTML tree. This makes maintenance etc easier. However, if you do have images or binary files that are above the location of your HTML tree, then you can reference them by using the full server path:
        <img src="http://ping.pong.edu/images/foo.gif">
      
      (Note: do not do this with HTML files!)
    • If you are having trouble with URLs, try turning on the "Fix HTML" option in your group configuration pages. This will make the server perform some extra parsing of your pages to try and reduce problems caused by not-quite-right HTML.

    Things you can't do for technical reasons

    • Frames. Because the GSRC server takes the contents of your page and "wraps" them up with its own headers and footer, it really really wants to see a simple HTML page. For this reason, pages with frames will not come through properly.
    • Badly-formed HTML. Because we are processing your HTML and embedding it within the GSRC pages, the server requires correctly-formed HTML. In particular, things that can prevent your page from displaying are:
      • No <body> tag.
      • Too many <body> and </body> tags.
      • Malformed tables. Each <table>, <tr>, and <td> tag must have a matching end tag.

      If you have pages that don't show up or are strange, check that they have the following structure:

      <html>
      <head>
      <title>My Title</title>
      </head>
      <body>
      ...
      </body>
      </html>
      
      This will make life much easier for the parser... You can also try turning on the "Fix HTML" option in your group configuration pages. This makes the parser do some extra work to try and make sense of odd HTML, but there is no guarantee that it will fix anything.

    How do I customize my margins and footers?
    The gigascale server provides a facility that allows the pages that you author and that appear in your workspace to be customized. To customize your pages, you need to:
    • Turn on the "Fancy HTML" option in your group's configuration options.
    • Create a file called "toc.html" in the root directory of your HTML files. (That is, in the same directory as the index.html file you see when you go to http://gigascale.eecs.berkeley.edu/groupname.)
    The toc.html file can contain anything you want. However, the server parses it looking for pairs of HTML comments that look like this:
    <!--margin-->
    My margin text
    <!--/margin-->
    
    If the parser sees a block like this, it extracts the text between the comments and places it in the left-hand margin of the displayed page, instead of the links that it would normally generate.

    You can also append text to the existing auto-generated margin text by inserting the "+" sign, as follows:

    <!--+margin-->
    Text appended to margin
    <!--/margin-->
    
    Several fields can be specified in this way:
    • margin: text is inserted or appended to the left hand margin that appears on every page.
    • footer: text is inserted or appended to the footer that appears on every page.
    • header: text is inserted or appended to the header links that appear on every page.

    In addition, a block named options can be used to specify additional formatting and display options. the syntax is

    <!--options-->
    name1=value1
    name2=value2
    <!--/margin-->
    
    Currently the following options are supported:
    • margincolor: The background color of the left-hand margin.

    Once you have a custom setup for your site, you may wish to further customize subdirectories in your site. You can do this simply by placing another "toc.html" in sub-directories that you wish to customize. Note that you will need to duplicate all fields in your top-level file -- for example, if you put a footer in your top-level toc.html, you will have to put one in the sub-directory's toc.html too.

    Note also that toc.html is not "inherited." If /diva has a toc.html, and /diva/x has a toc.html, but /diva/x/y does not, then files in /diva/x/y will use the toc.html from /diva, not the one from /diva/x.

    How do I edit pages in a group with the "CVS Authoring" option?
    First, bear in mind that CVS authoring is generally useful for software projects, in which case we assume that you know the basics of CVS and can use SSH. For further information about CVS, see the GSRC CVS FAQ
    1. If you are a group administrator check that your group has cvs authoring turned on by going to your group pages and then clicking on the Admin link and then Configure Group link and verifying that CVS Module and CVS Checkin are selected.
      If these two choices are not selected, then select them and hit the Change Group Configuration button at the bottom. This will send email to webmaster, who will then set up your cvs repository and send email back to you. While you are waiting, you can proceed with the steps below.
    2. If you don't yet have a CVS account on gigasource.eecs.berkeley.edu, get one. You can request a CVS account.
    3. If ssh and cvs are not yet installed, then install the SSH client and CVS
    4. For Unix, download from ftp://ftp.cs.hut.fi/pub/ssh/
      Build and install. Note that you need not install the server side, you will only be logging out of your machine, not logging into your local machine.
      Download, build and install CVS. To download CVS, see: http://www.cyclic.com/cyclic-pages/howget.html
    5. For Windows, you can download cvs and OpenSSH as part of Cygwin. There are two methods of installing Cygwin: Downloading from the Ptolemy II website or downloading directly from the Cygwin website.
      The advantage of downloading Cygwin from the Ptolemy II website is that the download is somewhat smaller since only essential parts of Cygwin are included - hence this download would be useful if you are dialed in over a slow modem.
      The advantage of downloading from the Cygwin website is that you will get the most recent versions of the software.
      Download from the Ptolemy II Website You can download the version of Cygwin that is included with the most recent Ptolemy II distribution from http://ptolemy.eecs.berkeley.edu/ptolemyII/ptIIlatest/cygwin.htm
      OR
      Download from the Cygwin website Install the Cygwin toolkit from http://sources.redhat.com/cygwin/

      Complete installation instructions can be found at http://sources.redhat.com/cygwin/faq,

      The Cygwin installer tends to change over time, but below is a summary of the steps necessary:

      1. Create an empty directory
      2. Download http://sources.redhat.com/cygwin/setup.exe and save it in the empty directory
      3. Click on setup, then Next
      4. Select Install from Internet, then click on Next
      5. Use the default install root directory, which should be c:/cygwin
      6. Select 'Default Text File Type' of DOS.
        If you select Unix instead of DOS, then you will likely have Cygwin CR/NL problems.
      7. Hit the Next button.
      8. Use the default Local Package Directory, which should be the directory you created above.
      9. If you are not behind a firewall, click on direct connection, then click Next
      10. Select a nearby mirror, then click Next
      11. You will be presented with a list of packages to install.
        A complete installation takes at least 75Mb.

        At the minimum, select the packages below, which is roughly a 3Mb download that expands into an 8.7Mb installation.

        • Base Category
          • ash - For /bin/sh
          • bash
          • cygwin
          • diff - Configure uses cmp
          • fileutils - For ls
          • grep - Used by configure
          • sed - Needed by config.status, which is created by configure
          • sh-utils - For basename
          • textutils - For cat and tr
        • Devel Category
          • cvs
          • make
        • Net Category
          • openssh

        If you are serious about development you may also want:

        • Devel Category: autoconf - requires gawk and m4 from the Interepreters Category
        • Base Category
          • findutils
          • gzip
          • tar
        • Text Category: less
    6. Make sure that CVS_RSH is set to use ssh.
      To check under Windows, start up a Bash shell (Start->Programs->Cygnus Solutions->Cygwin Bash Shell) and type echo $CVS_RSH. If it is not set, then set it using Start->Settings->Control Panel->System->Advanced->Environment Variables CVS_RSH should be added and set to ssh.
      You may want to check that your path has been set to include Cygwin, which is located at c:\cygwin\bin

      To check under Unix, start a shell and type echo $CVS_RSH. If CVS_RSH is not set, then edit your shell startup files to set it automatically. If you run C-shell (csh), or some variant like tcsh, edit ~/.cshrc and add

      setenv CVS_RSH ssh
      
      If you run the Bourne shell (sh) or some variant like bash, edit ~/.profile and add
      CVS_RSH=ssh
      export CVS_RSH
      
    7. Create a file in your home directory called .cvsrc that contains:
      update -P -d
      
      This is necessary so that when you run cvs update, then the -P and -d flags are automatically appended. cvs update -P -d will prune any directories that contain no files and create any new directories that someone else has checked in.
      If you do not create a ~/.cvsrc file, then you will need to run cvs update -P -d instead of cvs update
    8. Once you have received email stating that your cvs account is setup, log on to the cvs server with ssh gigasource:

      ssh gigasource.eecs.berkeley.edu

      If your gigasource login is different from your local login then you should -l <i>username</i>:

      ssh -l username gigascale.eecs.berkeley.edu

      Once successfully logged in change your password with the passwd command to something more reasonable
      ptolemy@maury 67% ssh gigasource
      ptolemy@gigasource's password:
      
      Last login: Fri Aug 27 17:50:34 1999 from maury.eecs.berke^M
      No mail.
      $ passwd
      passwd:  Changing password for ptolemy
      Enter login password:
      
      New password:
      
      Re-enter new password:
      
      passwd (SYSTEM): passwd successfully changed for ptolemy
      $ exit
      Connection to gigasource closed.
      ptolemy@maury 68%
      
      
    9. If your group is called foo, then check out the foo (for example) module from CVS with:
          cvs -d :ext:gigasource.eecs.berkeley.edu:/home/cvs co foo
      
    10. The CVS repository for each workgroup has a web subdirectory that contains the website. There may also be adjacent directories that contain software projects.
      To change the website, cd into the foo/web directory and edit a file
    11. Commit your changes with
      cvs commit -m "message about your change" fileyouedited
      
      You will then be prompted for your password, and the change will be committed.
    12. Go to a page in that workspace. You will see an "CVS Update" link in the toolbar at the bottom of the page. Click on it.
    13. Press on the "Go back" link and make sure that the pages are what you expect.

    How do I update cvs-authored pages?
    You can only update the Web pages of a CVS-authored group if you are member of that group and are logged into the server.

    Assuming you are, go to the group's home page and press the "CVS Update" link at the bottom right of the page.

    Also, make sure that the group admin has selected "Home Page" in the Admin->Config page so that the website knows that you want to use a home page other than the initial default home page.

    Note: If you have made an error in your CVS web pages such that the home page of your group no longer shows (you accidentally checked in a page with frames, or you removed the standard footer, for example), you can update the page "manually" as follows:

    Add the string "?Action=cvsupdate" to the URL of your home page and press Enter. For example, to update the home page of group foo, enter the following in the Location box of your browser:

    http://gigascale.eecs.berkeley.edu/foo/?Action=cvsupdate
    

    How do I learn more about CVS?
    See the Softdevel CVS FAQ

    How do I find bad links in my group web page
    The search engine gets run every night and generates a list of bad links in http://www.gigascale.org/gsrc/private/9.html that can be viewed only by GSRC members.

