WWW Activity at Hypertext '93

November 29, 1993

This report is of the World-Wide Web Birds-of-a-Feather (BOF) group meeting at the Hypertext '93 conference in Seattle, Washington. BOF groups were informal meetings specific to common-interest topics which anyone could organize and attend. For comments, etc. on this document, contact Kevin Hughes at kevinh@eit.com.

Table of Contents


General BOF Notes

The WWW BOF meeting, called "World-Wide Web Brainstorming," was organized by Kevin Hughes and took place on November 17, 1993. Attendees were able to sign up on BOF sheets, and the Web BOF attendance quickly filled to its maximum capacity of 50 people. By far the group was the largest BOF meeting of the conference - even people who hadn't signed up attended.

Because of its size, conference organizers allowed the meeting to be held in one of the grand ballrooms, instead of a smaller meeting room on the floor above. In that room someone had written on the chalkboard:

	WWW moved to Grand Ballroom B
	(303 redirect)
Unfortunately, the group had to wait until a conference session finished and cleared the ballroom, starting the meeting roughly half an hour late. This turned out to be one of the biggest drawbacks of the meeting.

Kevin handed out a sheet outlining some current Web issues, controversial questions, and possible future developments.

Issues included:

Possible applications included:

Questions included: At this point Kevin suggested that because of its size, the meeting should be broken up into the following groups. The topics were only guidelines and were decided upon given their popularity in earlier sessions at the conference.

  1. The SGML/HTML/protocol tweakers' group
  2. The Web research/exploratory applications group
  3. Web "real issues and we'll solve them, dammit" group
  4. Web newbie group Everyone broke into groups - they all turned out to be roughly the same size. Jon Mittlehauser from the NCSA offered to go to the new users' group to help out.

    A WWW courseware group sprouted up and currently has about a dozen people on an email list. If you're interested in it, please send email to www-courseware-request@eit.com with the body subscribe.


    WWW BOF Group 1 Notes


    Kevin Hughes

    Some time after the conference, this group proposed a fix to the then-broken HTML DTD. They submitted their document to Tim Berners-Lee, who then replaced the "official" version with it.

    The proposal letter the group wrote is here.


    WWW BOF Group 2 Notes

    Web research/exploratory applications

    Kevin Hughes

    Topics bounced all over the place, ranging from wish lists to general comments. Here's what I was able to jot down:

    After the meeting was cut short due to time constraints (a new speaker had to set up in the ballroom), Kevin recommended that those who wanted to continue could move upstairs to the meeting room.

    A small group from sessions 2 and 3 formed upstairs and continued to discuss Web things for the next hour or so.

    The group included:

    Assorted notes:

    Some discussion revolved around Robert's views that the "information highway" will replace the phone system. He said that the Internet is held up by universities and individuals. He talked a little about how CERN's site went down for a while due to a collapsed roof! I suggested that he send the story to RISKS digest. There was the view that the replacement of the phone companies was not a fiber network issue, but an Internet issue.

    Some discussion about how Mosaic could use better, more specific error messages if a site is down.

    There was talk about corporate attitudes towards free software - at the conference quite a few Web-like corporate solutions reared their heads, and although they were expensive, we wondered if many companies would overlook Web solutions because of the fact that much work done on it is free. To a corporation, does "free" mean "of lesser quality"?

    Robert stressed the idea of content value. Individuals are making mini services, servers, and resources of their own. Brian said that the net gives a whole new opportunity for contact between customers and companies.

    Terje discussed his idea of being able to include pieces of HTML documents or media in other HTML documents in a media-independent manner. That is, instead of using <img src>, use something like <include>, and specify any URL, perhaps with a byte offset and length. In his speech at the conference, this is what Ted Nelson referred to as transclusion, something he emphasized greatly. I thought it was a really nice idea.

    There was some discussion of the creation of a yellow pages-like service for the Web. Could it be something distributed and updated voluntarily and regularly like FIDOnet, asked Adam. There were some opinions that using Web-walking robots (spiders) to do this was not such a good thing. Perhaps the use of templates that could be sent in regularly would do the trick?


    WWW BOF Group 3 Notes

    Web real issues

    Brian Smithson

    Notes from WWW Brainstorming BOF, "Real Issues" breakout group, 11/17/93 1:00PM - 3:00PM. Also includes notes from when this breakout group combined with the "research/exploratory" group.

    Attendees (for some or all of the breakout session):

    General thoughts

    This breakout group was formed to discuss current WWW issues which "must be solved for the Web to (a) become useful to the Internet community, and (b) become more robust and scalable". Instead, the group seemed to consistently drift toward discussion of larger issues which face the Internet as a whole.

    I've organized (using the term loosely) these notes by topic rather than chronology or speaker.

    Policy, funding, etc.

    We talked generally about the "future of the Internet" issues, such as acceptable use, backbone funding, and media policy. We didn't solve any of these issues :-(, but the discussion led to:

    WWW community and organization

    At present, there's no real governing body for the Web. The de facto centers are NCSA and CERN. We talked about the need for a clearinghouse for meetings, standards activities, etc. The recent developers conference helped, but we need a structure for connecting independent developments and for taking in-house solutions to the public.

