Go to ALU Welcome Page

Current Design

• tables: Tables versions of the site's contents.
• simple: Non-tables versions of the site's contents.
• frames: Frames versions of the site's contents.
• admin: Adminstrative information, including this page.
• img: .gif and .xbm files.

The same directory that holds these five directories also holds several more directories that have content that originated elsewhere. The material in them does not use the navigation scheme used by the "site proper," and they are not listed on the site map:

• HyperSpec: Harlequin's ANSI CL HyperSpec. (Not available at Elwood mirror.)
• mop: The MetaObject Protcol "hyperspec".
• luv: Information from previous Lisp User and Vendor conferences.
• humor: Brad Miller's collection of Lisp humor.

This current site is completely static -- it does not use any of the dynamic output-generation facilities of CL-HTTP, and is thus suitable for use with any webserver. A gnu compressed tar file of the entire site is available for download from the file alu.tgz in this (the admin) directory.

If you are reading this file from the www.alu.org or www.lisp.org domain, then the pages are being served by CL-HTTP running on Franz Allegro Common Lisp on a PC running Linux. The machine is at the MIT AI laboratory using space and an internet port donated by the laboratory. It is administered completely remotely over the internet, using the adminstrative facilities of CL-HTTP, rlogin, and a "telnet" process within the Lisp running CL-HTTP.

Further Work

• What is Lisp? Volunteer needed
• Learning Lisp Editor: Raja Sooriamurthi
• Applications Editor: Michael Tuchman
• Tools Editor: Bruce Tobin
• Community Volunteer needed
• Reference Volunteer needed
• Systems Editor: Don Naumann

• site in general webmaster needed

We need volunteers to do each of these. This is a chance to learn and work with such leading edge tools as CL-HTTP, as well as with some outstanding members of the Lisp community. For more information please contact the ALU board at alu-board@ai.sri.com.

CL-HTTP and the Site Overall

1. Search! Presumably this would use W4. The current dummy search page is only a suggested starting point for the user interface.

2. The dynamic output generation capabilities of CL-HTTP should be used to provide the three presentation formats (tables, no-tables, frames) from a single source. See proposal.

3. Discussion pages. Each page should have a mechanism for visitors to leave their comments, and to view the threaded comments which have been left. The CL-HTTP discussion mechansim should be used. (Example is National Performance Review Open Meeting.) A chat facility is also possible, or archived mail (see below).

   In particular, the current pages have a lot of editorial opinion which should be moved to such comment threads. See Learning Lisp and Systems.

4. Window dressing! Submit your graphics/animation! Given CL-HTTP's access to HTTP information from the client and the ability to generate mobile code, some sort of animated "your name in lights" applet based on client information would be cute.

5. The site could be used to host the common-lisp mailing list and/or comp.lang.lisp and comp.lang.clos newsgroups in a multiple-view archive hypertext archive style.

6. It ought to be easy to write a robot which periodically checks the links on the static pages and checks to see if they are still valid (404 - file not found). If not, the link would be changed to go to the dead links page, which would report the old link and which ALU page refered to it.

   we could also maintain a ``responsible-party'' database for each linked site, which creates an automatically maintained? listing of linked sites, and allows people to send mail which registers them as the responsible party for that site. The dead-link robot would automatically mail these people and ask for a new link.

   The idea is just to make it easy for our webmaster to be good citizen in term of keeping things up to date.

7. Rather than trying to maintain lists of certain resources by hand, we should use W4 to do it automatically. This is true for the listings of:
   • Commercial sites
   • Available software tools
   • Research projects
   • Research papers
   • Consultants
   • Jobs?
   For each of these, a separate W4 robot would be programmed to locate candidates for the given category. Some clever programming (dare I use the word 'AI'?) would:
   • Project these into single entries (i.e. reduce multiple pages found on ``the same thing'' to a smaller number of unique entires).
   • Locate the ``top'' or ``home'' page for each entry.
   • Determine if the page is already listed.
   • Produce the appropriate description, keywords for our site search engines, etc.
   • In the some cases (such as when gathering information about available source code or available commercial applications), it would also be desirable for the robot to attempt to determine dependencies on other systems. This can be done from textual descriptions, or in the case of source code, by examining the code itself. The information would be displayed in summaries of the application, and available to queries.
   This ``clever programming'' is a good opportunity for graduate student research!

8. Bug? In the tables-based hierarchy, returning back to a previously visited page does not come back to the same place in the page in at least one browser (Netscape 2.0 Sun). It is not clear if this is browser or html coding bug. If you have any information about this (even just to report that your browser does NOT have this problem), please contact us.

9. Automatically generated pages should have the date of last update inserted at the bottom, along with a mailto link for the editor of the corresponding area.

What is Lisp?

No planned changes (other than the discussion threads throughout the site.

Learning Lisp

1. We need someone to contact the publishers/authors and obtain the publisher's abstract for each book, which would be listed instead of the existing editory comments. The comments would then be moved to a discussion thread.

2. The Lisp family section is begging for a graphical tree which shows the various dialects, and approximate date of introduction. Clicking on the nodes of the tree should bring up the description of the dialect.