    You can also use the wget command, but you will need to set it up to use the cookie file from Mozilla.

    1. Install wget
    2. Log in to the website using Mozilla and then exit Mozilla
    3. Find your cookie file. Mine was at c:/Documents and Settings/cxh/Application Data/Mozilla/Profiles/default/lwhpscha.slt/cookies.txt
    4. Copy the cookies.txt file to a place with a shorter name.
    5. Run wget:
       
      wget -r --load-cookies cookies.txt -np http://www.gigascale.org/yourGroup
      
      This will produce a directory called www.gigascale.org that contains the contents of yourGroup
    6. Look for Not found in the output
    7. If you find a file that was not found, then grep the files for that file. For example, if foo.htm was not found, we would do
      find . -name "*.htm" -print > /tmp/files
      grep foo.htm `cat /tmp/files`
      

        What is download logging?
        A downloadable file, such as an attachment to a forum article, may be flagged so that any time someone downloads that file, an entry as added to a log showing the time of the download, the name of the file, and the name, email address, and organizational affiliation of the person downloading the file.

        If the person is logged on to the website, the information is taken from the user database, and logging will be invisible to the user. If the user is not logged in, s/he will be presented a form and asked to provide the information explicitly before the file is transmitted.

        Forum attachments

        The download logging feature is turned on or off for a forum attachment by the workgroup administrator, who simply edits the forum article and checks or un-checks a box next to the name of the attachment.

        Other files

        A downloadable file doesn't need to be a forum attachment to be logged. The URL one would normally use in a hyperlink to download a file can be modified to enable logging. The modifications are these:
        • Prepend "/download" to the beginning of the URL.
        • Append "?package_name=whatever" to the end of the URL.
        For example, if some software package can be downloaded by clicking a hyperlink that looks like this:
        <a href="/mygroup/blah/greatNewThing.tar.gz">Click here</a>
        
        To log downloads of this file, change the link to look like this:
        <a href="/download/mygroup/blah/greatNewThing.tar.gz?package_name=great_new_thing_1.0">Click here</a>
        
        The name used in "?package_name=great_new_thing_1.0" will be the name used to identify the download in the log.

        If you include the fully qualified URL,

        http://www.gigascale.org/mygroup/blah/greatNewThing.tar.gz
        
        (which is usually not recommended), then add "/download" just after the server name, "www.gigascale.org".

        Where is the log?

        The download log is viewable by administrative staff, that is, by members of the workgroup gsrcadmin from a link on the gsrcadmin home page.

    How to do group administration
    Questions about how to do things on the site.

    How do I start a new group?
    To start a new group, log in to the website and select Options in the upper right hand corner, and then select New Group under the Request menu on the left.

    Group names should be all lower case, so as to match the existing groups.

    After you fill out the form, your request will be forwarded to the GSRC website administrators who will authorize your group. Once your group has been set up, you will receive a piece of email with further instructions.

    What is a group admin anyway?
    Each group has one or more "admins", who are able to configure the facilities of that group, such as its Forum, FAQ, mailing lists, and so on. A group cannot be created without an admin, and that admin is responsible for managing that group, enrolling members into that group, approving (or not) requests to join that group, and in general responding to the needs of the people in that group.

    Being a group admin involves a certain amount of time involvement. Not a huge amount, because we have tried to get as much of the administrative overhead into the server code as we can. Nonetheless, it does involve some time (including, we hope, reading these FAQs!). If you don't have the bandwidth, please find someone else who you expect to be working with who does.

    The other side of this coin is that group admins have the freedom to configure and manage their groups. For example, they can give collaborators accounts on this site without needing to send email to some centralized "webmaster." They can turn on many of their group's features without having to wait for us to respond. (Unfortunately, we are limited in what we can do automatically in the Web server, so turning on features that require privileged access to the server machine will send email that requires a response from a person.)

    Note that some of the group admin interfaces generate things like passwords and secret keys that we need to supply to scripts that we run to add features to your group. Please use these forms if they exist -- if you just email us instead of using the web form, we have to impersonate you, fill in the form, and then read our own email!

    The goal of this approach is to make the GSRC Web Site scalable. By "out-sourcing" these (simple) administrative tasks to people who are working directly on the site, we hope that the site can expand to include many users and be a productive vehicle for your research goals.

    How do I set up my new workspace?
    Here is a quick how-to on initializing a new workspace:
    1. Log in to the GSRC web site.
    2. Follow the Admin link at the top right of the page. You will see your group in the left margin with a set of options. Under Admin, click on the Group Profile link and:
      1. Set the Group Title, One-Line Summary and Description.
      2. Set the Group Type to either Software Project, or Discussion Group. Note that if you do not select either one, then the new group will not appear in the main menu.
      3. Press the Change Group Profile button.
      Click on the Configure Group link at the left margin and:
      1. Select the "Is enabled" checkbox.
      2. Select the "Has members" and "Has administrator" checkboxes.
      3. Press the Change Group Configuration button.
      4. Choose other options and press the Change Group Configuration button For example, If you want a discussion forum, select the "Discussion forum" checkbox.
    3. Click on the Group Members link in the left margin.
      Add any new members by using the combo box under Add a new member and then press Add Member for each new member.
    4. If there are people you want to be in the group that do not have accounts on the GSRC web site, click on the Invite Members link in the left margin. Add their email addresses to the form and click on the Send Invitation button.

    My new group is not listed!
    You need to be sure that under Admin->Group Profile you have selected either Software Project, or Discussion Group and that under Admin->Configure Group you have selected Enable Group.

    For more information, see the How do I set up my new workspace? question.

    How do I control access to my workspace?
    Each section of a workspace (the part of the site devoted to a workgroup, such as its FAQ, forum, and so on), has a set of access control flags associated with it. The administrator of a workgroup can set these flags from the Admin page of that group according to the degree of access/privacy needed by the group.

    The access control flags are a two-dimensional grid, where one axis is the region of the workspace, such as the FAQ or forum, and the other is the class of user. User classes are as follows:

    • world. Users who are not logged into the site. As far as the server can tell, this could be anyone in the world.
    • gsrc. Users who are logged into the site as a GSRC member, but are neither a guest nor member of your group.
    • guest. Users who are logged into the site and have guest membership in your group.
    • member. Users who are logged into the site and are a member of your group.
    • admin. Users who are logged into the site and are are administrators of your group.

    Any given user has the highest class that applies to them (where "high" is alter in the above list). As a general rule, higher classes have higher permissions, where the permissions that can be assigned to each class of user are as follows:

    The flags are as follows:

    • Read. The user is able to read these pages.
    • Annotate. The user is able to add annotations to these pages.
    • Write. The user is able to add content to this part of your workspace. For example, a user with write permission can add an article to your forum.
    • Modify. The user is able to modify these pages (where the user interface permits). For example, a user with modify permission can edit an article in the forum. (Note: the author of an article can always edit it, even if they don't have modify permission in that part of the workspace.)
    • Execute. This is only useful in a few rare cases, and allows certain destructive operations that should only be done by an administrator.
    When you add a new feature to your workspace, you should check the access permissions carefully. Here are a couple of tips:
    • If the workspace is primarily for work in progress, you may want to make access more restricted, so that confidential results (for example) are not world-readable.
    • If you are a software development group, you may wish to make access more permissive. Allowing anyone to add to and annotate your forum and faq can be a valuable source of feedback from people who are using your software.
    We recommend that you err on the side of permissiveness when setting access permissions -- it is generally better in a research environment to have more information flow than less!

    Warning: incorrect default access permissions
    If your group was created before September 1st, 1999, the default access permissions set up for your group are incorrect.

    The correct default settings for each area of the group are

    world: Read
    gsrc:  Read Annotate
    guest: Read Annotate Write
    ...
    

    Prior to 1st September 1999, the default settings were

    world: Read
    gsrc:  Read Annotate Write
    guest: Read Annotate
    ...
    
    Please check your group settings and make sure that they are correct. (Go to your group's home page, click on Admin near the top of the page, and Access Control in the left margin.)

    What do the HTML Authoring options mean?
    If you have the "Home page" option selected, you can choose how to author that page. Currently, the available options are:
    • None. This is the default. When someone goes to youe group's home page, they will see a default page generated by the server.
    • CVS checkin. This means that you author HTML pages on your own computer and then check them into the GSRC code repository on gigasource.eecs.berkeley.edu. Before turning on this option, you muct have already turned on and received a CVS module for your group. See also the CVS authoring faq.
    • Remote URL. This means that your group's Web pages are fetched from a machine of your own. This is convenient if you already have some Web pages set up and it is easier to continue to author those pages. See also the Remote authoring FAQ.
    In the future, there may be a FrontPage option here as well.

    What is the "override" option?
    The administrative overrride option allows someone with suitable privileges to change their permissions without approval. The change of privileges only sets the permissions flags for that user for as long as they remain logged in, or until they choose to revert to their standard permissions. Joining a group using an override does not subscribe you to any mailing lists or appear on any member lists.

    Currently, only people that have Admin permission in the webmaster group can perform overrides. This is done from the memberships page of a logged-in administrator. Note that, after setting an override flag, your membership options will still indicate your "normal" membership, not your overridden one.

    This ability should be used sparingly, and only to correct errors and check for problems in pages that the admin would not normally have access to. After performing this task, always use the Restore permissions button at the bottom of the memberships page to revert to your regular permissions settings.

    How do I add members to a group?
    There are three methods:
    1. If the person already has a GSRC website account, go to the admin page of the group, and click on "Group members" link. Scroll down to the Add a new member section, select the person to add, select the membership privileges, and press the Add member button.

    2. If the person does not already have a GSRC website account, invite them to join the website. Go to the admin page of the group and click on the "Invite members" link. Fill in the form and click on the Send invitation button to send an invitation through email. The email will contain instructions for the new member; a copy will also be sent to you.

      Note that invitees do not get on the mail list until they respond to the invitation and actually get in the group. Nothing special happens when an invitation expires.