    We also talked about the pros and cons of specification-driven versus implementation-driven "standards". Specification-driven standards tend to take much longer to develop and become unwieldy in the process (like X.400, or maybe SGML :-); implementation-driven standards get to market more quickly but can be limiting by early implementations and are potentially dominated by one or a few implementors (like MSDOS, or maybe HTML :-).

    The usefulness of a WWW User's Conference was also discussed. WWW has become so large that it would be an expensive undertaking, and funding would be a big issue. If it was "invitation-only", then it might be too elite; if it was open to the public, then it might be too big to be effective.

    Using WWW for commercial applications

    Several companies were represented which are using or want to use WWW for commercial applications. So far, most commercial uses have been for delivering free or promotional information. To move on to some kind of pay-per-view information or other paid services, we'll need security, authentication, and ways to validate logins, charge for resources, and handle remittance.

    One thing which was brought up is the general fear of processing credit card information over the Internet. Is it that different from (e.g. less secure) telephone or mail order? The French Minitel system was cited as an example which might be useful when considering a commercial WWW.

    Commercial products for WWW

    A few commercial vendors at HT'93 mentioned that they were planning or considering support for HTTP or HTML. Does the WWW need commercial products in order to continue to grow and penetrate commercial applications? Some corporations won't use freeware, for various reasons. 3rd party support from companies like Cygnus makes it more "legitimate" to corporate users. Still, WWW is not a shrink-wrap product -- you need IP, you need to get through firewalls, and you generally need someone who is knowledgeable and can track fast-moving developments in the WWW community.

    Link maintenance and cross-server issues

    Cross-server issues: when one references a document on another server, how do you ensure that the other server continues to make the referenced document available?

    Authoring tools: manual editing HTML doesn't scale up well. It gets too messy and cannot be checked for consistency and link continuity. Paid-for commercial information services will help encourage quality and stability, but they will need authoring tools to do so.

    Indexing and directories

    Finding things on the Web was cited as a problem. Should there be One Big Index? Self-indexing servers which can be searched like Veronica/Gopher? Should indexes be automatically-generated or human-generated? Is there a useful business in value added index services?

    Service quality

    Server availability: recent sluggishness and downtime of critical servers brought up discussion of redundancy and replication of services.

    Content value: quantity and quality of information content on the Web varies widely from one subject to another. There is a large body of computer-related information, followed by other scientific and academic subjects. However, other subjects are not well represented; for example, there is little business-related information on the Web.

    A style and quality guide was proposed as a mechanism to help make the information content more consistent and raise quality standards. Paid-for commercial services would also raise the standard for quality.

    Dial-up access

    Some thought that PPP/SLIP is good enough, particularly with agile IP assignment. Others thought that a more simple dial-up mechanism (serial HTTP?) would make the Web available to a larger population, such as those who now access services like Compuserve or Prodigy.

    Multimedia issues

    An "include" function was desired which will let have an in-line instantiation of a referenced document.

    For those browsers that don't support graphics, or for those users who are at the end of a slow IP link, it was suggested that a nice feature would be to allow text in place of graphics. This text would not be displayed for browsers/users who get graphics.



    Drawbacks and Suggested Improvements

    Because this was the first such meeting of Web folk in such large numbers, obviously some things could have been done better. Here's a list of drawbacks to the meeting structure, and some improvements or changes that could be made:



    Despite the limited time, a lot of Web brainstorming occurred. Many people said that they'd like to continue the discussions over the network and/or at a later meeting. A small group that was interested in discussing Web courseware formed, and a similar movement has been forming on alt.hypertext. For those interested in the general BOF meetings mailing list, or the Web courseware mailing list, please send email to www-courseware-request@eit.com with the body subscribe.

    Individuals who wish to continue discussion will have to decide how to communicate with each other - if anyone begins a mailing list, IRC chat, annotation server, or similar resource as an extension of the BOF meeting please announce it in alt.hypertext and/or comp.infosystems.www.


    Web Sightings at Hypertext '93

    Courses, Posters, Demos, Papers, Panels, Speeches

    There were no Web-specific courses, papers, or research, and in the first few days after a lot of talking I got the impression that there was relatively little knowledge of the WWW among many corporate attendees and hypertext researchers. Many of the student volunteers were graduate students doing their Masters' on hypertext; many had not heard of the Web, and about a dozen showed up in the BOF meeting's new user group.

    No Web-related literature was readily available to conference participants. I made 200 copies of my Guide to Cyberspace and put them by the registration booths. Later, because of demand, the conference organizers generously made a hundred extra copies for me. All 300 copies were gone in three days; the first 200 were snatched up in about four hours!

    The course "Introduction to Hypertext and Hypermedia" by Jakob Nielsen of Bellcore had a good turnout; a general overview of the Web and related NSFnet statistics were shown. In "Forging the Business of Hypertext Publishing," Dale Dougherty from O'Reilly and Associates spoke of the Global Network Navigator and its details.

