User Guide to the Conference

SGML '95 Conference Chair:

Yuri Rubinsky, President,

Conference Co-Chairs:

Tommie Usdin, Vice President and
Debbie Lapeyre, Sr. SGML Consultant,
ATLIS Consulting

Sunday, December 3, 1995


Monday, December 4, 1995

9:00 am - Conference Welcome

Norman W. Scharpf, President, GCA

9:10 am - The View From the Chair

Yuri Rubinsky, President, SoftQuad and SGML '95 Conference Chair and
Tommie Usdin, Vice President, ATLIS Consulting Group and Conference Co-Chair

9:40 am - Guide to the Conference

Debbie Lapeyre, Senior SGML Consultant, ATLIS Consulting Group, and Conference Co-Chair

9:50 am - SGML and Related Standards: An Update

Dr. James D. Mason, Convenor, ISO Working Group 8 and Oak Ridge National Laboratories

10:10 am - What's Next For SGML?

Dr. Charles F. Goldfarb, Consultant, Information Management Consulting
SGML's inventor analyzes some key information processing directions and the challenges and opportunities they present for SGML users. How will trends like online information services, Windows 95, and "SGML purity", affect you and SGML in the coming year?

10:30 am - Poster Session

12:00 pm - Lunch

TRACK 1: Newcomer

1:00 pm - Keynote: SGML on a Strictly Business Basis

Eric Severson, Vice President, Market Strategy and Corporate Development, Interleaf
SGML is often championed by technical enthusiasts and sold to others primarily on its technical merits. While these arguments are often valid, they are not always persuasive. Especially when technicians begin fervently debating the fine points, SGML often appears to business people more as a religion than a practical implementation strategy. However, the business reasons for structured documents and SGML are in fact quite compelling, and can be explained in terms that both senior managers and technologists will understand. This presentation takes a powerful look at SGML in its broader business context, covering the following critical points: The basic business issues motivating the need for structured information and vendor-independent standards -- how these requirements can be specifically related to operational management's need to improve business processes and to senior management's ultimate objectives of increasing revenue, market share, profitability and cash flow.

1:45 pm - Tools for Implementing SGML- Based Information Systems

Dianne Kennedy, Consultant, SGML Resource Center;
Bob Barlow, Consultant, SAIC; and
Kurt Conrad, Consultant, Sagebrush

Implementing SGML can be an enormous task. To be successful, an implementor must have a good technical background in SGML and must have a clear understanding of data flow and SGML system functionality. Gaining an understanding of the key components of an SGML system is critical. This presentation is designed to provide the SGML newcomer with an overview of the major classes of SGML tools and a brief review of the products commercially available today. Presenters for this session are independent SGML consultants who specialize in the design and implementation of SGML-based information systems.

2:30 pm - Break

3:00 pm - Tools for Implementing SGML- Based Information Systems - cont.

TRACK 2: User

1:00 pm - Keynote: Using SGML -- The Pain/Gain Ratio

Pamela L. Gennusa, Director, Database Publishing Systems, Ltd.
Most companies implement SGML in order to see an improvement in their business, to realize a benefit. For this improvement, they are willing to make an investment. But how much? This presentation examines the trade-offs when deciding to implement SGML. Examples are used to illustrate underestimated efforts that fail to realize any benefits, overambitious endeavors that demonstrate the principle of diminishing returns, and well-measured implementations that provide the company with both timetable and intangible benefits for their investment.

1:45 pm - A Different Way to View Your Data Using SGML Content Tagging

Chris Wheedleton, Information Technology Engineer, SAIC
This presentation will discuss the prototype system that SAIC has created for the Defense Mapping Agency. SGML is used to place DMA navigational documentation on CD-ROM which will give navigators a greater functionality in using navigational information. The presentation will show the DTD and its content tagging schemes and demonstrate the prototype model of the CD-ROM data.

2:30 pm - Break

3:00 pm - Entity Management Across Groups and Platforms

Beth Micksch, Senior Manager, Standards-based Solutions Group, Intergraph Corporation
Systems and general text entities can easily be managed and manipulated by an SGML author or technical writer. Some of these entities require standardization within a corporation so that they can be shared, but not modified by a random user. Others must be customized for a particular document or group. A description of tools and methods used to accomplish this task as well as a description of actual user environments where it is implemented will be outlined. Special features used in document translation will also be discussed.

