The Davenport Group (aka Initiative for On-Line Documentation and Publication) Meeting summary for September 24 and 25, 1991 by Dale Dougherty Meeting location: Interleaf, Waltham, MA Next Meeting: As yet unscheduled. ******************************************************** CONTENTS: ATTENDEES MEETING SUMMARY NEXT MEETING AND OTHER OPEN ISSUES HaL PRESENTATION SLIDES ******************************************************** *ATTENDEES The following people attended the meeting: Dale Dougherty, O'Reilly & Associates H. Garrett Long, UNIX System Labs Kathy Murphy, Digital Eswar Priyadarshan, Sun Microsystems Matt Koehler, Sun Microsystems Susan Windheim, Prime Computer Joan Karp, Prime Computer Marcia Allen, HaL Computer Gary Epple, HaL Computer Rich Pusateri, HaL Computer Linda Walsh, O'Reilly & Associates David Scarbro, Interleaf Donna Tramontozzi, Digital Phyllis Bregman, Prentice Hall Stephen Uitti, Interactive Systems Mark Hamilton, Frame Technologies Lar Kaufman, Technical Arts Group Dave Dow, Fulcrum Technologies Martha Hughes, Technical Arts Group Ken Leese, Fulcrum Technologies * MEETING SUMMARY On the first day, Rich Pusateri and Gary Epple from HaL Computer made a presentation showing that a standard for data interchange is vital to the development of *open* on-line documentation systems and hypertext browsers. They proposed that we support the work being done by OSF/UI to develop an SGML DTD for technical documentation and that we incorporate the work of the HyTime standard for hypertext documents. They also proposed to lead an effort to demonstrate the viability of an open-systems architecture based on that standard. They are looking for other members of the Davenport Group to participate in the development of supporting technology. Furthermore, they'd like to aim for a public proof-of-concept demo: that documents from multiple sources can be processed into and out of a single format that can be processed by a variety of browsers. For instance, we discussed taking several types of UNIX documentation, including man pages, and preparing them for display on several different browsers. It is not enough to propose a standard, especially one as complex as SGML; we must develop the tools and the expertise that allows that standard to be incorporated into existing documentation and publishing processes. For instance, we can help provide filters for translating into and out the standard. Similarly, we need to work with developers of on-line documentation systems and browsers to understand their needs and gain their support for the standard. Also, an SGML-based interchange standard does not touch upon some areas that are critical for electronic display: o graphics o formatting o indexing Another issue is that there are are areas in which, despite on-going standards efforts, there is yet no specific standard that one can implement. Thus, in absence of a standard, we must come up with interim solutions if we want practical results in the short-term. For instance, because the OSF/UI DTD is not yet available, HaL proposed that we use the Prime DTD as an interim standard. It is publicly available now -- and has the additional virtue of being simple. In a nutshell, we'd like to have a working group that pursues the implementation of systems based on the proposed standards. This effort can also lead to providing critical feedback to standards bodies while the standards are still under development. The areas included in such a system are: o Writing/Authoring Environments o Filter Technology o Hypertext Markup Tools o Text Retrieval Technology o Graphics Rendering o Browsers In discussing these ideas, someone identified the interchange format, which was represented in the middle of an architectural diagram, as a "tuna can" and that terminology stuck throughout the meeting. The tuna can was where we wanted to concentrate our efforts. It was said that an objective of the Davenport Group was to agree upon what goes in the tuna can, not necessarily upon a particular approach to open systems. Also, we do not want to become a standards body; we want to build on existing efforts. The role of the HaL working group is to put something in the tuna can and begin working with it on different systems to discover its strengths and limitations. That is why they proposed using the Prime DTD with a subset of extensions from the HyTime standard. David Scarbro of Interleaf proposed that we could avoid standardizing on an interchange format if we could instead agree upon a distribution format. This format could be read directly by browsers, and might support multiple modes. Any browser might be equipped with certain "drivers" that allow it to read certain kinds of data (ASCII, MIF, Printerleaf, PostScript, TIFF, SGML). The idea, I believe, was that you'd supply the data in a number of different formats, capable of being read directly by a number of browsers. The browser would specify which formats it could handle and the driver would deliver it in that format. We spent a good deal of time discussing the merits of these two proposals. Finally, we had to decide upon a particular direction for the immediate focus of our efforts. I asked if we wished to foster the development of an interchange format for data containing links OR a distribution format for CD-ROM data that can be read by multiple browsers. By vote, the interchange format was chosen decisively. I would like to add that it was important that we choose a particular direction. However, that does mean that the two choices