Hypermedia Research and the Web Workshop

The Hypermedia Research and the World Wide Web Workshop was held at Hypertext '96 in March of 1996. The overall purpose of the workshop was to set the stage for better communication between the hypermedia research community and the Web development community. Below is information about the workshop and some of the results of it. Update September 3, 2005: I pulled a few key files out of my archives and added to my new site.

Participant position papers

Applying Hypermedia Research to the World Wide Web

By Keith Andrews, Graz University of Technology, Austria. Drawing upon experience with Hyper-G, Keith points out the need for external link databases to help alleviate some of the Web's current problems.

An Evaluation of the World Wide Web as a Platform for Electronic Commerce

By Dan Connolly, World Wide Web Consortium. Dan evaluates the architecture and implementation of the Web with respect to Douglas Engelbart's requirements for an open hyperdocument system, which are derived from experience in using CSCW to support large scale electronic commerce.

A Web of Objects

By Paul De Bra, Eindhoven University of Technology, The Netherlands. Paul describes how, by using object oriented database technology for the storage level of Web servers, it is possible to incorporate many features envisioned by some well-known hypertext reference models. Unlike Hyper-G, this would preserve the architecture of the Web.

Extended Linking Facilities for the WWW

By Gary Hill, University of Southampton, UK. Gary talks about how, by adding some of the features of Microcosm to the Web, authoring facilities and navigation would be improved. The Distributed Link Service is presented as an example.

Research on Usability-Based Facilities for WWW Browsers

By Stewart N. T. Shen, Old Dominion University, USA. With a focus on developing browser features based on user needs (as opposed to advertiser needs), Stewart describes ways to assist beginning users, how to help experienced users find their favorite sites, and how to make it easier to create personal annotations.

Structured Web Site Design

By Daniel Schwabe, PUC-Rio, Brazil. Daniel describes the Object-Oriented Hypermedia Design Model and demonstrates how it can be applied to a World Wide Web site. OOHDM is comprised of four incremental and iterative activities, each of which involves building a set of object-oriented models.

World Wide Web Benefits and Dangers for (Traditional) Hypermedia Research

By Andreas Dieberger, Georgia Institute of Technology, USA. Andreas focuses on three main issues: navigation (applying navigation research to the Web), the user interface (the Web as a globalized user interface) and prototyping (using the Web for hypermedia research). He stresses some of the problems of applying existing research to the Web and using the Web for hypermedia research.

The Eastgate Web Squirrel

By Mark Bernstein, Eastgate Systems. Mark describes Web Squirrel, software to help users manange all of their Internet resources. It is a great example of how hypermedia research can be applied to the Web, since research on spatial hypertexts was applied to Web Squirrel.

Simplicity and Extensibility: What we can learn from the Web

By Roy Fielding, University of California, Irvine, USA. As a Web developer (Roy wrote the HTTP specification), he first asks why he even wants to be at the workshop. Although the position paper is not completed yet, you can gather from the outline and title that its focus is on explaining why the Web has been successful and how it will continue to be successful. It is simple by design and extensible by design. So, incorporating more hypermedia research won't be that hard. In fact, the designers are waiting for the help, in some sense. Just be careful to avoid the "non-solutions" that come about if you don't understand the social context of the Web.

Report on the workshop (by me)

(A version of this report appeared in the June, 1996, issue of the SIGLINK Newsletter.)

A one-day workshop at Hypertext '96 on Hypermedia Research and the World Wide Web was held March 17 in Washington, DC. I organized the workshop and was very anxious during the position paper phase before the conference: I had sent out over 30 personal invitations, but nobody was willing to participate in my workshop. I was beginning to wonder if I was the only person who saw this huge gap between the World Wide Web and hypermedia research communities. I even thought about canceling the workshop because of lack of interest. But after extending the submission deadline as late as possible and making a few more contacts, I was finally able to get a good group of participants together and the workshop was held. Below are some notes of what transpired.

As it turns out, the reason few people could attend my workshop was NOT because they did not also see this problem. Rather, everyone was already talking about it in other places. For example, both of the other workshops at Hypertext '96 spent time talking about these same issues. The discussion I had with others in the hallway during the technical program also told me that I was on the right track and the hypermedia community was beginning to feel like it should take some action to forge a closer tie to the Web community. Many people re-iterated the same questions: "Why aren't the Web developers looking at the existing research? Why are they reinventing the wheel?" I agree that the Web community is not paying close enough attention to the existing research, but since the workshop (and after addressing similar issues at CHI 96) I have developed comeback questions for these: "Why isn't the hypermedia research community more active in the Web? Why aren't they submitting more papers to the Web conferences? Why aren't they applying their research to the Web to show how useful their research is?" It is a two-way street, folks.

