[alt.hypertext] HYP8803A.BIB Bibliography Addenda

deh0654@sjfc.UUCP (Dennis Hamilton) (03/15/88)

%A Ed McNierney
%T The User at the Controls
%J PC Tech Journal
%V 6
%N 3
%D March, 1988
%P 65-75
%K Windows SAA PS/2 Presentation Manager Standardized
Interface Keystrokes
%X REVIEW: This author fails to note that newly-released
Windows 2.03 incorporates many of the concepts of
OS/2 Presentation Manager and will provide substantial
look-and-feel similarity (though still keeping the
MSDOS-Executive appliance instead of OS/2's alternatives).
  The importance of this information is that there is
going to be great effort to unify window-based application
interfaces as part of IBM's SAA.  Some of the conventions
are familiar to Macintosh and Windows users.  Some of the
new ones, especially those which reduce gear-shifting
(switching between keyboard and mouse) will be important.
  In particular, the standard-interface conventions of
Windows 1.03 have been changed for PM and 2.03.  These
have several beneficial results.  This article only
tabulates the new conventions, in the following organization:
   Standardizing Keystrokes -- reserved combination-key
sequences for PM interception (in Table 1)
   Recommended Keystroke Conventions -- suggested for use
by applications in consistently handling standard application-
implemented actions (in Table 2)
   The conventions for recommended mouse usage (in Table 3) are
pretty thin, and will be adjusted to suit applications anyhow.
   The shortcut key sequences (accelerators) now also being introduced 
with Windows 2.03, including the "sticky" ALT-key, make it much easier 
to operate mouseless except when drawing and editing.
   All of the window framing and controlling operations and
appearance items are the same in Windows 2.03 as in this
description of OS/2 Presentation Manager.  Window sizing,
moving, scroll-bar controls, menu-selection accelerators, menu
expansion, dialog boxes, message boxes, and the system control
panel should be familiar.  The additional OS/2 utilities involve
multi-tasking and different setup and filer applications.
   However you encounter this interface, because of SAA it is
going to be increasingly supported.  It will be very important
to conform to it whenever that is possible as a matter of
arbitrary choice.
   [dh:88-02-27]

%A Michael Brian Bentley
%T A Consistent API
%J PC Tech Journal
%V 6
%N 3
%D March, 1988
%P 78-89
%K API OS/2 SAA GPI Application Services Windows
%X REVIEW: The OS/2 Presentation Manager (and Windows 2.03)
user interfaces present users with a consistent
set of conventions and forms for commonly-needed
interactions with applications in a multiple-window
setting.
  The Application Program Interface is the same thing
but set up for application programs to use in interfacing
with users and data systems in a device-independent way.
This is complicated but beneficial.  At the moment, using
API's (as reflected in struggling to use the Window's
Development systems, for example) is burdensome and
difficult.  It remains to be seen whether the API
conventions can be supplanted (but while serving the
necessary purposes).
  In any case, the OS/2 API uses the same basic scheme
used in Windows (1.03 or 2.03) and covers the following
activities:
  APPLICATION ARCHITECTURE
     organization of programs is seriously impacted by the API interface method
     more about this below (the effect is about as dismal as the OS DL/I interface)
  INTERACTING WITH THE USER
     is actually about the PM "shell" or Windows interface that the user sees
     user is reached via API operations and management of the interface is done by system
     there are also API operations which give application some of the same functions that users have
     basic self-conscious access to shell via API is for task management
  WINDOW MANAGEMENT
     API services provide for creating, sizing, moving, and otherwise controlling windows
     All windows fit in classes and adhere to class-specific styles, with a number of supported window-specific style features
  HANDLING INPUT
     All input is asynchronous and API-using programs are input driven
     message-passing is the chosen method for communicating to applications via API
  HANDLING OUTPUT
     Involves communication via objects for device contexts and presentation spaces
  ARTISTIC SERVICES
     Basically via the Windows GDI (Graphics Device Interface) extended to GPI
     Features world coordinate space, modal space, page space, and device space
     Lots of modal stuff, plus control for clipping, defining regions of space
     Everything can go to a bitmap and there is a metafile capability
   My biggest objection to the application architecture forced by API (and the
Microsoft Windows counterpart) is what Bentley, and many others, take as a
desirable feature:  "This structure imposes another level of centralization
within the application.  All input messages [from PM through API] are sent to
one procedure within the application that owns that window. ... The result is
a much more maintainable program than the traditional sprinkling of I/O calls
and input processing procedures among the computationally intensive sections of
code that form the core of a standard application."  [Standard applications are
the current crop of conventional, non-API ones.]
   While conceding that many applications are badly structured and the effect is
to have input-output interactions inbedded at inappropriate and difficult to
maintain points, I contend that forcing all input to a program, from all sources,
through a single interface point will only make the problem worse.  This means
that every application must be designed with control coupling at its topmost
level, hardly an encouragement to improved structure.
   I suspect that this process will be survived, if it can be, by imposition
of even more "middleware" between applications and API so that applications
can be decomposed into rational, functional portions with maximum decoupling
and asynchronous usability.  The middleware, hopefully based on 
object-oriented technology, will facilitate that and do whatever is necessary 
to bury the single-contact-point-knows-all-handles-all API interface that is 
required by the system deep in the system where no one has to admit to its
questionable existence.
   -- Dennis E. Hamilton
   [dh:88-02-27 because this is going to influence our ability to build and
manage hypertext systems and applications.  It is as serious as having to
know about Xanadu back-end conventions.]