3:45 pm - From Cut and Paste to SGML: Building a Single Source Library

Maria Miller, Senior Technical Publications Specialist, and
Bob Tabone, HEADS Program Manager, Hitachi America

Hitachi America has defined a strategy for electronic document production and publishing of its technical documentation that supports their objective of "create once, publish many". This presentation will discuss this strategy which is called the Hitachi Electronic Applications Document System (HEADS). It will outline what the strategy calls for and discuss how the Pinnacles DTD is a foundation. It will cover the steps that were taken to create a single source library for technical documentation. Hitachi America receives its technical documentation from authors in Japan who work in a mixed hardware and software environment. In addition, the technical documentation is both authored in English by Japanese speakers and authored in Japanese and then translated into English. In order to move to a single source library, Hitachi America had to work within the limits imposed by the infrastructure that exists in Japan. The challenges of this process will be discussed.

4:30 pm - Graduating from File-based Document Assembly to Info-Based Document "Construction"

David Sklar, Director of Applications, Electronic Book Technologies
SGML's entity features support the assembly of a document from a collection of modules that must be precisely enumerated and absorbed in toto. Breaking out of these limitations enables publishers to graduate from mere assembly to highly sophisticated "construction", in which novel SGML documents are auto-generated via combinations of queries (in SGML infobases and/or relational databases) and SGML transformations. This presentation discusses the limitations of assembly via entities and an introduction to the concept of document construction (DC). A description of current proprietary implementations of DC engines is given as well as a proposal for the role of standards like HyTime and DSSSL in the continuing evolution of DC technology.

TRACK 3: Expert

1:00 pm - Keynote: The HyTime Corrigendum -- SGML's Future Passed

Dr. Charles F. Goldfarb
The ISO recently approved changes to the HyTime standard (ISO/IEC 10744) that have profound implications for general use of SGML as well. These result from the alignment of DSSSL and HyTime, and from the independent specification of "general facilities" on which the HyTime architecture depends. These facilities, which can be used by documents and systems that do not necessarily support the HyTime architecture, include: architectural forms, including derived architectures with multiple base architectures, entity management that supports Formal System Identifiers with URLs, multibyte character sets, and archives (e.g., TAR, SBENTO), and a common model -- shared by HyTime and DSSSL -- of information recognized by an SGML parser. These and other related issues will be discussed.

1:45 pm - SGML for the Programmatic Reconciliation of Legal Documents to Arbitrary Time Points

Nick Carr, President, Allette Systems
This presentation will discuss the challenges managers face with the requirement to see what a consolidated document looked like at any specific point in time. An example of this is the requirement for lawyers to see the consolidation of an Act and its Amendments at the time an offense was committed rather than the present. While document databases will store and manage versions and changes (deltas) to documents, managing the actual semantics of the change is a problem that has been very difficult to address. This presentation will demonstrate a technique for solving this issue regardless of the database or editing environment.

2:30 pm - Break

3:00 pm - Publishing in Eleven Languages

Jean-Pierre Gaspart, Managing Director, ACSE
This presentation will discuss the document process for simultaneous publication in 11 languages of the 1500 page Budget of the European Commission. The creation, translation, proofreading, editing and publishing process is highly complex and involves various parties each using their own SGML or non-SGML system to feed data into a central SGML repository. The system is operational and provides the responsible editor, the Office for Official Publications of the European Communities, with permanent capacity for checking the status of each of the linguistic versions that are being prepared for publication.

3:45 pm - SGML Transformations -- the IDM Approach

Sean McGrath, Technical Director, Digitome Ltd.
This presentation will introduce an integrated SGML transformation architecture called IDM - Intelligent Document Manager. The philosophical approach to SGML transformation taken by IDM is similar to the STTP of DSSSL. This presentation will show how this approach simplifies SGML transformation programming, allows transformations to be reused in different transformation projects, simplifies project management, and increases the quality of the production process.

4:30 pm - Parsing In A New Key: On Beyond Annex H

Ron Turner, Co-Owner, Soph-Ware Associates
It is a curious irony that a portion of the SGML standard as arcane as Annex H addresses some of the most practical issues among SGML systems: machine resources, possible intractability, visual context checking, and overall simplicity. Model group notation, as the Annex says, resembles the regular expression notation of automata theory. This presentation suggests Petri Nets as a modeling and design framework for such complex parsing tasks. Petri Nets are not simply another clever notation. They are mathematically equivalent to FSAs, thus assuring a true mapping of SGML content models. Some visual comparisons between content models and their corresponding FSAs and Petri Nets offer a striking argument to consider Petri Nets as viable contenders in SGML systems design.

7:00 pm - International SGML Users' Group

Tuesday, December 5, 1995

TRACK 1: Newcomer

8:30 am - If SGML Is So Smart, How Come It Ain't Rich?