    Ted Nelson's speech on the 15th gave everyone who was wondering the official word on Xanadu (tm): it's still alive, and he's working on it with his own company in Sausalito. He spoke of wanting to use it via a client and server solution with Gopher and telnet, but that's as far as he got regarding the Internet and his own visions.

    He talked of the decades-long history of his hypertextual ideas, some of which could stand to be researched a little better and even implemented on the Web. His speech was called "Above and Beyond Hypertext: The Inexorable Logic of Metamedia Publishing," and he talked some about the idea of online publishing and distribution with Xanadu.

    A few people I was with realized that he was talking about hypermedia authoring in a Web-like environment. I was by no means the only person that wanted to ask him his opinions of the Web. Vladmir Chaloupka, who runs a physics Web server at the University of Washington, asked him and he got a very brief reply. Ted thought the Web was great: it was missing one or two of his hypertext constructs, but it was otherwise fine in his book.

    Nevertheless, quite a few people got the feeling that Ted Nelson was running a little behind these days.

    A review of Nelson's speech in TidBITS #204 is available, as is an older TidBITS issue describing Xanadu in more detail.

    In the demo room, nearly a whole wall of booths were comprised of Web-related presentations:

    "Navigational Search in the World-Wide Web"

    Renier Post demonstrated an extension to the search capabilities of Mosaic that followed links. One can specify the depth of the search and many other parameters, and I thought it was quite nice. For more information, email Dr. Paul De Bra at debra@win.tue.nl.

    "Using World-Wide Web Hypertext as a Generic User Interface"

    Steve Putz, using a 14.4k connection, was able to show off his interactive map at Xerox PARC fairly well. He can be emailed as putz@parc.xerox.com.

    "A CAI system for linear algebra based on World-Wide Web, Xmosaic and a few special purpose programs"

    There was no setup at the demo booth, but Terje Norderhaug showed his poster on the MATHrix project from the University of Oslo's Department of Mathematics. The project explores the possibilities of learning math on the Web, and he showed off how extensions were made to HTML to allow it to include TeX and plots and animations from Mathematica by inserting kernel calls. It looked great, and the contact is Bjorn Remseth at rmz@math.uio.no. Try the server at http://math-www.uio.no/mathrix/top.html.

    "World-Wide Web Hypertext in Physics Research"

    Vladmir Chaloupka from the University of Washington brought a NeXT cube and showed off his physics server locally. He's experimenting with collaborations in large-scale physics projects over the Web and can be reached at vladi@u.washington.edu. His server is at http://web.phys.washington.edu.

    "Developing Global HyperMedia: The NCSA Mosaic System"

    Both Chris Wilson and Jon Mittelhauser, WinMosaic developers, were on hand with a PC and a 14.4k SLIP connection to show off their creation. I don't think they got any rest. You can reach Chris at cwilson@ncsa.uiuc.edu.


    Although it wasn't printed in the program, Dale Dougherty had a booth to show off O'Reilly and Associates' Global Network Navigator and hand out little GNN brochures. He kept busy, was asked a lot of questions, and showed off a prototype commercial form or two that looked really nice. He's at dale@ora.com.

    "The SuperBook (R) System for Enterprise Efficiency"

    Bellcore showed off their "corporate-strength" SuperBook system, a multiplatform hypertextual electronic document distribution system that runs off the client-server concept and supports SGML DTDs. Sound familiar? The most-often asked question at the booth was "Yeah, but can it read HTML?" Contact Carol Lochbaum at ccl@bellcore.com.


    Seen On The Freebies Desk

    Join ACM SIGLINK, the ACM's Special Interest Group on Hypertext/Hypermedia. Email eyoder@clarit.com for details.

    A brochure about IRIS InSight, a multimedia multiplatform client-server distributed electronic document system from Silicon Graphics. Call 1-800-800-7441 in the US.

    A flyer selling three documents - "Museums and Interactive Multimedia", "Hypermedia and Interactivity in Museums", and "Interactivity in American Museums". Virtual curators can call (412) 683-9775 or fax (412) 683-7366.

    Call for papers to be published in a special section of the Journal of Management Information Systems. The topic is "Accessing Large Textual and Multimedia Information Bases." Contact Tomas Isakowitz, tisakowi@stern.nyu.edu for more information.

    Brochure for SGML '93, the annual conference of the SGML Technical Community, December 6-9, 1993, Boston Massachusetts. Phone (703) 519-8162 for registration quick!

    The preliminary call for participation in the second ACM International Conference on Multimedia, October 15-24, 1994, San Francisco, California. For more information, contact Professor Patrick Mantey, mantey@cse.ucsc.edu.

    A call for participation in the European Conference on Hypermedia Technology (the 6th international conference on hypertext), September 18-23, 1994, Edinburgh. Contact IRitchie@acm.org for details.

    Call for participation in the ACM 1994 Conference on Computer Supported Cooperative Work, October 22-26, 1994, Chapel Hill, North Carolina. Email cscw94@cs.unc.edu or ftp ftp.cs.unc.edu in the /cscw94 directory.