are exclusive. We could decide to support efforts to develop a distribution standard sometime in the future. On the second day, we clarified the statement of purpose that we had voted on the previous day. "Establish an ASCII format for data containing hypertext links that can be produced by multiple vendors and can be interpreted by on-line documentation systems and browsers from multiple vendors." Finally, we talked about our representation in efforts to develop standards. Several people gave us a progress report on the efforts of the OSF and UI to work jointly to develop an SGML DTD for technical documentation. As their meeting overlapped with ours, there is undoubtedly new information of interest to us now available. We identified other standards efforts that we'd like to know more about and asked a volunteer to report by e-mail. CD-RDx (Dale) DSSSL (Donna Tramontozzi) NIST (Donna) Hypertext '91 (I will distribute program description.) HyTime (Rich/Marcia) Lar Kaufman gave a presentation on the first day of the meeting about his efforts to standardize a common man page format. His Committee for the Common Man will be developing a man page DTD and making specific recommendations for UNIX system vendors. Garrett Long said that USL is committed to providing man pages in SGML somewhere in the future. *Next Meeting and Other Open Issues The next meeting is as yet unscheduled and I will send out a separate notice. The meeting will be held on the West Coast. I am trying to coordinate our dates with OSF/UI DOC SIG meetings and prevent overlap. We talked about having a meeting on the West Coast to brief those who were unable to attend the East Coast meeting. In particular, we wanted to go over the HaL proposal in detail. Please respond if you'd like attend such a meeting, which would be held in the next week or two in the Bay area. We can do it in a couple of hours -- the intention being not to open new ground in this meeting. In the meantime, Rich and Gary have promised to develop the HaL working-group proposal in writing. This should provide more of the technical details to help others evaluate the proposal and determine their level of involvement. *HaL Presentation Slides Here are the slides from the HaL presentation. 1. Strategy Proposal HaL Selected Technology o SGML o Prime-based DTD o HyTime o PostScript (display of graphics & printing) HaL Committment o Product Plan / Development Team Set o Starter Kit to be available o Results of research and analysis o Coordination for Proof of Concept Demo 2. Architecture Major Components o Writing Environment o Filter Technology o Hypertext Markup o Text Retrieval Technology o Browswers 3. General Architecture (diagram) Authoring Environment --------------------------------------- | | | | non-SGML docs | | | (non-HyTime links) | | | | | | Filter Technology | | | | | | Hypertext | | Markup Tool | | | | | | | ________________________ ( ) ( SGML/HyTime ) ( Docs ) ( ) <-[called the (________________________) "tuna can"] | | |-------------> Full-Text Indexing Tool | | | | | | --------------------- | | | (-------) (Format Prefs) -->| Browser | ( ) | |<-------( Index ) | | (_______) ---------------------_ | | (end user) 4. Standard Information Format Goal: o Hypertext document interchange among various information providers and consumers. Achieved by: o Identification and support of a standard hypertext document interchanged format. Format Components: o SGML o Select DTD's o HyTime o PostScript 5. SGML Description o Describes document structure with generic markup Criteria / Rationale o International standard o Industry trend with growing list of ISV products o Represents information content Current Status o ISO Standard 8879 o Endorsed by over 37 international organizations o Target of OSF and UI DocSIGS Supporting Technology o ARC SGML Parser o Amsterdam Parser and SGML to LaTeX filters 6. HyTime Description o Hypermedia/Time-based Structuring Language o Based upon SGML Criteria / Rationale o Draft International Standard o Non-proprietary o Provides model for hypertext interchange Current Status o ISO Committee Draft Standard 10744 o ANSI Project Number X3.749-D Supporting Technology: o Currently none. 7. DTD Support Description o Prime-based DTD -> for technical reference data o Other DTD's supported (e.g., man page DTD) Criteria / Rationale o Currently exists and publicly available o Simpler than most public DTDs o Requires changes to support other documents Current Status o Prime DTD undergoing minor revisions o HaL DTD will be revised at new document types are identified Supporting Technology o ICA Devegram DTD building tool o SGML parsers 8. Graphics Data Format Description o PostScript o X Raster Bitmap Format Criteria / Rationale o Existing standards o Simple Implementation Current Status o Concretely defined standards Supporting Technology o X Window System o Display PostScript o Publicly available troff to Postscript filters o NeWS 9. Proof of Concept Demo The Demo o 2Q92 Target o Demo Multi-vendor Interoperability o Public Demonstration (Show / Conference) o End-to-end solution o "Real" Data used as basis for Demo Logistics o Participant meeting o The plan (Who, What, When) o Spreading the message If you want more information, contact Rich Pusateri, HaL Computer Systems, Inc., 8920 Business Park Drive, Suite 300, 512-794-2855, (rtp@hal.com). As always, please send me your comments or suggestions. Thanks, Dale Dougherty O'Reilly & Associates, Inc. 1936 Coffee Lane Sebastopol, CA 95472 707-829-3762 dale@ora.com