      You can also add the new member to the GSRC group as well (except in the unusual case in which you yourself are not a GSRC member). By default, membership in your group does not create a membership in the GSRC group; you can use the radio buttons on the form to change this.

    3. As a last resort (or if the person has already used the account request page to send you email), click on the "Create account" link on the group's admin page. Fill in the form and press the Create account button.

      Note: do not use this facility unless you really really have to. Read the text on the account creation page for more info.

    How do I reset a member's password?
    Any member can reset their own password. Just have them go to the login page and click the "reset password" link in the text below the form. That will take them to a form they can use to request that their password be reset. This will happen immediately, and they will receive email telling them what their new temporary password is.

    How do I disable someone's account?
    Unless you have webmaster admin privileges, you can't. You can, however, remove a person from a group for which you are an admin. Go to the group's admin page, click on the "Group members" link, and scroll down to the section labelled Remove a member from .... Select the name of the person and press the Remove member button.

    If you are a webmaster and you need to disable an account, go to the Options page and click on the Reset password link. Select the account to disable and press the Disable account button. Note: you will not be able to disable an account if this person still has any group memberships. If they do, you must either ask them to resign from all groups, or you will have to assume admin privileges for each group in which they are a member and revove them. (Yes, this is tedious, later on we could make this easy but it's very much a low-priority nicety.)

    How do I rename an account?
    In general, it is best not to try to rename accounts, but occaisionally, it is necessary. The best thing to do is to:
    1. Go to the People page and find out what groups the old account was in. Make a note of them, you will need them when adding the new account.
    2. Remove the old account from any groups.
    3. Disable the old account via https://www.gigascale.org/options/power/password
    4. Create the new account in gsrc workgroup and send an invite to the user.
    5. Once the user has logged in as their new account, then go to each of the groups that the old account was in and add them. You may need to use the power override to do this.
    Note that if you have the old and new accounts active at the same time, then you may run into problems because the workgroup invitation combo box lists the user by real name, not by account name, and if the user has the same name for both accounts, then you may have a hard time differentiating between the old and new account.

    Mail list administration
    Questions and answers on managing mailing lists. As much as possible, we are giving groups admins control over their mailing lists, but you do need to read these FAQs carefully!

    How do I create mailing lists?

    There are basically two kinds of mailing list you can get on the gigascale.org server:

    1. Standard lists. (See the rest of this question.)
    2. Custom lists. (See this question.)

    Standard lists are created by the group admin(s) on the admin page for that group. There are four types of standard list:

    1. group@gigascale.eecs.berkeley.edu. A private mailing list for the people on the group. Being a member of a group automatically puts you on the list, and there is no separate way of subscribing to this list. Guests of a group do not get this email.
    2. group-notify@gigascale.eecs.berkeley.edu. A private mailing list for notification of events on the group's subsite (additions to the forum, and so on). In general, if a group of people is actively working in a group this list should be turned on.
    3. group-devel@gigascale.eecs.berkeley.edu. A private for members of a group that have the "Developer" flag set. This flag is supported for software groups only.
    4. group-cvs@gigascale.eecs.berkeley.edu. A private mailing list for CVS notification. This is a private list that should only be turned on if you are doing software development with CVS on the GSRC servers.
    5. group-interest@gigascale.eecs.berkeley.edu. The public interest list for the workgroup. By default, the list is configured to be visible to the public and to allow anyone to subscribe, although this can be changed by the admin. Typically this list is used by people who are not directly involved in your research or software project but who have expressed interest in it. Eventually, we would like to subscribe guests of a group to this list automatically.
    6. group-announce@gigascale.eecs.berkeley.edu. The public announcement list for the workgroup. This list is intended to be a low-bandwidth announcement list for research results or software releases. By default it is configured as a moderated list.

    Note: in all cases, the person requesting the list becomes administrator of the list. This entails a certain degree of, well, administrative overhead that you need to be prepared to deal with. If you are not, please find someone else you are working with who is. Thanks :)

    How do I create a mailing list with a non-standard name?
    To have a mailing list created with a non-standard name, send email to webmaster at gigascale dot org with the name of your mailing list. The list name, if it contains a dash, must end in -cvs, -interest, or -announce. It appears that names like foo-bar-announce do not work, so use foobar-announce.

    The webmaster will

    1. Become root on gigascale and do
      cd ~mailman 
      bin/newlist mailing-list-name email-address-of-owner password
      
      For example
      bin/newlist gsrcworkshop-announce www@gigascale.org foobar
      
    2. Take the output and place it at the end of /etc/aliases and run newaliases
    3. Go to http://www.gigascale.org/mailman/admin/listname
    4. Edit Base URL for Mailman web interface (Details) and change it to http://www.gigascale.org/mailman
    The text below here documents a not yet implemented feature

    You can create a mailing list with a non-standard name. To do this, use the admin page of the relevant group, and follow the "New list" link in the left hand margin. In the bottom section of the page, there is a form that enables you to request a non-standard or "custom" mailing list.

    Warning: a custom list must be entirely configured by you, the server will not give you any help. If you do not believe you can handle hand-configuring a mailing list, do not use this option.

    When you submit the request, email will be sent to webmaster at gigascale dot org with your request. When the webmaster has created and enabled your list, you will be able to configure the list. We suggest you look at a standard list to get a feel for what the configuration should look like. In particular:

    • On the general options page, the hostname and base URL of the list will be wrong by default. Set these to gigascale.org and http://www.gigascale.org/groupname/ respectively.
    • [Add more here.]

    More on mailing list configuration
    Note the following carefully:
    • If this a private list (diva, diva-cvs, diva-devel, or diva-notify), do not use the Membership Management page to subscribe people. People are automatically subscribed if they are members of this group, and manually subscribing them will get them on the list but they will get weird configuration problems.
    • Leave the following fields on the General Options page blank:
      1. "A terse phrase identifying this list"
      2. "An introductory description"
    The following tips may be useful:
    • With your private member list (eg groupname@gigascale.eecs.berkeley.edu), you might prefer to set the option titled "Are replies to a post directed to the original poster or to the list?" to "List". That way you can reply to each message and everyone is automatically in on the discussion. (Don't do this for public lists!)
    • The options with the list title and description are now redundant, so don't bother setting them.

    Why do I get messages about "Implicit destination" from the mail manager?
    MailMan tried fairly hard to trap mail that looks like spam. If a message comes in that doesn't look it is correctly configured, MailMan will hold it for approval, and send you (the list administrator) mail requesting that you approve or discard that message.

    If you are forwarding mail from other lists or have aliases set up elsewhere to point to your list, then you will need to explicitly tell MailMan not to reject these messages. Go to the list administration page and select the "Privacy options" page. In the field labelled "Alias names (regexps) which qualify as explicit to or cc destination names", add the name of the forwarding list or alias that is sending to this list.

    For example, if you have an alias foo-list@mymailserver set to point to a-gsrc-list@gigascale.eecs.berkeley.edu, then the configuration option for a-gsrc-list needs to have "foo-list" added to it.

    How do I send email to all the GSRC Principal Investigators?
    All of the Funded GSRC Principal Investigators are members of the faculty workgroup.

    To send email to this group, use the faculty mail list. The address of that mail list can be found in the private forum article, "The Faculty mail list", where it is protected from the view of spammers.

    I sent a message, but it did not go out!
    The mailman program tries to filter out messages that are over a certain size, and that have too many people in the To: field.

    If a message seems to be held up, use the website to check for administrative requests (note you must be an administrator for that group):
    Click on Admin in the top menu bar,
    then select the appropriate mailing list from the list on the left,
    then select Tend to pending administrative requests. at the bottom. This should lead you to a URL like http://www.gigascale.org/workgroup/admindb/gsrc

    How do I see who is on a mailing list
    If you are a member of the workgroup whose list you are interested in, simply go to that group's "mail" page, http://www.gigascale.org/workgroup/listinfo, (or navigate there via the navigation menus), click on the name of the list in question, then click on "(Show member list)". (You must be a logged-in member of the group. This protects members from having their addresses harvested from our site by spammers.)

    You may also want to try the "overview" page for a workgroup, which will list all the members of that group. Try http://www.gigascale.org/workgroup/overview

    If you are the adminstrator of a group, you can use the mailman interface on the http://www.gigascale.org/workgroup/admin page to view, add and remove members of a mail list.

    There's the "mconnect" command that will show you aliases that are handled by the mail server, using the "expn" command from within mconnect:

    
    % mconnect eecs.berkeley.edu
    connecting to host eecs.berkeley.edu (169.229.60.28), port 25
    connection open
    220 EECS.Berkeley.EDU ESMTP Sendmail 8.12.10/8.9.3; Wed, 15 Oct 2003
    10:45:43 -0700 (PDT).  Unsolicited electronic mail advertisements
    strictly prohibited, subject to fine under CA law CBPC 17538.45.  This
    electronic mail service provider's equipment is located in the State of
    California.
    
    expn listname
    250 2.1.5 <listname@ic.EECS.Berkeley.EDU>
    quit
    
    % mconnect ic.eecs.berkeley.edu
    connecting to host ic.eecs.berkeley.edu (128.32.48.202), port 25
    connection open
    220 ic.EECS.Berkeley.EDU ESMTP Sendmail 8.12.9/8.12.9; Wed, 15 Oct 2003
    10:47:33 -0700 (PDT). Unsolicited electronic mail advertisements
    strictly prohibited, subject to fine under CA law CBPC 17538.45.  This
    electronic mail service provider's equipment is located in the State of
    California.
    
    expn listname
    250-2.1.5 <bar@EECS.Berkeley.EDU>
    250-2.1.5 <bif@EECS.Berkeley.EDU>
    
    If the alias is one of the mail lists that's handled by Mailman on gigascale, you'll get a result like this:
    %  mconnect gigascale.org
    connecting to host gigascale.org (128.32.32.199), port 25
    connection open
    220-InterScan Version 3.8-Build_1080 $Date: 01/31/2003 16:12:0037$: Ready
    220 gigascale.org ESMTP Sendmail 8.12.8p1/8.12.6; Wed, 15 Oct 2003
    10:50:27 -0700 (PDT)
    
    expn listname@gigascale.org
    250 2.1.5 <"|/home/mailman/mail/truncating_wrapper post  listname">
    
    In that case, you have to either check the membership of the workgroup on the Gigascale website (from the "overview" link for that workgroup), or use the shell command on gigascale:
    % /home/mailman/bin/list_members list_name
    
    ... that is, if the workgroup is on the Gigascale website.
    If it's on the Chess website, the command is:
    % /home/mailman_2/bin/list_members list_name
    
    and if it's on the Embedded website, the command is:
    % /home/mailman_3/bin/list_members list_name
    

    Why doesn't the GSRC Mailman system handle large attachments?
    Q: The mail handling package that manages email for the various GSRC workgroups is configured to bounce email messages that are larger than 40k. This makes it difficult to send attachments. Why?

