%M C.HYSTAN.90.1
%T Proceedings of the Hypertext Standardization Workshop
%A Moline, Judi
%A Benigni, Dan
%A Baronas, Jean
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 1
%K Hypermedia, Hypertext, Standards
%X This report constitutes the proceedings of a three day workshop on Hypertext
Standardization held at the National Institute of Standards and Technology
(NIST) on January 16-18, 1990.  Efforts towards standardization of hypertext
have already been initiated in various interested organizations.  In
recognition of these existing efforts, NIST sponsored the Hypertext
Standardization Workshop organized by the Hypertext Competence Project of the
National Computer Systems Laboratory.
   The major purpose of the Hypertext Standardization Workshop was to provide a
forum for presentation and discussion of existing and proposed approaches to
hypertext standardization.  The stated workshop goals were to consider
hypertext system definitions, to identify viable approaches for pursuing
standards, to seek commonality among alternatives whenever possible, and to
make progress towards a coordinated plan for standards development, i.e. a
hypertext reference model.  The workshop announcement solicited contributed
papers on any aspect of hypertext standardization, including assertions that
standardization is premature or inadvisable.  Approximately 30 contributions
were received and distributed to the 65 workshop participants on the first day.
   The workshop included plenary sessions and three discussion groups.  This
proceedings includes the papers selected for presentation in plenary sessions,
reports of the discussion groups, and supplementary materials.  Major
conclusions of the workshop were that the discussion groups should continue
their technical efforts, and that NIST should sponsor at least one more
workshop to provide a forum for public discussion of progress.

%M C.HYSTAN.90.3
%T Introduction
%A Gallagher, Leonard
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 3-4

%M C.HYSTAN.90.9
%T Reference and Data Model Group (RDMG): Work Plan Status
%S Hypertext Models Discussion Group
%A Parunak, H. Van Dyke
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 9-13
%X A reference model is a structured description of some domain that can be
used to compare existing implementations in that domain, design new
implementations, and (most important for our purposes) map out possible areas
for standardization and show their relation to one another.  The main output of
the RDMG during the NIST workshop was a work plan for arriving at such a
reference model.  The work plan that we propose has the following structure,
where the flow of activity is down the page (except for the single feedback
loop), and where activities marked by '*' received significant attention during
the workshop. <pre>
      +---------------+-----------------+-------------+
      |               |                 |             |
      v               v                 v             |
   *Define       *Brainstorm    *Compare Existing     |
   "Hypertext"   Concepts       Models (DTL)          |
               \      |       /                       |
                \     v      /                        |
             *Organize Ontology                       |
                      |                               |
                      v                               |
           Rank Concepts by Centrality                |
                      |                               |
                      v                               |
           Inventory Existing Systems                 |
                      |                               |
                      v                               |
        Construct "Implementation" Model              |
                      |                               |
                      +-------------------------------+
                      v
           Select Areas for Standards </pre>
   The rest of this document defines each of these steps, and reports what we
have done in each of them.
   This document summarizes the portion of the final RDMG presentation that I
delivered on 18 January 1990.  It represents my perception of the deliberations
of the group, but has not been reviewed or formally approved by the other
members.

%M C.HYSTAN.90.15
%T Reference and Data Model Group: Comparison of Three Models
%S Hypertext Models Discussion Group
%A Leggett, John J.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 15
%X The Reference and Data Model working group spent 45 minutes comparing and
contrasting the R-model, Dexter and Lange reference models.  David Stotts,
Danny Lange and John Leggett spent another 90 minutes over dinner discussing
the three models.  A summary was provided by John Leggett during the final
plenary session.  As these three models are currently under development, the
comparisons are rather broad in nature.  It is interesting to note that the
three models were developed independently and with varying levels of
collaboration.  The results of these discussions are presented below in mostly
tabular form.