%A Charles Sheffield
%T The Courts of Xanadu
%J Isaac Asimov's Science Fiction Magazine
%V 12
%N 4
%D April, 1988
%P 20-59
%O Whole #129
%K Science Fiction Xanadu
%X "To search for Xanadu.  How can I say no?"
   What does this have to do with hypertext?  Doesn't EVERYTHING
have to do with hypertext??
   [dh:88-03-01]

%A Michael J. Miller
%T Sophisticated Text-Retrieval Program Offers Hot Links, Keyword Outlining
%J InfoWorld
%V 10
%N 9
%D February 29, 1988
%P 52
%O First Look column
%K IZE Memory Mate Ask Sam text retrieval outliner Guideline Keychanger Hot Links Persoft
%X "... Ize can search through a series of documents and find those that
contain a specific word or combination of words.  Ize typically stores
documents in a `textbase,' along with the list of up to 300 keywords
you assign to each document. ..."
   "Rather than just going through and showing [keyword matches] one
by one, Ize displays the matches in an outline form, showing documents
that match the word you are searching for as organized by their other
keywords. ... By moving through your outline, you can retrieve specific
documents. ...
   "Ize gives you several ways to control the outline. ...
   "Keychanger lets you add or delete a specific keyword to a text
or to make certain words automatically marked as keywords as the
text is being saved.  This latter feature is indispensable when you
import files from a word processor [including ASCII, Wordstar, Word
Perfect, Multimate, and DCA formats].
   "Hot LInk feature.  Rather than importing in a document, Ize
creates a special summary file with any associated keywords so it
can find the document when you search for those keywords.  Then you
simply `hot link' to the document and Ize will actually run the
program with it. [I take this to mean it will run a program that
is associated with the particular file.] ... Hot Links doesn't
display the actual file within Ize, only the name of the file you
link to.  [Miller points out that this could involve a communication
access or other procedure for reaching non-resident data.]"
   IBM PC's with at least 512k.  $445, scheduled for release in
March, 1988, by Persoft Inc., 465 Science Drive, Madison, WI 53711;
(608) 273-6000.

%A Frank Bican
%T The WORM Turns
%J PC Magazine
%V 7
%N 6
%D March 29, 1988
%P 199-224
%O Mass-Storage Devices product reviews
%K WORM CD-ROM ISI 525 WC Maximum Storage AP-3000 Micro Design LaserBank 800
Optotech 5984 Symphony Systems N/Hance 525 Optical Disk System
%X Describes current crop of available Write-Once/Read-Many optical disk
drives and subjects them to series of tests.  This evaluation's preference
is for the $2795 ISI 525 WC (122MB) or the $2950 N/Hance 525 (120MB) using 
the same drive but with excellent TextScan text-retrieval software package.  
The $2950 Optotech 5984 (200MB) and the $2675 APX-3000 (122MB) were not able 
to perform as well in the tests applied to all drives.  The $9,995 Laser 
Bank 800 (400MB) includes a 25MB hard-disk staging unit, performs well on
all tests but takes up to twice as long.  The 120-122MB units are all based
on packagings of the ISI drive.
   Information Storage Inc., 2768 Janitell Road, Colorado Springs CO 80906;
(303) 579-0460.
   Maximum Storage Inc., 5025 Centennial Boulevard, Colorado Springs,
CO 80916; (303) 531-6888.
   Micro Design International Inc., 6985 University Boulevard, Winter Park,
FL 32792; (305) 677-8333.
   Optotech Inc., 770 Wooten Road, Colorado Springs, CO 80915; (303) 570-7500.
   Symphony Systems Inc., N/Hance Systems Division, 908 Providence Highway,
Dedham, MA 02026; (617) 461-1970.
   [dh:88-03-04]