Chet Ensign, Director of Electronic Documentation, Logical Design Solutions
No one will deny that obstacles and roadblocks are encountered when trying to adopt SGML. Why won't writers give up WYSIWYG word processors? Why can't the corporate communications department and the print publication department talk to each other? Why doesn't the President see that there is a problem here? This presentation will outline these issues from the top of the organization to the bottom and back again. It will conclude with strategies for identifying the issues that will be important at your organization and suggest ways to get around them.

9:15 am - From Word Processor to SGML to HTML and Beyond: Processing Text for On-Line Databases

Maureen Prettyman, Computer Specialist, National Library of Medicine
The National Library of Medicine uses an object-oriented search and retrieval system that relies on SGML-encoded data to present to the searcher the contents of a document and context of the retrieved data. Although retrofitting paper-published documents into an SGML-dependent system is not a trivial task, it has proven to be well worth the effort as the data are readily prepared for publication electronically through a variety of systems and across multiple platforms. This presentation will detail the advantages that the Library has seen since choosing to implement SGML.

9:35 am - Financial Information Database -- A Case Study of SGML in the Financial Sector

Russell Vann, Consultant, Chartered Association of Certified Accountants
This case study will show how and why the Association of Certified Accountants (A.C.C.A.) converted its documents to SGML and the tangible financial benefits obtained by investing in SGML. It will discuss how SGML has allowed flexibility in marketing the A.C.C.A.'s product to a wider audience. Further, it will answer the following: why SGML is the answer to publishing financial information; how investment in SGML "future-proofs" documentation; why there is an apparent lack of knowledge about SGML within the financial sector; and whether there can be a consensus on a set of standard DTDs.

10:00 am - Break

10:30 am - Producing Presentation-Oriented Technical Manuals from Content-Oriented Information Models

Ken Holman, Chief Technology Officer, Microstar Software Ltd.
This presentation is a case history describing the technical manual production system of a supplier of military equipment to a Canadian Department of National Defense (DND) Project Office. The nature of the structure of the information model is content-oriented, based on the physical equipment breakdown structure (EBS). The nature of the structure of the technical manuals is presentation-oriented, based on a book paradigm. The system architecture and philosophy of operation are described.

11:15 am - Standardized Coding for the Publishing Industry in Norway

Erland Overby and Petter Thorsrud, Center for Information Technology Services, University of Oslo
NOR/MARK is a markup standard for the publishing industry in Norway based on "SGML compliance not SGML dependency". This presentation will discuss the background of why tools were selected and the benefits that they add to further developments of conversions to NOR/MARK or other DTDs, and from other stylesheets than the NOR/MARK-specific DTD. Future plans and hopes on the effect of new editing, conversion and layout/pagination tools enhancements and specialized versions of NOR/MARK '96 translation will also be discussed.

12:00 pm - Lunch

1:00 pm - The World's Cheapest SGML Database Management System

Brian Travis, President, Information Architects
There are advantages and disadvantages to each of the various approaches to SGML database design: relational, full-text, object-oriented, and object-relational. There are places where each may be appropriate.

1:45 pm - Since When Did SGML Become So Sexy? -- An Editorial View

Carol Ellerbach, Senior Database Developer and
Bill Trippe, Lead Database Developer, Inso Corporation

It is SGML's ingenious distinction between form and processing that provides the scaffolding for Inso Corporation's entire electronic publishing operation. The presentation will outline the benefits that have been seen at every stage of processing. The fact that SGML is independent from any particular processing requirements has repeatedly worked to their advantage in unexpected ways. These will be highlighted.

2:30 pm - Technology Showcase

TRACK 2: User

8:30 am - One Application - Five Ways, Pinnacles

Bob Yencha, Sr. SGML Systems Analyst, National Semiconductor;
Tom Jeffery, Principal Technical Writer, Hitachi;
Alfred Elkerbout, Manager Product Information Development, Philips Semiconductors;
Melanie Yunk, Program Manager, Strategic Publishing Processes, Intel Corporation;
Jeff Barton, Manager, Information Systems and Application Support, Semiconductor Marketing Communications, Texas Instruments;
Tommie Usdin, Vice President, ATLIS Consulting Group

The Pinnacles Component Information Standard is an SGML application standard for the interchange of technical information about electric components such as semiconductor products. The standard specifies how the SGML data for interchange is to be organized and tagged, and expects that variations on the published standard will be used for data creation and by display applications. Representatives of five organizations that are creating PCIS documents will discuss their approaches including SGML in the production cycle of their technical documentation, conversion of existing documents to PCIS form, development of their PCIS Authoring-set DTDs, tool selection and customization, and user training.