%M C.HYSTAN.90.17
%T Hypertext Reference Model Group: Responses to "Issues for Discussion Group
Consideration"
%S Hypertext Models Discussion Group
%A Black, James
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 17-18

%M C.HYSTAN.90.21
%T Summary of the Hypertext Interchange Group
%S Data Interchange Discussion Group
%A Oren, Tim
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 21-22

%M C.HYSTAN.90.23
%T Note on Representing Anchors
%S Data Interchange Discussion Group
%A Oren, Tim
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 23-25
%X An ad hoc subgroup of the Interchange working group met to compare various
proposals for archival interchange.  It was composed of Ed Fox, Steve Newcomb,
Tim Oren, and Victor Riley.  These notes are the result of that meeting.  They
are a first pass which has not been considered by any other group.  See the
summary of the Interchange group for context and definition of terms.
   We chose to proceed by focusing on the anchor or "anchor-like" portion of
each proposal.  We began by considering how the features of the Intermedia
Interchange could be added to the HIP proposal, and expressed the result in
HIP-like terms.  We then attempted to reconcile this result with the formalism
and language of the pertinent sections of HyTime.  Note that this applies only
to anchors, and there may be additional difficulties in reconciling layering
strategies when we look at the link layers of the various proposals.

%M C.HYSTAN.90.29
%T Report from the User Requirements Working Group
%S User Requirements Discussion Group
%A Glushko, Robert J.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 29-35
%X This report summarizes meetings held on January 16-17, 1990 during a
workshop on Hypermedia Standardization held at the National Institute of
Standards and Technology in Gaithersburg, MD.  In addition to the author, the
members of the Working Group for User Requirements were Carol Adams, Peter
Aiken, Jean Baronas, Denise Bedford, Tim Berners-Lee, Valerie Florence, Kevin
Gamble, Louis Gomez, Seymour Hanfling, Kathryn Malcolm, Cathy Marshall,
Fontaine Moore, Dan Olson, Duane Stone, Clifford Uhr, David Wojick and Don
Young.  The group followed an agenda set by NIST to identify the current state
of affairs, important driving and constraining factors, potential areas for
standardization, and research needs.
   Complete consensus on these complex topics was impossible in two days for a
group this size, so this report emphasizes the majority themes for the issues
that received the most attention.  I apologize for my own biases, which
undoubtedly show through.

%M C.HYSTAN.90.39
%T Hypertext Interchange Format -- Discussion and Format Specification -- Draft
1.3.4
%S Papers
%A Bornstein, Jeremy
%A Riley, Victor
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 39-47
%X The Hypertext interchange format described here is based on the work of the
Dexter group, an industry coalition of hypertext researchers interested in a
standard for hypertext data exchange.  This paper describes the result of a
collaboration towards this end between Jeremy Bornstein and Frank Halasz, with
significant input from other members of the Dexter group, most notably Tim
Oren.  The work took place during the summer of 1989, and a demonstration is
planned for the Hypertext '89 conference in November of 1989.

%M C.HYSTAN.90.49
%T Standards for Hypertext Source Files: The Experience of UNIX Guide
%S Papers
%A Brown, P. J.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 49-58

%M C.HYSTAN.90.59
%T Standards: What Can Hypertext Learn from Paper Documents?
%S Papers
%A Cole, Fred
%A Brown, Heather
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 59-70
%X Hypertext literature tends understandably to concentrate on what is new and
to ignore, or take for granted, the properties of hypertext that are also
present in paper documents.  The purpose of this paper is to consider how the
expertise that exists in standards and models for paper documents can be used
to save effort when designing a standard for hypertext, and how to make
hypertext and paper document standards compatible.  Section 2 discusses some
relevant similarities between paper and hypertext documents.  Section 3
introduces relevant aspects of the Office Document Architecture (ODA) [1] and
suggests ways to build on ODA to create a standard that combines the strengths
of the two areas.