    A: There are several reasons that Mailman is configured this way:

  • Sending Attachments to large groups of people on a mailing list is not always appropriate. Very large attachments can bog down mail servers. If a 10mb attachment is sent to a mailing list that has 10 people on one server, then that message will require 100Mb. This could cause problems
  • Attachments are very insecure. The recent 'ILOVEYOU' attack is an example of how attachments can be misused
  • Email that contains attachments are not indexed by the website search engine, whereas some files (such as .pdf files) that are placed as web pages will be indexed.
  • Mailman does not handle large messages gracefully - the web interface tends to be very slow or crash.
  • However, many people use attachments for collaboration, and we intend to make it easier to use attachments with the website.

    In the short term, the workaround is to place the file in the workgroup forum and ask people to go to the forum to upload the article. For more information, see How do add an article to a forum? and How do I attach or upload a file to an article?

    One workaround would be to have an alias that is not indexed for large attachments, or to have a way to mark large messages as approved mark them as not-to-be indexed.

    Another solution is to use CVS for collaboration and sharing of files.

    What mailing lists are available?
    Most of the mailing lists can be found at http://www.gigascale.org/people/lists/

    How do I fix 'Post by non-member to a members-only list'
    Most of the mailing lists are configured such that a the sender must be a member of the mailing list if their post is to go through without requiring approval from a group admin.
    The reason we have this restriction is to avoid spam.

    If a non member posts, then the group admin will get email requesting that the approve the post.

    If a member posts from more than one address, then the group admin may want to add the other addresses to the list of approved addresses. This will mean that the member gets only one copy of the email, but can post from multiple addresses.

    The way to do this is to be a group admin and go to the admin section of the group, then go Mail Admin on the left side for the appropriate mailing list, then go to Privacy and add the user's other address to

      Addresses of members accepted for posting to this list
      without implicit approval requirement. (See
      "Restrict ... to list
      members" for whether or not this is in
      addition to allowing
      posting by list members (Details)
    

    Hardware Software Infrastructure
    This FAQ section covers GSRC Hardware/Software Infrastructure.

    What can I do to improve the quality of the video captures?
    Powerpoint tips
    • As of 9/02, we are still using PowerPoint 97. Unfortunately, if you use PowerPoint 2000 or later to develop your presentation, then there is a chance that your animations may not work as you intend.
    • If you would like to preview your talk on a PowerPoint 97, send email to infrax at gigascale org.
    • The process of converting animated Powerpoint slides may result in different results. For example, slides that have separate text sections that are displayed as the presenter clicks the mouse are all combined into one .jpg file with all the animation completed.
    • We are working on using a different encoding system that will permit us to use a more recent version of PowerPoint.
    • Combining multiple talks into one long talk can help speed up sessions.
    • Make your slides available to capture crew the week before the presentation. Interrupting the crew while they are capturing the talk before your talk is only going to cause problems.
    • Once you submit your slides, don't submit updated versions, it makes it much more difficult to keep track of the proper version of the slides.
    • Powerpoint slides that contain non-standard fonts may not display properly on the presentation machine. When in doubt, get your files onto the presentation machine well in advance of your talk and check for font problems.
    • Transferring Powerpoint slides larger than a single floppy is tricky. Potential workarounds:
      • Emailing the presentation before the day of the talk
      • Use Winzip (http://www.winzip.com/) to created multiple floppy images.
      • Sometimes a single image on a slide will consume megabytes of space.
      • Usually the capture crew will have an IBM 600 zip drive available for use.
      • The capture crew also may have a Flash Memory Card available. Flash cards work best with Windows 2000, you may have problems with NT.
    Presentation tips
    • The less you move around while speaking, the more we can zoom in. Sitting behind a desk can sometimes help.
    • The lavolier microphone should be 8-12" below the chin.
    • Use a pointer, stay away from the screen. The screen is very bright, so if both the speaker and the screen are in the image, then the speaker will not be well lit.

    Tell me about telephone conferencing.
    Currently, we are using the Polycom Teleconferencing Telephones.

    For use in conferences, we have a Polycom soundstation premiere (2 mics plus wireless mic ~$1600)

    GSRC will purchase one Polycom Soundstation, (~$400 ea) for each for each site. The following locations have conference phones:

  • Professor Richard Newton's Office, 563 Cory Hall, (510) 642 2967
  • Professor Edward A. Lee's Ptolemy Lab, 337 Cory, UC Berkeley, (510)643-5892
  • CAD Lounge, 550 Cory, UC Berkeley,
  • Professor Andrew Kahng, UCLA (In Process)
  • Note that to use the Polycom Soundstation, you will need to have an analog phone line (aka a Plain Old Telephone Service or POTS line) A PBX line is not likely to work. It may be possible to use a PBX to analog converter, which usually runs about $150. We tried a converter locally, at it worked for a time, and then stopped working.

    To obtain a Polycom Soundstation: Send email to rnewton@ic.eecs.berkeley.edu and requesting a phone. Please include the following:

    1. The room number where the phone will be primarily used.
    2. The phone number that the phone will usually be connected to. By listing the phone numbers here, it will make it easier to set up conference calls
    3. The name and the address of the person the phone should be shipped to.

    Note that to use the Polycom Soundstation, you will need to have an analog phone line (aka a Plain Old Telephone Service or POTS line) A PBX line is not likely to work. It may be possible to use a PBX to analog converter, which usually runs about $150.

    Ideally, the Polycom Soundstation should be located near a PC with a projector for use with Teleconferencing Video software and other collaborative applications.

    For information about setting up conference calls, see See the GSRC Private Forum Question How Do I Schedule a telephone conference Call?

    How do I get access to commercial CAD Tools?
    Various companies have offered GSRC academic sites access to low cost or free licenses.

    The details are covered in the GSRC CAD Tools Forum.

    How do I use ICQ?
    ICQ (www.icq.com) is an "instant messaging" product that has two neat features of interest to collaborative workgroups:
    1. You can see when someone you are working with is currently active on their computer, and
    2. You can initiate a NetMeeting with them through ICQ.
    The rest of this FAQ explains how to install the ICQ client and set it up for these purposes. Our recommendation is to only use ICQ for contacts that you are working with very closely and who you don't mind interrupting you at any time. The setup notes here assume that you use it this way.

    Installing ICQ

    First, download ICQ from one of these locations

    Install it -- you will need to create a user ID, which should all be fairly explanatory. Enter the minimum amount of information possible.

    Open the ICQ client. From the ICQ menu/button at the bottom, select Security and Privacy. On the Security tab, select "My authorization is required." On the Ignore List tab, select "Accept messages only from users on my contact list."

    From the ICQ Menu/Button, select the Find/Add Users menu and then select the Find User - Add to List entry. use the dialog to locate the person you are working with and add them to your contact list. They will be sent a message requesting authorization.

    Contacting people

    When your ICQ client is running, you will see the people in your contact list, and a small icon indicating whether they are connected to the network and they are active on their computer. If they are, click on the user name to send a message. They will be interrupted from their work by ICQ beeping at them and displaying your message.

    (You see why you should keep your contact list very small. However, for keeping in close contact with the right people, this can be invaluable.)

    NetMeeting with people

    Assuming you have NetMeeting installed (see the NetMeeting FAQ), click on the name of the person you want to netmeet with. From the menu, select Internet Telephony/Games, and then select Microsoft Netmeeting. ICQ will start NetMeeting (if it isn't already running) and "call" the other person's NetMeeting.

    What camera do I use for desktop videoconferencing?
    If you do not already have a camera installed in your PC, we recommend that you purchase the Winnov Videum Conference Pro in the PCI bus version. These run about $400 retail. We use this successfully and without problems on Windows NT 4.0. (In fact, the Winnov card is what we use for recording presentations at the GSRC workshops.)

    For laptops, the only camera/card that we have used is the Winnov Videum Traveler. This is a PCMCIA card with a small camera. It works under Windows NT 4.0, with the following caveats:

    1. It does not work correctly with CardWizard.
    2. The video quality is quite poor.
    As it costs about $350, we cannot recommend the Videum Traveler unless you absolutely positively must have video-conferencing on your laptop. If you find a camera with PCMCIA card that has good picture quality and works under Windows NT 4.0, please let us know.

    For more information about cameras, see the GSRC Video group cameras page For more information about PCMCIA Video cards, see the GSRC Video group PCMCIA page

    How do I use NetMeeting?
    NetMeeting is Microsoft's free Internet video/audio conferencing program. If you have a camera installed into your computer (see this FAQ), you can use NetMeeting to conduct brief and/or impromptu meetings with other people from your desktop.

    Installing NetMeeting

    By default, NetMeeting is installed in Windows 2000. " To use NetMeeting on Windows 2000, click Start, point to Programs, point to Accessories, point to Communications, and click NetMeeting. Windows NetMeeting"

    Or, you may download NetMeeting from the Microsoft Web site: http://microsoft.com/windows/netmeeting/download/.

    Installation is straight-forward. To the question about logging into an ILS server on startup, answer No. (And in general, anytime Microsoft asks you if you want to be listed in an ILS directory, say No.)