10:00 am - Break

10:30 am - Database Evolution: The View From over Here (A Document-Centric Perspective)

Sebastian Holst, President, Texcel, USA
This presentation will position the recent and classic DBMS technologies (relational, object-oriented, object-relational, extended relational, etc.) in light of their relative merits in supporting document database information. Specific topics to be discussed will be document and database management evolution and the relationships/dependencies between the two technologies, the core competencies that have arisen out of each of these disciplines and the importance of integration, the role of standards such as SGML, SQL, CORBA and others in building broad information management solutions, and pitfalls and opportunities that have cropped up as technology and commercial products rapidly unfold.

11:15 am - Component Based Information Development

Mark Baker, Manager, Information Engineering Methods, Meridian Information Products, Nortel
Nortel has developed a Component Based Information Development system (CBID) to move away from using SGML to manage their authoring. CBID is a method for capturing and managing information as a collection of generic components which can later be selected and assembled to create specific information about products in any format desired. The presentation will also discuss the two roles that SGML plays in CBID. Complex components are tagged in SGML, and the database queries produce an SGML instance which is processed to produce the desired form of output.

12:00 pm - Lunch

1:00 pm - The SRC Document Processing System -- "make"-ing SGML Work

Mark Shelman, Manager, Electronic Document Delivery, Semiconductor Research Corporation, and
Robert B. Warren, Ph.D., Computer Systems Consultant, Computing Works, Inc.

In converting the results of the SRCs research efforts from exclusively paper form to electronic distribution (including electronic books on CD-ROM to multiple platforms), SRC developed a document processing system to deal specifically with the issues that arose. This presentation discusses these issues.

1:45 pm - Using SGML and Document Management to Share and Reuse Information

Mike Maziarka, Xyvision, Inc.
Since its inception, proponents of SGML have heralded "data reuse" as a key benefit of using SGML. But prior to the widespread use of document management systems, this statement was more visionary than reality. The reason being that many still used file mechanisms for storing data. As a result, information reuse was minimal. Now, with the introduction of document management systems which support storing information at a much finer granular level, document creators and users are able to use SGML as a mechanism for defining the storage units for their data. This presentation will give the attendees an appreciation for the issues associated with reusing information within the scope of a document management system. It will also enable them to explore the use of SGML modeling constructs to achieve their desired results.

2:30 pm - Technology Showcase

TRACK 3: Expert

8:30 am - A Tables Manifesto

Marcy Thompson, Manager, Consulting and Training, Passage Systems
SGML should be used to enrich data with whatever is most important about that data. Typically, this is structural or content-related information, except when the presentation of the content is "table-like". For "table-like" content, markup typically enriches the content with presentation information. This presenter does not believe that what is important about 95% of "table-like" information is what it looks like. This presentation explains why tables may not be a useful information category and discusses the pitfalls of the currently popular approach of "Well, let's just use CALS tables.". The talk proposes alternate approaches to capturing data that may ultimately be presented as a table.

9:15 am - Using Document-Oriented Interfaces for Revision Control and Object-Oriented Databases

Jean Paoli, Technical Director, Grif S.A.
Multiple strategies for handling revision control in SGML can be envisioned: the "a-posteriori" technique involves the comparison of newer and older versions of a document, following modification. The "a-priori" technique, by continuously tracking the modifications made to a document, makes it possible to identify with a very fine degree of granularity, each successive modification made to a document. This presentation examines in detail the latter "a-priori" strategy and studies its implications in terms of the User Interface. The presentation will detail Document-Oriented Interface techniques for generating revision control information in two cases: a basic case where different versions are generated and stored within a single SGML document and a more advanced case where the versioning function of an object-oriented database wired to an SGML Editor is used to store the revision control information.

10:00 am - Break

10:30 am - Object-Oriented Development of SGML Applications

Bob Ducharme, Consultant
This presentation will discuss the similarities and mutually-exclusive characteristics of SGML and object-oriented systems. It will examine the common work of Dr. Charles Goldfarb, Grady Booch, and Charles Rumbaugh as well as potential mutual benefits.

11:15 am - The SGML Link Feature: Why I Demand It

Eliot Kimber, SGML Architect and HyTime Analyst, Passage Systems
This presentation discusses the reasons why the SGML Link feature is not only a good idea but a key aspect of SGML's protection of the rights of data owners from the extortive tendencies of tool vendors. In addition, the Link feature can make traditional SGML products more efficient. Mr. Kimber gives an overview of link mechanisms and presents a proposed convention for integrating link rules with typical style-based processors such as editors and browsers. The presentation will show how the use of Link can simplify style-based processors by providing early binding of some style choices, thereby making product-specific style sheets simpler and runtime processing more efficient. Finally, it will discuss the relationship between Link and DSSSL and explains why the two mechanisms are not in conflict.