%M C.HYSTAN.90.71
%T Standards for a Hypermedia Database: Diachronic vs. Synchronic Concerns
%S Papers
%A Crane, Gregory
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 71-81
%X This paper outlines the perspectives of a professor in one traditional
branch of the humanities (Classics).  My colleagues and I are engaged in
creating a hypermedia database on ancient Greek civilization, but our work is
intended to explore the generic issues of building a complex hypermedia
database, and Perseus was conceived as a model for what should (and no doubt
should not) be done.  We have encountered a number of problems along the way
that must be solved before information disseminated in a hypermedia environment
can have more than marginal impact on intellectual activity.  This paper
addresses hypermedia databases: although much of our work revolves around texts
and still images, we can see that sound, animation, and motion video are also
basic categories of information.  This paper at least views hypertext as a
subset of hypermedia.
   The argument of this paper can be summarized simply.  Standards for
hypermedia must emerge before hypermedia databases can be fully useful, but
long-lived standards can only emerge after we know much more about how people
will use hypermedia databases.  Since we can do qualitatively different things
in a hypermedia environment, we must assume that usage patterns will emerge. 
Practically speaking, we can expect to see short term interchange tools so that
we can move data from one hypertext system to another, but we should be
prepared to abandon these standards if they prove too inflexible.  The rest of
this paper outlines some pragmatic concerns.

%M C.HYSTAN.90.83
%T The Trellis Hypertext Reference Model
%S Papers
%A Furuta, Richard
%A Stotts, P. David
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 83-93
%X We describe a hypertext "meta-model" -- one that provides an organization
for the architecture of a hypertext model.  The specific meta-model presented
was developed in the context of the Trellis hypertext model.  However the
organization seems generally applicable to other models as well.  As such the
meta-model may be a good candidate for a hypertext reference model, and so we
call it the Trellis hypertext reference model.  In this report we first
describe the Trellis hypertext reference model, and then discuss the
relationship of some hypertext-defined concepts to the reference model.

%M C.HYSTAN.90.95
%T The Dexter Hypertext Reference Model
%S Papers
%A Halasz, Frank
%A Schwartz, Mayer
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 95-133
%X This paper presents the Dexter hypertext reference model.  The Dexter model
is an attempt to capture, both formally and informally, the important
abstractions found in a wide range of existing and future hypertext systems. 
The goal of the model is to provide a principled basis for comparing systems as
well as for developing interchange and interoperability standards.  The model
is divided into three layers.  The storage layer describes the network of nodes
and links that is the essence of hypertext.  The runtime layer describes
mechanisms supporting the user's interaction with the hypertext.  The
within-component layer covers the content and structures within hypertext
nodes.  The focus of the model is on the storage layer as well as on the
mechanisms of anchoring and presentation specification that form the interfaces
between the storage layer and the within-component and runtime layers,
respectively.  The model is formalized using Z [19], a specification language
based on set theory.  The paper briefly discusses the issues involved in
comparing the characteristics of existing systems against the model.

%M C.HYSTAN.90.135
%T Standardization of Hypermedia: What's the Point?
%S Papers
%A Hardt-Kornacki, Shoshana L.
%A Gomez, Louis M.
%A Patterson, John F.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 135-144
%X In this paper we present multiple views on the issue of standardization of
Hypermedia systems that operate over a global heterogeneous information
network.  To aid our analysis we introduce a reference model that captures the
information flow and the information control aspects from the viewpoint of the
user.  This model is then used to focus the analysis of Hypermedia systems from
a variety of perspectives, such as overall resources, network communication,
interface building, and application writing.  Based on our analysis we conclude
that at this time, the components of Hypermedia systems that are ready for
standardization are not necessarily Hypermedia-specific.  Moreover, we strongly
believe that the Hypermedia-specific aspects of these systems are not yet ready
for standardization and we question the wisdom of ever standardizing certain
Hypermedia specific components such as the user interface or the navigation
tools.  In addition, we conjecture that it may be desirable to standardize a
generic set of tools that can be used to build these components so as to
guarantee that the access to the information stored in future Hypermedia
systems will not be impaired.