3. A gentle "common examples" page, written by an educator or author. It has been suggested to me that people who know one or more programming languages like to see examples of doing common things before learning about a new language. The "hello world" and factorial examples that I listed in Learning Lisp are as much as I dare. Things to keep in mind are:
   1. We don't want to confuse beginners with multiple ways of doing things.
   2. We don't want to put off people who prefer a particular programming style by only demonstrating things in one style that might be different than the one a particular reader favors. See the ``Object Oriented and Procedural Lisp'' page.

   This has to be done very carefully. My feeling is that if we can't get something very good, we should not do this at all.

4. A "learn-more-about" section. This would have a number of entries, each of which is a "home" page for material on that topic. Topics might include such things as:
   • packages
   • setf-expanders
   • pathnames
   • multi-threading
   • foreign-function interfaces, etc.
   Each topic would have an editor, who is responsible for receiving submissions and formatting the ``home page'' for that topic. It is likely that all topics would have each certain features such as:
   • a friendly overview
   • pointers to the relevent parts of the hyperspec and CLtL2
   • a listing of pending proposals for changes to ANSI
   • a listing of applicable papers
   • a listing of related available Common Lisp source code.
   • contact information for the editor.

Applications

1. No one should be listed unless some sort of contact information is provided, including a name of someone from the project who is willing to talk about the use of Lisp in the project.

2. The information should be set up in such a way that it should be easy for users, using search engines or other means, to locate users with combinations of properties such as:
   • Geographic location
   • Industry
   • Vendor
   • Utilities used
   • Additional software
   • On-line availability of describing documents
   • Availability of resulting software, commercially, free, or otherwise.

3. Wherever possible, the home webpage of the corporation should appear as a link. When this happens, we should also have a button which would use W4 to search that hierarchy for references to lisp. This would make it easy for people to locate all the people within an organization who are making web-public comments about Lisp.

4. Twistdowns are probably a good idea here. See Crafstman Tools search for a non-CL-HTTP example. This should be coordinated with activities for the Tools Area.

Tools

1. The FAQ should be reviewed, with any missing available tools added to this page.

2. A different organization which broke this into a number of smaller pages wouldn't be a bad idea.

3. Twistdowns are probably a good idea here. See Crafstman Tools search for a non-CL-HTTP example. This should be coordinated with activities for the Applications Area.

4. See also, the automated maintainance planned for the site.

Community

1. The listing of research papers needs to be overhauled into a proper research bibliography. This is a major task. However, it might be automated with a robot.

2. Who's Who in Lisp. This provides a place to list mailing addresses, websites and other contact information for people in the Lisp community, as well as brief descriptions of their interests. Options include:
   • List by invitation only. A board of editors (the ALU board?) would decide who is to be listed. This option is further qualified as to whether or not people should be listed without their permission. Of course, contact information should not be listed without permission under any circumstances.
   • List by request only. Anyone can submit their own entry to the Lisp Community Editor, who (after possible editing), would post it on the page. Perhaps this could be limited to dues paying members.
   • List by request or by editorial selection. In other words, we start with a list of notables and allow people to add themselves.

Reference

1. The CL-HTTP documentation should be available at the ALU site. This is necessary for anyone maintaining part of the site remotely if they don't have their own working copy of CL-HTTP. It is also a service to the community.

2. It would be nice if links to the local hyperspec were inserted as appopriate.

3. The MOP has been converted to HTML by hand. It woult be nice if this were done using more rigourous automatic tools, with indexing, etc, as was done for the hyperspec. (Any volunteers from Harlequin, Kent?) Extra credit for inserting links from the MOP to the local hyperspec pages.

Systems

1. The editor for this section should establish a uniform format, including contact information, for each system entry, and contact the vendors to have them supply the appropriate information. The current material, some of which is editorial from the FAQ or CMU, should be moved, if appropriate, to a discussion thread.

   Note that we want to have entries for all implementations, even if a vendor has a website, because we want searches limited to the ALU site to cover the system descriptions.

Other Issues

1. A language comparison section should include material covering the differences between Lisps. In particular, I would like to see a guide to two-way porting between Common Lisp and each of ISLisp and Scheme (perhaps others). The guides should include text and compatibility code, and would be used:
   1. In helping people understand which language to use for a given situation. (For example, it seems to me that ISLisp is suitable for people learning Lisp, before they move up to Common Lisp, but this is before actually writing such a comparison, so...)
   2. To actually make use of source code writen in a different language, with the least amount of fuss.
   It is not clear which area of responsibility this falls under.

2. It would be nice to include material related to other Lisp dialects, including Scheme, ISLisp (including ILOG), XLISP, AutoLisp, and Interleaf. However:
   1. It can make it difficult for readers to locate what they are looking for. There are a number of techniques we might use:
      • Search engines, with keyword systems, neither of which are yet implemented.
      • A color-coding scheme which distinguishes material as being of general Lisp interest, Common Lisp, Scheme, ISLisp, etc. If enough material on a subject is distinguishable by dialect (such as in Tools, Implementations, or Research, then separate (color coded) pages would be used. Otherwise, the color coding would be used for the text. How this can be integrated with other color coding schemes for this site is not clear.
   2. I am hesitant about including material related to other languages in some areas, but not in others. For example, if we include Scheme material in general, but not for Tools or Research, then readers might get a false impression of the quantity of material available.
   I personally believe that ISLisp could be allowed to sneak in without fundametally changing the concept and without introducing the above problems. Of course, other people might want to let in ``just'' XLisp, or ``just'' ...