*************************************************************** ****************** WELCOME TO SGML NEWSWIRE ******************* *************************************************************** * * * To subscribe, send mail to sgmlinfo@avalanche.com * * * * (Please pass along to interested colleagues) * * * *************************************************************** HYPERTEXT '93 SUMMARY ===================== The following is a summary of the Hypertext '93 conference. It was posted it to the alt.hypertext USENET newsgroup on 23 March 1994. Its author, Lynda Hardman (lynda@cwi.nl) has particularly good things to say about SGML and the WWW. Whither Hypertext? Some personal impressions from Hypertext 93, Seattle, Washington It is now 6 years since the first hypertext conference, HyperTEXT '87, was held in Chapel Hill, North Carolina. The attendance of this and subsequent conferences has always been a good mix of people from technical areas and the creative arts, but the actual numbers have not grown significantly. A European Conference on HyperText was introduced on the even years in '90 and '92, and in '94 will be held in Edinburgh, although ECHT '94 will be European Conference on Hypermedia Technology (more about this later). The mood of the conference has gone from an initial period of introspection, with questions about what hypertext is (or should be), to one of a more mature discipline where we can take time to look at the underlying information issues rather than being caught up with details such as what visual representation link markers should have, and whether text should scroll or not. Hypertext is to be found all over the place these days - for example as a recommended topic in conferences on Human Computer Interaction, Databases, Information Retrieval and even Artificial Intelligence. In the commerical world too there are plenty of word processors which have basic link following and creation capabilities. But what does the community have to show for its existence? Here is a summary of the contributions I think the field has made (and has still to make). (1) Mature projects There are a small number of examples of projects that have had the time and resources to develop from trying out new ideas to having many people use the system for an extended period. One such project which has influenced and inspired the field is the Perseus Project (presented in a Technical Briefing at the conference by Elli Mylonas) The goal of the project is to collect together works of ancient Greek literture, history and archaeology and make them available via a flexible hypertext environment. The project has had two main phases, the second of which was due for completion at the end of '93 with the release of Persues 2.0, two years after the release of Perseus 1.0. The differences between the two releases are perhaps not in themselves surprising, but serve as examples of how such specialised information environments can be made robust, in terms of information ageing. A fundamental decision in the project was to concentrate on the structuring of the data rather than the software. This was stored as SGML which could be converted into formats required by the final delivery system, for which the developers chose HyperCard. The navigation and index mechanisms were developed in close collaboration with users of prototypes of the system, ensuring they were well suited to carrying out typical tasks. Examples of applications in the commercial arena are sparsely scattered through the academic literature, but the "hypertext engineering" work of Robert Glushko can be found in the Hypertext 89 conference, and the industrial requirements for hypertext are described in Hypertext 91 (Malcolm et al.). (2) Dexter model The Dexter model was developed to get at the "essence" of hypertext structure. Developers of different existing hypertext systems started work on the model in 1988, and Halasz and Schwartz presented a paper at a NIST meeting in January 1990. The model is published for the first time in an accessible journal in the Communciations of the ACM in Feb 94. While this has made it difficult to get hold of the Dexter model (how many of us have a photocopy of a photocopy?) it has not stopped further work on improving and extending the model (which is reported alongside the Dexter model in the CACM issue). The question is what happens next with this work. In order for these models to be useful there need to be systems which support them. Perhaps SGML is one approach... (3) SGML --> HyTime SGML is much more tangible than the Dexter model (or its descendants). And what is more, everyone seems to be using it, or a home-grown version therof. We are at least learning that it is useful to record the structure as well as the content of information, so that when the current generation of computers has been replaced we can still access the information we put together many years ago. (A good example is the documentation needed for aeroplanes. This needs to last the life of the plane - a lot longer than your average computer, or even software package. Paper is sufficient for the storage-term problem, but it is physically unwieldy and cannot be easily searched.) So even though I have my doubts about the immediate applicability of the Dexter model, there is something to be saved through the SGML route. The hope is that the more complex HyTime will also catch on. It cannot have escaped your attention that in the same way that "hypertext" systems blossomed 3 or 4 years ago, we now have hoardes of "multimedia" systems escaping into the market place. People *must* be noticing that once they have spent a *lot* of time, energy and money on creating a great multimedia presentation, when they want to show it to their friends/colleagues/fellow researchers it is no problem at all - so long as their friend/colleague/fellow researcher has the same computer, same videoboard, same screen..... Creating a system that supports (the authoring or the playback of) HyTime documents is no easy task, but it is one that will have to be addressed at some point. (4) WWW A successful enterprise that has been growing from the ground up is the World Wide Web, started at CERN, Switzerland. The number of users and contributing sites has been growing slowly but steadily, until, that is, a group at NCSA built an XWindows based reader they call Mosaic. This allows access to the network of information stored at various sites on the internet using a point and click user interface. Problems with ftp, gopher (you name it) protocols are kept well hidden from the end user. Since the introduction of Mosaic the numbers of WWW sites is growing FAST. If you don't know how to access WWW then "telnet info.cern.ch" for a simple text-based browser, and if you would like a copy of the Mosaic browser look in "ftp FTP.NCSA.UIUC.EDU". The language user by the developers, HTML, is an SGML style language. The language specification can be found in: http://info.cern.ch/hypertext/WWW/MarkUp/HTML.html You can always read exactly what was discussed at HT '93 about WWW: http://www.hcc.hawaii.edu/hypertext/ht93.report.html ----- There is great a need for communication between academic and commercial developers of hypertext systems and documents. Not that it is likely that both use the same tools, but there are long-term problems that need to be solved that the commercial world has no time for, and many short-term problems that have already been solved by commerce but haven't been written up in a conference somewhere. A commerical symposium was organised at HT '93, but attendees had to pay, making it unattractive for academic participants. We aim to find a better solution for this at ECHT '94. We are fortunate that the field already has a number of active commercially-oriented participants - Bob Akscyn, Knowledge Systems, Ian Ritchie, Heriot-Watt University (ex president OWL), and Bob Glushko (Passage Systems), to name but a few. Bob Akscyn, in particular, would like to get people together to tackle a BIG project, looking at cooperative hypertext/hypermedia, compatible information, and networked information. He would like you to get in touch with him if you are interested in participating in such a project, with ideas for how to go about it, and possible sources of funding. Contact: rma@centro.soar.cs.cmu.edu The next conference in the "hyper" series is ECHT '94. One of the themes for this conference is to integrate the participation of commerce and academia, so that those of us in the academic world can see what has already been done in the commercial world, and can get a view of the long term problems that need to be solved. Similarly, academics are already tackling big problems with success (WWW for example), through their strength in numbers, and without the constraints of proprietory information which could prove valuable test grounds for pre-commercial projects. Note that the deadline for submissions is 1 April 1994. Check the previous conferences out: HT 87 Chapel Hill HT 89 Pittsburg ECHT 90 Versailles HT 91 San Antonio ECHT 92 Milan HT 93 Seattle Lynda Hardman, CST, CWI, Kruislaan 413, 1098 SJ Amsterdam, The Netherlands email Lynda.Hardman@cwi.nl tel. +31-20-592 4127 fax. +31-20-592 4199 ************************************************************** * SGML NEWSWIRE LIST MANAGER * * * * Linda Turner * * Corporate Communications * * Avalanche * * 947 Walnut Street * * Boulder, CO 80302 * * sgmlinfo@avalanche.com * * linda@avalanche.com * * Vox: (303) 449-5032 * * Fax: (303) 449-3246 * **************************************************************