%M C.HYSTAN.90.145
%T A Formal Model of Hypertext
%S Papers
%A Lange, Danny B.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 145-166
%X In this paper a formal specification of an abstract model of hypertext is
presented.  The Vienna Development Method (VDM) is used in this specification.
Experiences with a prototype hypertext system and studies of other existing
hypertext systems are captured in this formal specification.  Basically
datamodel of hypertext is suggested.  In this model three main abstract data
types of hypertext are formally defined: nodes, networks and structures.  The
abstract data types are applied to the concepts of object-oriented databases
and a "hyperbase" is defined.

%M C.HYSTAN.90.167
%T A Multi-Tiered Approach to Hypertext Integration: Negotiating Standards for
a Heterogeneous Application Environment
%S Papers
%A Marshall, Catherine C.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 167-177
%X Hypertext is most useful as a technology when it is embedded in an
application: a paperless technical manual, a notetaker, a specification
management system, or any other task domain where it is useful to represent and
manipulate the structure of text.  We feel that it is important to connect
system requirements for hypertext with the situation of use; thus
standardization efforts should be directed at enhancing the ability to embed
hypertext in heterogeneous applications environments.
   This paper addresses a specific application and task environment -- using
hypertext as a medium for a shared notetaker that will be used in the
intelligence community -- and how it suggests a protocol-driven approach to
integration.  The work described in this paper includes an informal work
practices study of the task environment, and the development of a functional
specification for a hypertext system for notetaking.
   From the study and the development of a specification, we postulate that
standardization of a multi-tiered system of linking protocols will help address
the closed-world problem that we have encountered in NoteCards and many of the
other second-generation hypertext systems without specifying rigid standards
for applications that want to share information to a greater or lesser extent
with a hypertext substrate.  Such a system of protocols can be based in part on
existing work on hypertext exchange and hypertext reference models.
   First we will briefly describe the task environment and present an informal
model of the task.  Then we will go on to describe linking and anchoring
requirements in support of this task.  Finally, we will argue that a
multi-tiered system of linking protocols will not only meet the needs that we
have already identified, but will be adaptable as the environment changes and
will facilitate information sharing.  It is this set of protocols that we
propose should be standardized based on negotiations between applications
developers and the hypertext community.

%M C.HYSTAN.90.179
%T Explanatory Cover Material for Section 7.2 of X3V1.8M/SD-7, Fifth Draft
%S Papers
%A Newcomb, Steven R.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 179-188
%X The mission of the ANSI X3V1.8M Music in Information Processing Standards
(MIPS) committee is to develop a Standard Music Description Language (SMDL) to
enable interchange of musical documents.  The committee has chosen to represent
the structure of the information represented by SMDL as a Standard Generalized
Markup Language (ISO 8879-1986) Document Type Definition (an "SGML DTD").
   In the course of its work (which began in 1986), the MIPS committee
developed a general model for the representation of schedules for the execution
of events.  When it confronted the problem of representing music in several of
its normal contexts, such as the interdependently synchronized lighting,
staging, and orchestra cues in musical comedy and opera, the MIPS committee
developed SGML-based means of representing links within and among documents. 
These means are what is set forth in the following extract (Section 7.2
["General Links"] of the fifth draft of X3V1.8M/SD-7 ["Hypermedia/Time-based
Document Subset"].
   When it became clear that this model would be useful for the representation
of the scheduling of non-musical (as well as musical) events multimedia and
hypermedia documents, the committee extracted the time model from the other,
strictly music-related portions of SMDL, gave the model a name ("HyTime"), and
placed it in its own Standing Document, X3V1.8M/SD-7.  In the current draft of
SMDL, Standard Music Description Language (SMDL) is an application of HyTime. 
(The rest of SMDL is described in X3V1.8M/SD-8.)
   When HyTime's "General Links" facilities were discussed at the NIST
Hypertext Workshop, it turned out that the Dexter, Intermedia, and HyTime
models all decomposed the problem of document addressing in much the some way,
although their jargon was dissimilar.  The "Room 705 Ad Hoc Group" (Ed Fox,
Steve Newcomb, Tim Oren, and Victor Riley) succeeded in showing how the
"anchor" concept in the three models could be merged.  It is anticipated that
the NIST Hypertext Workshop will have significant impact on succeeding drafts
of HyTime.