12:00 pm - Lunch

1:00 pm - How to Get the Most From Your Document Management System

Brad Chang, General Manager, Publishing Line of Business, Xsoft
This presentation will differentiate information creators, transformers and producers within an organization and show how they will use SGML tools. It will evaluate existing business systems and how they relate to the SGML system. Mr. Chang will identify training requirements, define implementation processes and focus on steps to ensure that your goals will be met.

1:45 pm - Improved Database Technology for Document Management

Albert D'Andrea, Vice President, UniSQL
This session will discuss how a new genre of object-related database technology can provide far improved data management support for SGML-based document management systems and SGML publishing applications. The general issues and technology that address today's complex document and multimedia data modeling will be included.

2:30 pm - Technology Showcase

Evening Sessions

TRACK 1: Newcomer

7:00 pm - SGML Q & A

Clive Carpi and Michelle Butler, InfoDesign Corporation
This informal session will use the most frequently asked questions about SGML as a point of departure for a free-wheeling discussion.

TRACK 1: Newcomer

TRACK 2: User

7:00 pm - SGML Free for All

Chair: Yuri Rubinsky
An informal session intended to draw out rant, cant and argument, this 90 minutes is geared to those people for whom the basic questions and answers are too basic -- but they still feel there are questions that need addressing. A loose gaggle of experts will be lured into the room to provide both targets and anchors.

TRACK 2: User

7:00 pm - Public SGML Applications

Elaine Brennan and Tony Graham, ATLIS Consulting Group
Public SGML applications have been developed for a variety of industries. The session will discuss what a public application contains and when and why someone might use one.

TRACK 3: Expert

7:00 pm - Making SGML Better and Better

Chair: Pam Gennusa, President, Database Publishing Systems Inc.

What's Wrong With SGML

David Silverman, Chief Scientist, Innodata
SGML has revolutionized the way in which we view text information. However, some of its core assumptions have slowed its acceptance and usefulness. This presentation will describe ways in which these core assumptions can be turned on their head, such as allowing DTD validation to be turned on and off in a document including relevant formatting and display information in documents to make SGML even more useful and harmonize it with the other evolving standards.

SGML 2000

Michael Leventhal, Vice President, Text Science, Inc.
This presentation will discuss where SGML should be by the year 2000. It will include three different points of view: revision issues from those involved in the review process, revisions from those that are not involved in the process and from those that feel that there is something better than SGML.

Wednesday, December 6, 1995

TRACK 1: Newcomer

TRACK 2: User

9:00 am - ABC's of DSSSL, An Introduction to the Concepts

Sharon Adler and Anders Berglund, Electronic Book Technologies with
Steven Newcomb, President, TechnoTeacher and
Dr. Charles F. Goldfarb

Document Style Semantics and Specification Language (DSSSL) is the newest international standard in the family of SGML-related standards. Sharon and Anders will cover the introductory concepts including the transformation language and query. The character model will also be discussed. Areas of commonality between DSSSL and HyTime will also be covered.

10:30 am - Poster Session

TRACK 3: Expert

9:00 am - DSSSL for Developers: Advanced Topics in DSSSL

James Clark, Consultant
This talk will discuss some of the aspects of DSSSL that are hard to understand just from reading the DSSSL standard. Topics in both the transformation language and style language will be covered. This is intended for those who need to acquire a detailed knowledge of DSSSL. A basic knowledge of the DSSSL standard will be assumed.

10:30 am - Poster Session

12:00 pm - Lunch

TRACK 1: Newcomer

1:00 pm - They Said It Couldn't Be Done

Mark Gross, President, Data Conversion Laboratory
This presentation will warn against assuming too much when you convert legacy documents to SGML. Because this is a new field, there might be cost-effective conversion options that you have never heard of, and data that you thought was impossible to convert might be convertible after all. This presentation discusses a few composite case-studies that incorporate such troublesome issues as tables and cross references. It will also go over some general principles about how to analyze your legacy documents and then how to analyze what you need from your target documents in order to determine the best way to get there.

1:45 pm - Practical Issues Facing a Conversion Project

Bryan K. Caporlette, Manager, Applications Engineering, Passage Systems
Many organizations which have adopted SGML have had to make a difficult decision: switch to an SGML authoring system or filter their information from a WYSIWYG authoring tool into a target document type definition (DTD). This presentation will address the issues facing an organization that has selected the latter solution. We will address the selection of an authoring tool, how to develop a template geared for conversion into a target DTD, issues facing developers writing a conversion environment to perform the translation, and problems faced by the authors when running the conversion filter. The presentation will provide practical examples of pitfalls for conversion, and provide a frank representation of what it takes to develop a robust, high-quality conversion environment.