    Setting up NetMeeting for video

    NetMeeting can only be used for point-to-point conferencing. (Some parts of NetMeeting, such as the shared whiteboard, and application sharing work with a multi-person meeting. The audio and video are only point-to-point -- see the multi-point video FAQ.)

    To enable video, start Netmeeting (it should be in the Program Files menu), and then select Tools then Options then select the Video tab. On this screen, select the checkboxes to automatically send and receive video. Set video quality to the high end of the scale, and verify that Videum Video Capture is set for the video camera.

    To check that your video is working with NetMeeting, click on the button with the triangle and two vertical bars drawn on it. If you see yourself in a little window, you're fine. If not, you have a problem.

    Starting a meeting

    There are two types of meetings:
    1. Impromptu point to point meetings
    2. Scheduled one to many meetings and sharing an application

    Impromptu point to point meetings

    From the Call menu, select New Call. Enter the name of the machine that the person you are calling works on. For example, edalap89.eecs.berkeley.edu. If you do not know the name of their machine, you will have to call them and ask! Or use ICQ.

    If you get the machine name right and they have NetMeeting running, you will get a window that says waiting for a response from machine. (If they are not running NetMeeting, you will get a message saying that they are unable to accept NetMeeting calls. In that case, call them and ask them to start it.) When the other person answers the call, you will see their video, and as long as network bandwidth is good, hear them.

    Note: if you are having problems with the audio, you may prefer to use a speakerphone than voice over IP. This is rather more reliable, but less cool.

    Scheduled Meetings

    NetMeeting can only share video and audio in a point-to-point manner, but it is possible to share applications from one presenter to multiple participants.

    We use a NetMeeting directory server currently running on vallejo.eecs.berkeley.edu as a meeting place.

    To use NetMeeting, send email to your participants, stating

    • The time of the meeting
    • The name of the directory server, which in this case is vallejo.eecs.berkeley.edu
    • The name of the machine that the host user will present from. For example:
      We will be having a NetMeeting/teleconference meeting on Month, Day, Time Time Zone

      The audio portion of the meeting will be accessible via
      XXX-XXX-XXX
      Participant Code: XXX XXX
      Host Code: XXX XXX

      We will use Microsoft NetMeeting to share applications. Our directory server is vallejo.eecs.berkeley.edu

      The hosted meeting will appear as Select a meeting name

      The password will be something short and easy to type

      NetMeeting is included in Windows 2000.

      To start NetMeeting, do Start -> Programs -> Accessories -> Communications ->NetMeeting.

      We will use vallejo.eecs.berkeley.edu as the NetMeeting Directory Server.

      NetMeeting can have multiple directory servers. To use vallejo as your directory server, follow these steps:

      1. Start Netmeeting, Select Tools -> Options -> Directory Settings
      2. Overwrite the current entry with vallejo.eecs.berkeley.edu.
      3. Select Logon to directory when Netmeeting starts.
      4. Close the windows and Netmeeting should immediately log into Vallejo.
      5. Then select the Find Someone icon from the main Netmeeting console.
      6. Select vallejo.eecs.berkeley.edu as the directory server if it is not default.
      7. Then they should an entry with a with red star, which means that that machine is hosting a meeting.
      8. Double click on the appropriate host and you will be automatically connected to the meeting..
      For details about NetMeeting, see http://www.gigascale.org/gsrc/faq/75.html

      Using the whiteboard

      While the video portion of NetMeeting only works between two people;that is, it is point-to-point, NetMeeting does include a whiteboard facility that can, be used between two or more people.

      To use the whiteboard, start up NetMeeting and select Tools->Whiteboard.

      Note that the whiteboard works best if all the NetMeeting clients are version 3.01. Anyone who is running Version 2.0 should upgrade to 3.01

      Multiple Netmeeting users

      The whiteboard can support multiple users, but regular net meeting cannot.

      In January, '02, Marvin Motley said:

      Netmeeting coupled with correct software(MS ILS/Directory Server/Conference Server) can be used to host multiple meeting participants. This setup takes minimal hardware but much configuration.

      The Sony system will be able to fit within that framework but to what extent is yet unknown. Sony is about to release v 3.0 for their PCS6000 so we should get some help there. Also the PCS6000 is easily manipulated via the OS so down the line... I have only hosted point to point using the Sony camera & netmeeting and it worked ok. Performance can be improved via configuration.

      I have hosted multiple meetings participants using Netmeeting but it is something we need to develop and it is not ready "Out of the box".

      For further information about NetMeeting, see the Video workgroup NetMeeting Page

      What about DSL?
      DSL

      Digital Subscriber Line (DSL)

      GSRC Faculty and Students have expressed interest in getting DSL. This FAQ is targeted at DSL users at UC Berkeley, but GSRC Members at other sites might find this information useful.

      Contents

    • Guides
    • Vendors
    • Current Status
    • UC Berkeley: How to get DSL
    • Troubleshooting
    • D-Link Systems
    • Setting up a WebRamp Router
    • DSL vs. Cable Modem
    • DSL Features we would like
    • Multiple IP addresses
    • DSL Modem/DHCP/NAT box all in one
    • Separate Ethernet Routers
    • Guides

    • Yahoo DSL Companies
    • Yahoo DSL
    • dslcenter - part of dnai
    • www.dslreports.com
    • Navas Cable Modem/DSL Tuning Guide - Good list of how tos and Alcatel compatible modems
    • Sharing a modem
    • www.2wire.com
    • Vendors

    • Alcatel
    • Covad
    • NorthPoint
    • Pac Bell DSL
    • Pac Bell XP DSL page
    • PacBell FasTrak
    • Current Status

      As of 2/00, basic DSL from Pac Bell uses PPPoE which provides a dynamic address. A few of us are having it billed through UCB Telecom

      Open Questions The next step will be to try to support Multiple IP addresses by purchasing the appropriate hardware that may need to support PPPoE

      UC Berkeley: How to get DSL

      1. If your advisor or supervisor agrees, you can have the DSL tacked on to your existing phone line and the DSL portion of your phone bill paid separately by campus directly, see the UCB Telecom DSL Page for details.
        Either follow the separate billing procedure, or, if you are paying for the bill yourself, go to http://public.pacbell.net/dedicated/dsl/dsl_basic.html and sign up.
        If your DSL bill is being paid directly, you will need the name of your grant administrator and a fund number for recharge.
        Most people will want to select Basic service, which consists of 1 dynamic IP address and costs $54/month. Enhanced service consists of 5 Static IP addresses and costs at least $84/month. In the past, Enhanced service was billable directly to UCB.
      2. PacBell will make an appointment to come by for installation and then show up to do the installation.
      3. During installation, you or the PacBell Tech will install a PPPoE program that acts like a dialer and gets your dynamic IP address, Gateway and the Primary and Secondary DNS servers.
      4. Pac Bell DSL instructions for XP
      5. Currently, PacBell Basic DSL gives you one dynamic IP address. Unfortunately, many campus services use the IP address to control access. When Basic DSL provided a static address, we made the follow adjustments in our campus environment:
        • To send email as you@eecs.berkeley.edu you will need to have your systems administrator configure sendmail so that it will relay email from your IP address.
          For example, on the mho cluster with Sendmail 8.9.3, we add the IP address of the DSL machine to /etc/mail/access:
          nn.nn.nn.nn     RELAY
          
          and then run makemap dbm /etc/mail/access < /etc/mail/access
        • If you are using hostname authentication for your website, then you will need to add the DSL hostname to the .htaccess file.
        • (On the mho cluster, this would be /usr/local/apache/etc/ohmCluster
          If you are using ssh and using RSA authentication to log in, then you will want to add your hostname to ~/.ssh/known_hosts to the line that corresponds to your key. Hostnames are separated with a comma, so the new line might look like
          adsl-nn-nn-nn-nn.dsl.snfc21.pacbell.net,ashwene.eecs.berkeley.edu 1024 37 123124213...
          
      6. The PacBell Quickstart guide says that you need to Download the DSL software, but this is really a customized version of Netscape.
        The PacBell FAQ at http://dialup.pacbell.net/help/faq/dialup_faq.html#customized says
        What is the difference between your customized Netscape software and the version I can download from Netscape's home page?
        Pacific Bell Internet Services' customized version of Netscape software includes a customized version of Netscape Navigator or Netscape Communicator, Dialer, TCP/IP stack, Registration Wizard, and supporting documentation. However, we do not include nor provide technical support for Cool Talk or other plug-ins. Our software gives you all you need to connect to the Internet using a modem.

        The Pacific Bell Internet Services customized version Netscape Navigator and Netscape Communicator has the same functions as the standard Netscape software. The customized features include easy access to the Pacific Bell Internet Services subscriber home page, with navigation aids.

        Do you have to use your customized versions of Netscape software with Pacific Bell Internet Services? Can I use Microsoft Internet Explorer?
        No, you do not need to use Netscape software with Pacific Bell Internet Services, but we do supply a complete software solution to register for our service and surf the net. So, yes, you can use Microsoft Internet Explorer with Pacific Bell Internet Services. However, you may need to configure the Microsoft Internet Explorer settings so that it will operate with Pacific Bell Internet Services. For Pacific Bell Internet Services settings see our network settings page.

        Please note that most users will need the Pacific Bell Internet Services software to register and we only provide customer support for our bundled software.

        If you are feeling brave, you can skip the download.
        If you are not feeling brave you can download via http://dialup.pacbell.net/download
        BTW - When downloading, Netscape showed that I was getting about 120K/sec.
      7. PacBell also throws in a DSL Dialup account via https://secure.pacbell.net/DSL.
      8. If you are dialed in using a PacBell modem or DSL connection, then you can access the PacBell.net DSL page at http://dialup.pacbell.net/dsl/. Note that dialup.pacbell.net is not accessible if you are connected to the net via an internet provider other than PacBell.
        dialup.pacbell.net includes links to the Enternet PPPoE binary.
      9. How do I removing the annoying DSL window? The PacBell DSL software EnterNET, installs with a default setting whereby an EnterNET "Profile" window opens at every reboot. Here is some info found online which provides the procedure for disabling this daily annoyance:

        Q: How can I get the EnterNet Folder to stop launching every time I reload Windows NT?