%M C.HYSTAN.90.189
%T Toward Open Hypertext: Requirements for Distributed Hypermedia Standards
%S Papers
%A Oren, Tim
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 189-196
%X Much discussion of hypertext standards has centered on the transfer of
closed, static hypertext document bases among various platforms and
organizations.  While there is an undoubted need focused on the use of
hypertext with optical media and technical documentation, the thesis of this
position paper is that any standard based primarily on this limited application
will be necessarily flawed.

%M C.HYSTAN.90.197
%T Toward a Reference Model for Hypermedia
%S Papers
%A Parunak, H. Van Dyke
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 197-211
%X A necessary first step in discussing standardization in a domain is the
development of a reference model for that domain, a high-level framework within
which specific topics for discussion can be defined and discussed.  This paper
offers a "straw" version of such a framework as a basis for discussion, and
discusses the "standardizability" of various detailed subjects within that
framework.

%M C.HYSTAN.90.213
%T An Interchange Format for Hypertext Systems: The Intermedia Model
%S Papers
%A Riley, Victor A.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 213-222
%X Realization of the potential for information sharing that is inherent in
hypertext systems depends on the ability to readily exchange data between those
systems.  A format for exchanging link-related data between first-order
hypertext systems has been designed, and partially implemented, for the
Intermedia system.  The design is described to the individual field level.  An
example of usage for Intermedia link-related information is provided.  The
import, export, and verification utilities created for the interchange format
are also described.

%M C.HYSTAN.90.223
%T Strawman Reference Model for Hypermedia Systems
%S Papers
%A Thompson, Craig W.
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 223-246
%X This paper provides a strawman reference model that can be used for
comparing and reasoning about hypertext/hypermedia systems.  It begins with a
glossary of hypermedia terms.  Agreeing on these provides a common vocabulary
for developing the reference model.  The reference model itself is presented in
terms of basic features all hypermedia systems have, advanced features some
hypermedia systems have, and open features that hypermedia systems share with
other computer systems.  These features represent independent dimensions which
can be used to classify or compare existing hypermedia systems and to contrast
them with near-miss related systems.  Based on the features, the architecture
of an ideal hypermedia system is described that covers existing hypermedia
systems.  The architecture is modular.  A consequence is that discussion of
standards or a more detailed reference model can focus on one module at a time,
avoiding movement toward a portmanteau standard.  The final section of the
paper evaluates some areas where consensus and eventual standardization of
hypermedia systems is possible and would be valuable.  An appendix references
some standards related to hypermedia systems.  Another appendix is an initial
document log listing references important to hypermedia standardization.

%M C.HYSTAN.90.249
%T Hypermedia Bibliography
%S Appendices
%A Kahn, Paul
%B NIST Hypertext Standardization Workshop
%D 1990-01-16
%P 249-264
%X Since the last time we compiled this bibliography in November 1987 for the
Hypertext '87 Workshop, there has been an explosion of hypertext literature. 
When we started the bibliography project at IRIS in 1983, we thought it would
be possible to collect every book, conference paper and journal article on the
subject of hypertext.  In 1989, that seems an impossible goal.  We hope our
collection includes a large portion of the current literature, but every day we
learn of new papers that are not part of our collection.
   This version, prepared for distribution by NIST, contains only references to
material we have been able to collect over the past six years.  The reference
list differs substantially from the 1987 version.  In 1987 there just were not
that many papers focused entirely on hypertext, so we included in the
bibliography many papers that, while only tangentially related to the topic of
hypertext, had been influential in helping us think about the subject.  Now
that there are so many papers focused solely on hypertext, we have opted to
narrow the scope of the bibliography and include only those references that are
exactly on the topic.