2:30 pm - Break

3:00 pm - Legacy Conversion, Tipniques & Pratfalls

Dale Waldt, Director of Publishing, Research Institute of America
This presentation will provide tips, techniques and recommended practices for addressing information conversion. The discussion will include a description of tools and their strengths and weaknesses, resources, project planning, and a few tricks to make survival possible. Several practical case studies will be used for illustration.

3:45 pm - Everything You Ever Wanted to Know about Selling SGML to the Boss but Were Afraid to Ask

Steve Anderson, Project Manager, TiMS Project, Rover Group Limited
Based on a case study of introducing an SGML solution into a major British manufacturer, this presentation will discuss the development of a "Project Plan", describing from inception to delivery and beyond the process of selling SGML through internal channels to a large company. A sophisticated SGML-based Technical Information Management System (TiMS) integrates all of the elements of a complete SGML solution. The need to market the solution, to justify the rationale for this culture change, is one of the key aspects of the internal selling process. This presentation will also recommend a structured process for addressing the culture change, aligned with the development and implementation activities more normally associated with a systems development project.

4:30 pm - Technology Showcase

TRACK 2: User

1:00 pm - Top 10 Common Pitfalls in SGML Conversion

Richard Barth, Director of Operations, Data Conversion Laboratory
Based on numerous actual conversions, this presentation emphasizes the various technical and organizational problems that must be anticipated as part of a conversion plan.

1:45 pm - Automated Production of Truly Customized, One-Time Documents (When Parameter Entities Aren't Enough)

Cynthia N. Lee, Principal, Emerging Markets, Inc. and
William A. Barth, Principal, Strategic Visions, Inc.

This presentation will discuss the use of strategies beyond parameter entities to produce a customized document. It will outline what happens when the document production is many orders of magnitude more complex, including and excluding text based upon a complex set of heuristics and calculations. Examples from real-world problems and solutions will be highlighted.

2:30 pm - Break

3:00 pm - Level, Plumb and Square: Techniques for Designing and Building Major Content Processing Environments

Sam Hunting, Information Architect, Martin Hensel Corporation
Large, dynamic publishing and electronic distribution environments have great needs for efficient content processing yet have DTD requirements that are open-ended. This presentation will describe how the Martin Hensel Corp. has designed and implemented systems in which sophisticated DTDs are easily produced from architectural form-based DTD fragments. The advantages of this approach will be highlighted.

3:45 pm - Xerox Service Manuals are Smart Documents -- A Case Study

Mark Harmison, Technical Specialist/Project Manager, Xerox Multinational Customer and Service Education
For the past 18 months Xerox has been implementing SGML-based delivery of electronic documents. The development efforts have focused on using SGML to enable "smart documents." This presentation will discuss the following key areas of implementation: design and development of a DTD-independent, multimedia viewer, tailored for procedural documents, the development process used, technologies employed, issues faced, and results of implementation in the field. This presentation will include live demonstrations of the technologies and documents described.

4:30 pm - Technology Showcase

TRACK 3: Expert

1:00 pm - New Applications of SGML Technology: The Collaborative Environment for Language Learning

Michael Leventhal, Vice President, Text Science
The Collaborative Environment for Language Learning (CELL), a completely new application of SGML technology, is used for the text-centered multimedia study of foreign languages. SGML plays a critical role in permitting the language- and application-independent definition of a language-learning database.

1:45 pm - Creating DTDs Via the GB-Engine and Fred

Keith Shafer, Senior Research Scientist, On-Line Computer Library Center
This presentation describes the motivation and tools that have been built to automatically create reduced structural representations of tagged text. These tools are novel in that they let one use the basic tenets of SGML without creating DTDs by hand. The presentation also describes the process used to go from specific document instances to generalized grammars (DTDs).

2:30 pm - Break

3:00 pm - Visual SGML: Linking SP to Fresco

Tim Prinzing, Senior Software Engineer, Fujitsu Network Transmission Systems
Fresco is a set of publicly-available remotable interfaces and a user interface toolkit conforming to CORBA. Platform- and resolution-independent, it supports structured graphics and direct manipulation of the user interfaces. The work represents a newly-available way to have SGML objects that users can interact with; for example, to create radically different views of the same document.