        A: Auto-launch is greyed out on NT because the feature is controlled from the PPPoE services. By default it loads at startup. If you wish to stop this behavior, from the Control Panel, click PPPoE services. Click both the icons and make sure that the Subservice loads checkbox is cleared. It doesn't always catch so verify by closing and re-opening the PPPoE services after changing them.

        There is also a bit more information on the same page at: http://support.efficient.com/KB/NTS/windows.html#autoNT

      Troubleshooting

      The Troubleshoting section is now at http://www.gigascale.org/gsrc/faq/94.html

      My Static DSL Configuration

      Alcatel 1000
      adsl-63-197-18-103.dsl.snfc21.pacbell.net
      router	63.197.18.254
      DNS 206.13.31.13
      

      D-Link Systems

      See the Wireless Ethernet FAQ - D-Link

      Setting up a WebRamp Router

      Web Ramp is no longer in business . . .

      We are using WebRamp 700s machines with static Pac Bell Basic DSL that was installed in 1999. These boxes will not work with dynamic Pac Bell Basic DSL that uses PPPoE. If you have PPPoE, see the PPPoE section below.

      What you will need

    • Network addresses, which can be obtained from ipconfig /all
    • Default Gateway
    • IP Address
    • DNS Server
    • Note that if you have already set up you PC for DHCP, then you will need to get this information from the information PacBell gave you during the initial install
    • The Webramp box
    • The DSP box
    • Your PC
      1. Plug in the webramp, DSL box and a PC. The webramp and the DSL are connected on the WAN (Wide Area Network) side of the webramp. The webramp and the PC are connected on the LAN side.
      2. Boot the PC, bring up Netscape, type in 192.168.1.251
      3. Login with
        Username admin
        Password password
      4. Goto General->Networking
      5. Standard Addressing Mode:
        SelectNAT Enabled
      6. Lan Settings:
        WebRamp 700s Web Address 10.0.0.1
        LAN Subnet Mask should remain as is 255.255.255.0
      7. Wan Settings:
        WAN Router Address should be set to the address of the DSL Router Under NT, you can find this by typing ipconfig /all and looking at the Default Gateway setting. For me, the value was 63.197.18.254
      8. NAT Public Address should be set to the address of your machine. This is that address that PacBell gave you for your machine. Under ipconfig /all it can be found under IP Address 63.197.18.103
      9. DNS Server should be set to the first value from DNS Servers from the ipconfig /all output. For me, this was 206.13.31.12
      10. Press Update, but don't restart yet
      11. Select the time tab, and set the time
      12. Select the password tab, and change the password
      13. On the Left side select DHCP. Under Global Options, make the following changes
      14. Client Default Gateway: 10.0.0.1
      15. We will set up dynamic DHCP for addresses 10.0.0.3 through 10.0.254. Note that the webramp will only do NAT for 5 machines. To handle more machines it would be necessary to buy a software upgrade.
      16. Select the current Dynamic range and delete it
      17. Add a new range from 10.0.0.3 to 10.0.0.254, and press update
      18. On the left side, select Advanced, then select Intranet, and select "WebRamp 700s's WAN link is connected directly to the Internet router"
      19. Restart the router, it will take almost a minute to restart
      20. Set up the PC to use DHCP
        Under NT, select Start->Settings->Control Panels->Networking->Protocols->TCP/IP Protocol->Properties->Obtain an IP Address from a DHCP server. You may need to reboot, but you could try running
        ipconfig /release
        ipconfig /renew
        
        From a DOS shell or bash window
      21. After the webramp has come back up, try using the browser to connect to 10.0.0.1
      22. Then try connecting to http://gigascale.org
      Troubleshooting Webramp setups
    • Try pinging the webramp router with ping 10.0.0.1
    • On the webramp, select Tools -> Diagnostics, and try pinging remote hosts like brahe (128.32.171.111)
    • Power cycle the DSL hub and then the webramp
    • Power cycle your PC
    • Use ipconfig /all to look at your network settings
    • To get a new DHCP connection, use
      ipconfig /release
      ipconfig /renew
      
    • Use route print to look at your routes
    • Webramp power connection

      The Webramp takes the following wall wart:
      PowerDec +5v 1500ma tip +, ring -
      

      DSL vs. Cable Modem

      The @home pricing page says that DSL is more expensive. Other than that, a search of DSL on the @home weba site returns nothing.

      Jul 1, 1999 zdnet article about Cable modem problems

      The ADSL Forum says:

      How does ADSL compare to cable modems?

      ADSL provides a dedicated service over a single telephone line; cable modems offer a dedicated service over a shared media. While cable modems have greater downstream bandwidth capabilities (up to 30 Mbps), that bandwidth is shared among all users on a line, and will therefore vary, perhaps dramatically, as more users in a neighborhood get online at the same time. Cable modem upstream traffic will in many cases be slower than ADSL, either because the particular cable modem is inherently slower, or becasue of rate reductions caused by contention for upstream bandwidth slots. The big difference between ADSL and cable modems, however, is the number of lines available to each. There are no more than 12 million homes passed today that can support two-way cable modem transmissions, and while the figure also grows steadily, it will not catch up with telephone lines for many years. Additionally, many of the older cable networks are not capable of offering a return channel; consequently, such networks will need significant upgrading before they can offer high bandwidth services

      DSL Features we would like

    • Ease of maintenance -

      Some have the time to hack around with Linux, some people do not

      A web based configuration tool would help.

    • Multiple IP addresses: Either purchase multiple addresses, or use Basic DSL for one address will be sufficient for many users. However, some users will want multiple addresses to support more than one PC, or to support a network printer.

      Network Address Translation: (NAT)

      Hides your LAN behind one Internet IP address by translating IP addresses on a LAN to a single static or dynamically assigned IP address on a WAN.

      Dynamic Host Configuration Protocol: (DHCP) A service that allows clients on a LAN to request configuration information, such as an IP address, from the router on a dynamic basis.

      Flowpoint NAT FAQ has some useful information about NAT
    • Ideally, we would like to have DSL billed through the University. At the minimum, we would like to have Separate Billing so that DSL users can have DSL added to their current phone line and easily get reimbursed from for their DSL costs.
    • Firewall - Some users would like to have a firewall at home. Note that UC Berkeley does not currently have much of a firewall, though some filtering is done.
    • DHCP support for home laptops
    • A common way to do this is to bring up a Linux box with DHCP, NAT and Firewall facilities. (ADSL HowTo for Linux Systems)
    • This is a good thing if you have an old PC, and you would like to mess with Linux.
    • However, old PCs tend to fail, and tend to need patching
    • Once the PC is broken into, a lot more damage can be done
    • Not everyone has the time or inclination to be a Linux sysadmin.
    • The right solution is to get a piece of hardware that one can just set up and forget.
    • There are several possibilities
    • DSL Modem/DHCP/NAT box all in one <---ethernet--->
    • DSL Modem <--->ethernet<---> Separate Ethernet Routers <---ethernet-->
    • Below we discuss some of the details

      Multiple IP addresses

      PacBell

      Pacbell Basic DSL Internet Access is
      $49: $39(phone) + $10 (ISP) for 1 static IP address at 384 Kbps-1.5 Mbps down/128 Kbps up Currently this address is statically assigned, though it may be dynamic in the future.

      Pacbell Enhanced DSL Internet Access is
      $79: $39(phone) + $40 (ISP) for 5 static IP addresses at 384 Kbps-1.5 Mbps down/128 Kbps up

      $199: $129(phone) + $70 (ISP) for 5 static IP addresses at 384 Kbps-1.5 Mbps down/128 Kbps up

      DNAI/Covad

      DNAI TeleSurfer -
      $67: 1 static IP address 384 Kbps down/128 kbps up
      $90: 1 static IP address 768 Kbps down/384 kbps up
      Speedstream 5250 DSL bridge $224

      DNAI Telespeed
      $122: $10(5 ip) + 112 5 static IP address 192 Kbps down/192 kbps up
      $158: $10(5 ip) + 148 5 static IP address 384 Kbps down/384 kbps up
      Flowpoint 2200 SDSL Router (DHCP and NAT) $395

      DSL Modem/DHCP/NAT box all in one

      PPPoE

      If you have Basic DSL from PacBell, then you probably have PPPoE. If you want to share your Basic DSL connection, then you will need a box that understands PPPoE. PPPoE is Point to Point Protocol over Ethernet.

      It seems like DSL vendors are moving away from a pure DHCP based system to a PPPoE System.

      PPPoE is causing a fair amount of concern among DSL users on the comp.dcom.xdsl newsgroup.

    • http://www.carricksolutions.com/pppoe.htm
    • http://www.rback.com/solutions/pppoe/wp_pppoe_comparison.asp
    • Nextlan and ZyXEL routers might handle PPPoE

      SonicWall

    • SonicWALL SOHO2/10-user Internet Security -Can be upgraded to VPN- $412
    • Linksys

    • Setting up Linksys with PacBell
    • BEFSR41 - EtherFast 4-Port Cable/DSL Router $117 at CDW
    • LinkSys BEFW11S4:
      LinkSys BEFW11S4 access point/router with 802.11b wireless support. This box has four wired ports, wireless 11Mbps with WEP encryption supported, and a built-in DHCP server. It is configured through a web interface and has several firewall options that allow you to redirect incoming connections, block outgoing connections, etc. It also supports PPPoE.