Anyway, back to the workshop. I had three main goals for the workshop:

1. bring together hypermedia researchers and Web developers
2. document hypermedia research's role in current Web development
3. lay out the Web's role in future hypermedia research

Goal #1 was accomplished, even though I did not get the 50/50 mix of hypermedia and Web people that I wanted. Representing the hypermedia research side were:

• Keith Andrews, Graz University of Technology and Hyper-G
• Paul De Bra, professor at Einhoven University of Technology
• Gary Hill, University of Southampton and Microcosm
• Stewart Shen, Old Dominion University
• Daniel Schwabe, PUC-Rio
• Andreas Dieberger, Georgia Institute of Technology and MOO/MUD navigation expert
• Mark Bernstein, well-known researcher and conference organizer from Eastgate Systems
• Marvin Pollard, graduate student at the School of Library and Information Studies at Florida State University

Representing the Web side were two big-time experts, fortunately for us:

• Dan Connolly, who works for the World Wide Web Consortium (W3C is the "referee" in the competitive game between Web vendors) and writer of the HTML specification
• Roy Fielding, graduate student at the University of California-Irvine and author of the HTTP specification (Actually, Roy could go in either camp, but he is more well-known for his HTTP work than his research in hypermedia-based software engineering environments, so I classify him as a Web expert in this context)

I want to publicly thank all participants, but especially Dan and Roy, since not many members of the Web community have realized that the hypermedia researchers do indeed have valuable things to contribute to the development of the Web.

The participants submitted position papers before the workshop. Everyone was given some time at the workshop to state their positions. We had worthwhile discussions on many topics. But what was most interesting was when one of the hypermedia researchers would identify a need for the Web (such as typed links) and then either Dan or Roy would come back with "It has already been specified, has been part of the Web definition for years, but we are just waiting from the vendors to implement it". It made me realize that you cannot get a good understanding of how the Web might evolve by simply looking at what Netscape Navigator does today and reading people's poorly-designed personal pages. You have to read the specs, see what new standards are being proposed, and you have to demonstrate to the vendors why they should implement some old feature from another hypermedia system. Why doesn't Netscape do typed links yet? Because their customers haven't told them they want typed links yet. Because Netscape does not see any financial reason to do typed links yet. It is up to the hypermedia research community to demonstrate to Netscape that typed links are good for users and for publishers and that they could sell more software if it utilized typed links.

Anyway, back to the workshop. Goals #2 and #3 were sort of merged together during the workshop. After the presentations, our discussion focused on "opportunities for improving the Web", a nice way of saying that in some ways the Web sucks. But the difference between saying "it sucks" and "here's how to make it even better" are important and crucial to getting people to listen to you. No one wants to hear a bunch of "old farts" whine about the "good old days." But companies are ALWAYS willing to listen to how they can enhance their products to make more money. So, we discussed different ways to improve the Web, being sure to mention both the existing research which suggests this is a good idea, and the existing Web specifications which would make realizing this opportunity possible today. We had to keep in mind that our audience for this list of opportunities was not other researchers, but rather the people at the big companies that are shaping the Web.

But we did not have nearly enough time. So, writing up the list of opportunities has been left as a post-workshop exercise. We have made some progress so far, but the going has been tougher than I anticipated. The first problem we ran into was specifying all of the things that need to be considered when suggesting an improvement. One should not just implement a new feature without considering the implications on the infrastructure of the Web, the users, and everything in between. Even the very best idea probably would not be worth it if it had huge, negative effects on network bandwidth and browsers, for example. You have to understand the entire Web to be able to make this kind of cost/benefit analysis. So, we came up with a template to fill out for each opportunity, being sure to include sections for existing research and existing Web specifications. There is also a heavy emphasis on users.

We also have a long, unorganized list of opportunities. The list needs to have different overlays on top so that opportunities are grouped in different ways. And each opportunity still needs to be expanded. But I feel this is quickly becoming too much for our little group of participants to handle. We could use some help from others, so if you see an opportunity that tickles your fancy, feel free to take charge of it and develop its criteria. If your opportunity involves the HTTP protocol, contact Roy to see what he thinks. If your opportunity might be implemented in HTML, contact Dan because he understands the history behind HTML and knows what is likely to come in the future.

Our goal is to get the W3C to put its "rubber stamp" on this document and have it presented to its members (the big vendors).

Also, one other document that we want to produce from the workshop is a reading list for Web developers: the key papers and books that they could read to make them aware of the larger field of hypermedia research and to make it easier for them to apply some of it to the Web. This list has not been compiled yet, so feel free to contribute your ideas. Send them to me.

So, in summary, the seed has been planted for better communication between the hypermedia research community and the Web community. But both sides will have to work hard to water and weed the plant to ensure that it will grow, bear fruit, and prosper. Please help out!