3:45 pm - Technology Showcase

Evening Sessions

TRACK 2: User

7:00 pm - SGML and the World Wide Web

All the Things that are Wrong with HTML, and Why They Don't Matter

Tim Bray, Vice President, Technology, OpenText
HyperText Markup Language, one of the bases of the World Wide Web, is a strange beast indeed. It is a formatting language used for screen presentation, but a poor one, lacking many of the features we have come to expect in such systems. The main thrust of this talk will be a defense, based on economics and illustrated by further examples from the Web Index project, of the following strong proposition: "Any person or organization planning any significant project involving the delivery of on-line documents must first prove that HTML cannot be used before considering other alternatives."

ISO 12083 Web Interface

Eric van Herwijnen, Founder, NICE Technologies
EPSIG, in conjunction with NICE Technologies, has laid the groundwork for a way to add structure to the World Wide Web. The Web is based on HTML, a specific SGML DTD; however, its capacity for management and display of large documents is virtually nonexistent. The publishers, editors, and authors who already utilize ISO 12083 in their databases can easily transfer their book length documents over the Internet, but the time involved in downloading is extraordinary, reading a book on a computer screen is unacceptable, and search capabilities are limited. The structure of SGML, particularly that found in ISO 12083, is useless today on the Internet. This presentation will discuss an approach which will allow clients on the Web to access the structure of an SGML database for search and retrieval, and then to present the information requested in a user-friendly environment.

TRACK 3: Expert

7:00 pm - Up to the Minute from the HTML Working Group

The Design Rationale for HTML3

Dave Raggett, World Wide Web Consortium
This paper will review the design rationale for the next generation of HTML, and describe how HTML3 meets the needs for a lingua franca for the World Wide Web. Proprietary extensions to HTML have focused on stylistic attributes, leading to worries about the long-term versatility of HTML. The HTML3 specification has been carefully designed to exploit associated style sheets and downloadable fonts. This, in association with mobile code, will greatly expand the opportunities for publishing over the web. This presentation will explain the relationship of the HTML table model to CALS, and how HTML has been extended to handle math.

How SGML Fits into the World Wide Web

Eric Severson, Vice President, Market Strategy and Corporate Development, Interleaf
Is HTML a blessing or a threat to SGML? If you have SGML, do you still need HTML? Or should the question be asked in the opposite order? Technically HTML is an SGML application, but the real issue will be decided by what people will accept and the true nature of their requirements. Using humor and analogy to explain the underlying technical issues, but then going on to explore the real implications of these technologies, this presentation provides important answers from both a developer's and manager's perspective. It will show how both HTML and SGML by themselves fall short of the real need, then go on to argue for a scaleable Web in which both formats coexist. In this vision, HTML will continue to be the primary data format for the Web, but generalized to form the backbone within which more sophisticated SGML applications will also fit.

Thursday, December 7, 1995

TRACK 1: Newcomer

9:00 am - Traversing Smart SGML in an IETM Environment

Neil Montgomery, Information Engineer and
Robert Fye, Vice President, General Manager, Aquidneck Management Associates, Ltd.

Aquidneck has been involved with supporting the US Navy's NAVAIR on legacy F/A-18 data migration and IETM development in accordance with the DoD Tri-service IETM specifications as part of a Phase II SBIR. An IETM Validation, traversement and presentation engine to traverse and present "smart" or "logic-enhanced" SGML that has been migrated from legacy data and structured in accordance with the IETM Content Data Model (CDM) generic and content-specific layers will be described in detail.

9:45 am - Developing Interactive Multimedia Training Applications for Use in Conjunction with SGML-Tagged IETMs

Michael Gallagher, Director, Commercial Sales and Marketing, Advanced Engineering and Research Associates, Inc.
This presentation will show multimedia computer-based training with interactive software applications which were developed to be used with interactive electronic technical manuals (IETMs). The demonstration will discuss ways by which technical documentation that has been represented electronically in SGML format may be enhanced by adding sound, video and animation in addition to state of the art CBT. The presentation will cover four areas: technical manual conversion and SGML tagged IETMs, CBT/ICW planning and development, the integration of the two applications into one cohesive training application, and issues relating to CD-ROM publishing.

11:00 am - Break

TRACK 2: User

9:00 am - Typesetting Direct from World Wide Web Documents

Peter Flynn, Consultant, Silmaril Consultants
Documents in the World Wide Web are intended primarily for on-screen display. The HyperText Markup Language used is limited in power, but still underutilized by implementors, and often written in a non-conferment manner. This paper describes the development of a tool to produce typeset (PostScript) output direct from valid, conferment instances of HTML. While it handles some invalid but well-formed tags, the objective is to demonstrate the power of making the markup work for the user as well as to provide a useful means of improving the quality of printed output.