      $250 at CDW

    • Alcatel

      PacBel DSL Faq seems to require a Alcatel 1000 ADSL Modem
      Equipment options for DSL are becoming more and more available, customers may use equipment from any providers, as long as the equipment meets the interface and other technical specifications of the FCC Pacific Bell's FasTrak DSL offering.
      Pacific Bell provides an Alcatel 1000 card, which is only a modem there is no reduction in price if you do not use it.
    • Alcatel 1000
    • Alcatel modems
    • Alcatel Speed touch office
      http://Cable-DSL.home.att.net/ says that the Flowpoint SmartSwitch Router 250 (see below) is the same as the Alcatel Speed Touch Office. The pictures look the same.
    • Efficient Networks

    • Efficient Networks SOHO DSL boxes
    • Separate Ethernet Routers

      Cayman

      www.cayman.com

      Netgear

    • www.netgear.com
    • Netgear RT314 is about $117 and has been used by admins in CS with PacBell:
      We purchased several of the Netgear RT314's for a couple of the CS department administrators. I haven't heard of any complaints. I checked one out when they were sitting in my office; they look pretty nice. And they have a 4 port 10/100 switch built in. All for ~$100.

      http://www.netgear.com/product_view.asp?xrp=4&yrp=12&zrp=55

    • Netgear Routers

      RT311 ENET TO ENET ROUTER FOR USE WITH ADSL AND CABLE MODEMS $355

    • Netopia

    • The Netopia 9100 has a built in 8 port 10 base-T hub. Discontinued 12/00
    • Red Creek

    • Red Creek makes the Personal Ravlin which has IPSec
    • CDW has it for $699 (Discontinued 11/00)
    • Single User?
    • WebRamp

      (7/01) It looks like the WebRamp 700 was actually made by SonicWall and has been discontinued.
    • 7/99 Information week article about the 700s
    • What hardware is available for loan to GSRC faculty?
      The GSRC has various equipment. This equiment may be borrowed by GSRC Faculty:

      Items available include:

      • 3 LCD projectors Sharp NV3, NV5 and NV6
      • 1 Nikon CoolPix 990 digital camera
      • 1 Polycom Soundstation

      What about Wireless Ethernet?
      The UC Berkeley EECS Department is using Lucent Wavelan II Gold PCMCIA cards to provide 11Mb wireless ethernet in Cory Hall.

      Links

    • ISDG Wireless ethernet policy
    • IDSG Wireless ethernet procedure page
    • Orinoco Website
    • Wavelan Orinoco RG-1000 Residential Gateway (~$379)
    • The basic idea is that for $360, you get the use of a PCMCIA ethernet card. The connection is encrypted, the department must type in a password to enable your card.

      Wireless Procedure

      1. Get approval for the cost and get an account fund number
      2. Fill out the form at http://www.eecs.berkeley.edu/idsg/forms/activate-wireless.html
        Affiliation
        The group name you are associated with, for example the login of your PI (eal etc.)
        Technical Point of contact
        The email address of your technical person, either yourself or CUSG if you have a CUSG contract.
        System Serial Number (or UC Property Number)
        The UC Property number is on a sticker on the bottom of the laptop
        System Operating System Type(s) (include version)
        Windows 2000
        SystemName
        Use your NetBIOS machine name, which is returned by running ipconfig /all at a bash prompt and reading the Host Name value, for example:
        bash-2.05$ ipconfig /all
        
        Windows 2000 IP Configuration
        
                Host Name . . . . . . . . . . . . : EDALAP03
                Primary DNS Suffix  . . . . . . . :
                Node Type . . . . . . . . . . . . : Peer-Peer
                IP Routing Enabled. . . . . . . . : No
                WINS Proxy Enabled. . . . . . . . : No
                DNS Suffix Search List. . . . . . : eecs.berkeley.edu
        
        Use Host Name, which in this case is EDALAP03
      3. Follow the instructions at \ IDSG Wireless ethernet procedure page which includes the steps below:
      4. Download the drivers by copying
        \\ntsww.cs.berkeley.edu\sww\wireless\Win2k to a temporary directory on your local machine and run setup. The reason to copy the driver directory is that IDSG will use that directory to update the firmware.
        OR
        Download the software from the manufacturer:
        1. Goto http://www.orinocowireless.com and select Software
        2. Under Select a Product, select Client Products
        3. Under Select an Operating System, select Win2000
        4. Hit the Search button
        5. Download the Client software. As of 8/01, this was called Client Software - ORiNOCO Rel 7.2 for Windows 2000 - Fall 2001 release. - 9/8/2001 and was 8.2Mb in size
        6. Optionally, you can update the firmware on the card by downloading the Firmware Update
      5. Goto 395/399 Cory or 321/331 Soda and get the card and have it enabled.
      6. If necessary, change your outgoing email server to smtp.eecs.berkeley.edu. This address will work for any machine on a network inside the eecs (and maybe cs) domain(s).

      Home wireless access points

      D-Link 713P

      In 9/01, we purchased several D-Link di713P wireless routers for about $220 each. It has NAT, DHCP, a firewall, a print server (Windows only), and 3 10/100baseT ports. The "P" is important. There's also a "DI-713" which is older and not as cool. (See also The GSRC DSL FAQ )

      To configure the D-Link 713P, most users will go to http://192.168.0.1/
      The password should be admin.
      Some users may have their networks configured differently, so an alternative is http://10.0.0.1

    • Most users will have Dynamic Addresses, so they will be using "PPP over Ethernet"
    • Setup -> Renew IP Forever should be checked
    • If you have a printer on your local network, then you may want to assign the printer a fixed IP address so that you can easily connect to the printer at the same address.
      To assign a fixed IP address, do DHCP -> Fixed Mapping
      • Enable MAC Address Control
      • Selection Connection Control
      • Add the MAC address of your printer and select the C Column
    • To set up wireless, go to the Wireless page and
      Network ID(SSID)
      A short name, for example MYNET
      Channel
      The default is channel 1, you may want to pick another channel.
      Security
      Enable IEEE 128 bit Shared Key security WEP KeyEnableIDSetting
      WEP Key
      A hexadecimal key of your choosing.
    • Then reboot the router.
    • Under XP, you may need to enter the same WEP key into your laptop with Start -> Network Connections -> Wireless Network Connections -> Properties -> Wireless Networks. Then add the Network ID (SSID) you chose above.

      Data Encryption (WEP enabled)
      checked
      Network Authenticatoin (Shared mode)
      checked
      Network Key
      Enter the WEP Key from above
      Key format
      Hexadecimal digits
      Key length
      104 bits (26 digits) - I don't know why this is 104 bits instead of 128 bits as above?
      Key index (advanced)
      0
      The key is provided for me automatically
      not checked
      Then hit OK

      3Com

      Steve Neuendorffer writes:

      I did a little bit more research. It looks like there is a consortium, WECA, that has implemented some 802.11b compatibility testing: http://www.wi-fi.org/

      There is a list of certified manufacturers on the website. The wireless access points in the $200 range don't appear to be certified, but at around $250 (on buy.com) you can get the 3Com access point which is certified compatible.

      For about $300, you can jump up to a broadband router/access point from 3Com that also does NAT, VPN, DHCP and firewall filtering and has a small ethernet hub in it: you can connect a broadband modem to the wired devices and wireless without any extra equipment. That's what I went for... I'll keep you guys informed of how it works.

      Debugging Wireless Connectivity Issues

      If you are having problems with connectivity from a specific location in Cory or Soda Hall, then send email to networks at eecs and dopsysadmin at eecs
      In your email, be sure to state which wireless access point base station you are connecting to and the mac address of the wireless card you are using.

      Using Windows XP to determine the access point you are connecting to

      You can provide the MAC address of your local Access Point by going to the Control Panel and selecting High Rate Wireless LAN Settings. Copy the Associated MAC address from the State information line.

      Using the ORiNOCO Client Manager to determine the Access Point you are connecting to

      The ORiNOCO Client Manager has a couple of features that will help characterize wireless connectivity issues. To start the Client Manger, either right click on the bar graph in the right side of the start bar, or run Start-Programs-ORiNOCO-Client Manager.
      Link Test-Test partner
      In Cory Hall, the wireless access points are name corywirelessFloorNumberCompassDirection
      so corewireless3sw is located on the 3rd floor in the south west corner of the building.
      Link Test-Test Results
      Usually the SNR should be at least 10dB
      The number of packets received and sent and the speed (11, 5.5, 2, 1)will give you an idea of the quality of your connection.
      Link Test-Test History
      Changing the Type of display to Signal/Noise on a one minute time scale will show connectivity drops as broken lines. This also give you feedback about signal strength as you move around.
      Things to try:
    • See if anyone else is having similar problems in the room.
    • Run the Client Mangager and try adjusting where and how you sit to maximize the signal.
    • Note things like packet speed: After running for a minute, what is the percentage distribution of packets between 11Mb, 5.5Mb, 2.2Mb and 1 Mb?
    • What access point are you connecting with? If you move to a different location in the building, is the connectivity better?
    • Try to get a different channel by noting the channel number, going to Advanced->Card Diagnostics->Test Card Now-> Cancel and then seeing if the channel changed.
    • See what other access points are in your vicinity by going to Advanced->Site Monitor
    • Follow the Download the software from the manufacturer instructions above and install new client software and update the firmware. It appears that it is not necessary to uninstall the drivers or client manager.
    • Mac wireless

      The Orinoco cards have MacOS8 and 9 drivers.

      The Dlink 713P will work with PC and Mac, though the printer software does not work under Mac, see: http://www.dlink.com/products/broadband/di713p/

      The Apple AirPort card ($99) should work with other base stations. The D-link support web page seems to indicate that the Apple AirPort card will work with at least one of their products, see http://support.dlink.com/faq/view.asp?prod_id=435

      The Apple AirPort Base is $279

      Orinoco Gold card under MacOs 8.6

      The Orinoco website has drivers for MacOS 8 and 9.

      However, when we clicked on the installer, we got CfgOpenTpt--OTCfgEncrypt could not be found

      The Descent game page says:

      "cfgopentpt" error: Some users have been reporting that the demo will not run unless you have the "Network Setup Extension" installed on your computer. This extension seems to be part of MacOS 8.5 or higher, so if you have 8.1 or lower, you might want to wait on downloading the demo until a solution is found. We are currently investigating this issue and will post more results here later.
      iBook-AirPort-MacFixIt has some notes about CfgOpenTpt

      Installing and Configuring Wireless Networking for Mac OS says that the drivers for the Orinoco card are included in Mac OS 9.0.4.

      The TSW Wireless page is a good source of info.

      ftp://ftp.orinocowireless.com/pub/software/ORiNOCO/PC_Card/MacOS/ includes an older driver (MacOS_63.bin) that does at least installs under MacOS8.6 and runs without encryption.

      Troubleshooting DSL
      The steps below are directed towards DSL users, but they could be used by anyone having network connectivity issues.

      Resources

      Tools

      Connectivity problems

      If you can't reach a host at all, follow the steps below.
      1. Take a deep breath
      2. Use ping
        Under Windows2k, try ping -t www.gigascale.org Other hosts to try dialup.pacbell.net, your DNS server (use ipconfig /all)
        If you are losing more that 3% of the packets, you are in trouble. Try to narrow down where the delays are by pinging different hosts
      3. Use traceroute
        Under Windows2k, traceroute is tracert Take a look at where your packets are going.
      4. Check for DNS problems by using IP addresses instead of domain names http://dialup.pacbell.net/dsl/dsl_tcpip.html suggests
                                     Primary      Secondary
        Los Angeles/Orange County    206.13.29.12 206.13.30.12
        San Diego                    206.13.30.12 206.13.29.12
        San Francisco/Bay Area       206.13.28.12 206.13.31.12
        Sacramento/Fresno/Stockton   206.13.31.12 206.13.28.12
        
        You might also try changing the last digit to .13 or trying one of the UCB EECS Departmental DNS servers listed in http://www.cs.berkeley.edu/idsg/network/ Three addresses to try are: 128.32.171.23, 128.32.239.23,128.32.156.24
      5. Try using a visual traceroute package like one of the ones listed above. See if you can determine what router is dropping your packets.
      6. Check the net traffic reports for bad traffic http://www.internettrafficreport.com or http://www.internetpulse.net/
        Look at the traceroute output for routers listed on the traffic report. For example, traceroute shows that my traffic goes through acr1-loopback.SanFranciscosfd.cw.net [206.24.210.16] core1.SanFrancisco.cw.net is worth a look.
        Since traffic is going through Cable & Western, looking at the Cable & Western Traffic site (http://sla.cw.net/) and trying ping and traceroute might indicate if the problem is pacbell, sprint, cw, or uc.
      7. Follow the Pac Bell DSL Slowness info notes at http://dialup.pacbell.net/dsl/trouble.html#4:

        I am experiencing slowness on my DSL Internet service.

        Power cycle Alcatel unit:

        1. Unplug your Alcatel unit - do not turn unit off.
        2. Shut down your computer.
        3. After 60 seconds reboot your computer.
        4. Turn the Alcatel unit off and plug the power source back in.
        5. Turn the Alcatel unit on and observe the lights.
        6. Wait 2 minutes for the unit to sync (the Power/Sync indicator light will flash red during the wait). 
        7. If the Power/Sync light does not turn solid green after 5 minutes repeat steps 4, 5, and 6 (up to three times).

        If this does not solve the problem, try the following:

        • Check our System Status page. Any reported Pacific Bell Internet Services Network issue that could affect your performance will be posted here.
        • Go to http://www.internettrafficreport.com. This site will show you if any of the major Internet backbone providers are experiencing packet loss. A serious outage on the part of one provider can affect the performance of others, as traffic is rerouted through the system.
        • Clear the Netscape disk and RAM cache, then restart Netscape. Click here for step-by step instructions.
        • Find and delete all netscape.hst files. The files may have become too large or obsolete. These files are located on your hard drive, in your Netscape directory.

      Bandwidth Problems

      If having problems with bandwidth,
    • http://promos.mcafee.com/speedometer
    • http://www.pcpitstop.com/internet/bandwidth.asp iCheckConnection from pcpitstop.com will test connectivity, including downloading ftp files and reporting back the timing. Note that iCheckConnection reports kilobits per second as opposed to kilobytes per second.
      Below is a simple test
      $ ./iCheckConnection ftp://ic.eecs.berkeley.edu/pub/Octtools/OCT5.1.document.ta
      r.Z
      PC Pitstop Internet Connection Checker 1.03
      Copyright (c) 2000-2 PC Pitstop LLC. All rights reserved.
      DHCP Server address not found
      DNS is working
      Ping ic.eecs.berkeley.edu (128.32.171.49) 32 bytes      30 20 20 30 20
      ftp://ic.eecs.berkeley.edu/pub/Octtools/OCT5.1.document.tar.Z returned 1530391 b
      ytes in 10 seconds (1194 Kb/s)
      
      1194 Kb/s is 149.5 KB/sec. (1194/8 == 149.5)
    • Take a look at some of the information at http://Cable-DSL.home.att.net/#ConnectionSpeed.
    • If you are running under Windows 2000, then you may want to experiment with the Windows Performance Monitor
      1. Start up the Windows Performance Monitor: Start -> Settings -> Control Panel -> Administrative Tools -> Performance
      2. Move the mouse over the graph, right click and select Add Counters
      3. Under Performance Object, select Network Interface
      4. Select All Counters
      5. Be sure that the appropriate interface is selected, in my case this was FE575 Ethernet Interface
      6. Then hit Add and then Close
      7. You may need to rescale the graph by right clicking, selecting Properties -> Graph and changing the Vertical Scale maximum to 10.
      8. The most interesting value is the Bytes Received/Sec value. Look at this value as you try the other tests.
    • Try downloading a 10Mb file by right clicking on the link below and selecting Save As http://www.gigascale.org/pubs/downloads/10meg
    • The speeds below are reported in KiloBytes per second.
    • The PacBell Basic DSL Page says that
      Speed Available* (downstream/upstream) Up to 1.5 Mbps/128 Kbps

      *Service not available in all areas due to factors associated with DSL technology such as line conditions or distance. Actual speeds will vary. Access speed is between customer's location and the DSL Internet-equipped Central Office or Gateway.

      Location Cable ID Machine SCP Download Rate (in KiloBytes) SCP Upload Rate (in KiloBytes) Internet pitstop Download Rate (IE only, results in kilobites/sec) Internet pitstop Upload Rate (IE only, results in kilobits/sec)
      DOP Center ? Sun Ultra 60 using wget 5.05MB/sec 5.05MB/sec
      400A Cory 10Mb Half Duplex ethernet 405-400A-006 Sony laptop with 3Com card 700-900KB/sec ?
      400A Cory 100Mb Full Duplex 405-400A-003 Sony laptop with 3Com Card 1000KB-1700KB ?
      336 Cory 10Mb Half Duplex 315-337-031 IBM 600X with 3Com Card ? ? 1200kb-2000kb 540kb-650kb
      Cable Modem El Cerrito P2 250 400KB-600KB
      Pac Bell DSL Albany IBM 600X with Bay Networks wireless card 145KB-156KB 12.7KB 1192Kb = 149KB 108Kb = 13.5KB
      Sonic Primary DSL (no minimum speed guarantee) Cotati IBM T30 with 10baseT (12/02) 19.2KB 12.5KB 162Kb = 20.25KB 110Kb = 13.75KB
      Pac Bell Dialup 49.2kbs Cotati IBM T30 with built in modem (12/02) 4-5KB 2.7KB? 40Kb = 5KB 62Kb = 7.75KB?
      If you have a Sony laptop, you may want to see: How do I configure the ethernet card?

    GSRC Workgroup Calendars
    Information about create, configuring and using GSRC workgroup calendars.

    How can I create and configure a calendar for my workgroup?
    You can create a calendar for a workgroup if you have Administrator privileges for that workgroup.

    From your workgroup's admin page, click on the "Configure group" link. On the resulting page click the "Calendar" checkbox in the "Workspace options" section of the form, and the click the "Change group configuration" button at the bottom to submit your configuration change. The dropdown menu for your workgroup should now include an item called "calendar", from which you can access your calendar.

    Access to the calendar can be controlled in basically the same way as access to the other sections of your workgroup (faq, forum, etc.), i.e., from the "Access control" subpage of your group's "admin" page.

    From your workgroup's dropdown menu, select "admin", and from the resulting page click on "Access control". This takes you to a page of forms for controlling access to the sections of your workgroup. You should see a section labelled, "calendar".

    The access settings have the following meanings:

    Read
    User can view calendar and click on events for details.
    Write
    User can add events to the calendar, and edit or delete only events orginated by him/herself.
    Modify
    User can edit or delete any event on the calendar, regardless of its originator.
    Annotate and Execute
    have no effect.

    Only logged-in website users can create or modify events, regardless of the settings for "world". This is because events must have a user's identity associated with them.

    How to use the calendar.
    You can view any workgroup calendar for which you have been giving read permission by the workgroup's administrator. (This includes the general GSRC calendar, which is just a workgroup calendar for the "gsrc" workgroup.)

    You can click on an event title on the calendar to view whatever detailed information has been provided by the originator of that event.

    To edit, delete or add an event for a given day, click on the date number for that day in the calendar display page. If the date numbers are not clickable, you do not have write or modify permission for that calendar.

    Clicking the date number takes you to a page containing a list of the events for that day and a form to enter a new event. If you have modify permission in the calendar, each of the events listed will have an "Edit" and "Delete" link that you can click for that event. If you only have write (and not modify) permission in the calendar, only those events that were originated by yourself will be editable and deletable. The others will show a description of the originator instead of the "Edit" and "Delete" links.

    Clicking "Delete" will take you to a page asking for confirmation before deleting the event.

    Clicking "Edit" will take you to a form for changing your event. The form does some validity checking -- the end time cannot precede the start time, and you cannot enter both detail text and a detail URL. If your event has "detail text", a user who clicks on that event will see that text in a popup window. If your event has a "detail URL", then the user who clicks on that event will instead be taken to a web page with that URL for further information about the event.

    Intellectual Property Issues for GSRC Participants
    This section discusses intellectual property issues for GSRC participants.

    How do I release software?
    NOTE: This is a provisional draft response
    If you have questions, send email to Christopher Brooks

    Unlike papers and other publications, software does not have a prepublication requirement.

    The GSRC contract with MARCO states:

    "5.2 Software Submissions" MARCO shall be informed of new software or updated versios within thirty days of creation [or] update via the Software Submission Form found at http://fcrp.src.org

    The GSRC software release procedure is documented on the Softdevel workgroup copyrights page.