10:00 am - Can and Should HTML be Used for Document Structure?

Tim Bray, Vice President of Technology, OpenText Corporation
HTML is an application of SGML. SGML is normally seen as a tool for "descriptive" markup, aimed at structure rather than presentation; the implication is that HTML can be used in this way. In general, it is not. The data-structuring capabilities of HTML are not only limited, they are often ignored or even seen as obstacles by HTML authors, who are overwhelmingly concerned with the visual appeal of their creations. Thus it is reasonable to ask: is there any need for document structures in the context of the WWW? This talk will introduce the data structuring capabilities of HTML and provide a frank discussion of their limitations in theory and in actual usage. Mr. Bray will demonstrate the importance of structural information in successful information retrieval and describe alternative strategies for working with, around, and against the defacto state of the WWW to gain some of the benefits of document structure.

11:00 am - Break

TRACK 1: Newcomer

TRACK 2: User

11:30 am - Semantic or Structural - What is the Best Industry-Wide Exchange Model?

Carla Corkern, President, Highland Consulting
This presentation discusses two ways of looking at industry-wide information modeling. It suggests that there are good reasons why each industry makes its own particular choices.

12:15 pm - Get off the Wave and Start Climbing the Ladder

Paul Jensen, VP, Chief Information Officer, Research Institute of America
A suggestion for keeping up with constant changes and improvements in new technology. Instead of "riding the wave" of new technology through its life and then catching the next wave, throwing out much of the old system in the mean time, let's recognize that it is time to view new technologies as steps on a ladder. New design approaches like object orientation and open standards like SGML provide organizations with a way to extend the life of their technology investments. A financial model will be presented to help explain and sell this approach to management. Techniques will be reviewed that can help techies to design and implement a system that will help them get to the next step up the ladder.

1:00 pm - Lunch

TRACK 3: Expert

9:00 am - SGML to SGML Transformations A Comparison of Four Techniques

Jeff Suttor, SGML Architect and Software Engineer, Passage Systems

The need to perform SGML to SGML transformations is growing. This is the result of a maturing SGML marketplace and diverse toolset that supports native SGML. Just several years ago, the major questions were how to get non-SGML information into SGML (up-transformation), and out of SGML (down-transformation). Our community is now dealing with the transformation of large bodies of richly marked up SGML to interchange DTDs, application specific DTDs and simplified DTDs in support of simple or minimally SGML-aware applications. This presentation will discuss and compare four techniques for transforming SGML: custom programming, architectural forms, link process definitions and DSSSL STTP. The transformation of HTML to ICADD will be used as the example SGML source and target DTDs. All examples will be done using two common SGML parsing tools.

9:45 am - A Theory of Documents: How SGML Can Change the Way We Think about Information

R. Alexander Milowski, President/Principle Researcher, Copernican Solutions Incorporated
This presentation will explore the representation of traditional application and storage formats as documents. In addition, it will set the stage for a theory of documents as human-based information representation and posit the argument that document-centered solutions are inevitably more productive and protective of the information that resides within.

10:10 am - Using HTML as the Basis for Information Sharing in Heterogeneous Computing Environments

Herman Kuiper, Numerical Mathematics and Application Programming Department at the National Aerospace Laboratory of the Netherlands and
H. van der Ven, Numerical Mathematics and Application Programming Department at the National Aerospace Laboratory of the Netherlands.

This presentation will describe how the HTML can be used to provide information within a heterogeneous computing environment. The different kinds of information that can be provided as well as an introduction to a more interactive use of HTML through forms is given. During the development of the information system a number of valuable lessons were learned and will be shared.

11:00 am - Break

11:30 am - Writing SGML Declarations

Wayne Wohler, SGML Product Development Manager, IBM
This session will begin with a description of SGML declarations and their description of document character sets, syntax character sets and the relationship of the two. It will also cover the SGML system model, some problems with it and some ideas for improving this model to support double byte character sets. We will discuss the Extended Reference Concrete Syntax to cover East Asian alphabets.

12:15 pm - SGML and HTML: Learning From Each Other

Murray Maloney, Product Manager, SoftQuad

Over the last two or three years, there has been an interesting and dynamic relationship emerging between and amongst the purveyors of SGML and HTML. Driven by a discerning market, SGML and HTML technology has been making rapid advances. Mr. Maloney will review some recent learning experiences and examine current and future SGML and HTML issues.

1:00 pm - Lunch

2:00 pm - Closing Keynote and Conference Wrap-Up

Michael Sperberg-McQueen, Editor, Text Encoding Initiative and Professor, University of Illinois at Chicago

Return to the SGML '95 Home Page