%A Stuart J. Johnston
%T Microsoft Steals Show at its CD ROM Meeting
%J InfoWorld
%V 10
%N 10
%D March 7, 1988
%P 19
%K CD ROM Microsoft Facts on File Guide Reader Hypertext Grolier
Electronic Encyclopedia BMUG anthology Arthur Young Hypercard
%X At the Third International Conference on CD ROM held in Seattle, there
were the following announcements:
  Small Business Consultant (SBC) CD ROM from Microsoft, including
more than 220 U.S. Government publications about starting and operating
a small business.
  Stat Pack CD ROM from Microsoft, giving users access to more than
10,000 tables of U.S. Government facts and figures, including the
Statistical Abstracts of the United States.  ($125 for Stat Pack,
$149 for SBC)
  Version 2.0 of the CD ROM extensions to MS-DOS, supporting both
original High Sierra format and the recent ISO 9660 DVI standard,
also endorsed by GE, Lotus, and Intel.
  CD ROM edition of Facts on File for both Macintosh and DOS.
  Guide Reader hypertext for Windows 2.0 from Owl International and
used as a read-only version of Guide.
  Grolier Electronic Publishing producing both Macintosh and PC-based
($395) versions of the Electronic Encyclopedia.
  Anthology of Berkeley Macintosh Users' Group (BMUG) shareware
library on CD ROM with Hypercard.
  CD ROM Version of Arthur Young's Hypercard-based reference materials.

%A Jerry Pournelle
%T WORMS and Friends: Words for the Ages and Pictures for the Millions
%J InfoWorld
%V 10
%N 10
%D March 7, 1988
%P 41
%O A User's View: column
%K ISI WORM Drive CD ROM Publications
%X Speaks of advantages of WORM, using ISI drive as example.  Figures
that if the cartridges don't last forever, by the time they need to
be converted, there will be something more durable available.  Sees
use of WORM as a daily backup device, so there is a deep, complete
history of state of one's working files on a day-by-day basis!  This
gives both backup and way to retrieve/restore old data for any reason.
  Mentions that NASA's Jet Propulsion Laboratory has two CD ROM disks
with planetary science data: one with 800 Voyager images, another
containing 1400 files of data on weather, Mars Landers, etc.
  Also cites CIA's The World database and Turbo Pascal engine for
accessing it.  The World is distributed by the National Technical
Information Service.  The access program is from Fred Pospechil,
3108 Jackson Street, Bellevue, NE 68005.

%A (Reuters)
%T Compact disk reinvented: This time for computers, TV
%J Rochester Times-Union
%D March 7, 1988
%K Philips CD-I players
%X Compact Disk interactive, also introduced at the Microsoft CD ROM
Conference, provides interleaved audio and limited full-motion video
on a CD ROM.  But CD-I is intended mainly for non-computer home use.

%E Greg Kearsley
%T Artificial Intelligence and Instruction: Applications and Methods
%I Addison-Wesley
%C Reading, MA
%D 1987
%Z ISBN 0-201-11654-5
%O 352pp, $32.25
%X This book documents the attempts of researchers to design and develop
ICAI programs for the real world of education and training.  State-of-the-art
work by leaders in the field of ICAI is presented, along with actual
methodologies for building ICAI systems.
  -- publisher's abstract.

%A P. K. Garg
%T Abstraction Mechanisms in Hypertext
%I University of Southern California, Computer Science Department
%C Los Angeles, CA 90089-0782
%D 1987
%R CRI-87-71
%O Available for $4.00 payment to USC Computer Science Department
%K Delisle Schwartz abstractions
%X "Abstraction is the means by which information can be stored and
retrieved from an information structure at different levels of detail
and from different perspectives.  As such, abstraction mechanisms in
hypertext are interesting to study and evaluate. In this paper we study
the abstraction mechanisms in hypertext from a theoretical perspective.
Abstractions then become various first-order logic formulae.  Specifically
we consider abstractions: sets, sequences, aggregations, generalizations,
revisions, and information structures.  Interesting results of this
work are the definition of level of generality of a hypertext node,
the demonstration of revision histories as a partial order, and the
notion of compatible--similar nodes.  Also defined in this paper is
the notion of primitive hypertexts versus application hypertexts,
and the usage of attributes of nodes (illustrated by the use of
keywords) across various abstractions.  An illustration of the concepts
is given using the context mechanism suggested by Delisle and
Schwartz.

%A P. K. Garg
%A W. Scacchi
%T A Hypertext System to Manage Software Life Cycle Documents
%I University of Southern California, Computer Science Department
%C Los Angeles, CA 90089-0782
%D 1987
%R CRI-87-72
%O Available at $4.00 paid to USC Computer Science Department
%K System Factory Hypertext DIF USC
%X "DIF is a hypertext system which helps integrate and manage the
documents produced and used throughout the life cycle of software
projects.  It was designed for use in the System Factory, an
experimental laboratory created at USC to study the development,
use, and maintenance of software systems.  DIF provides an interface
to a hypertext-based information storage structure and to a
structured documentation process.  A hypertext of software
information is built by the software engineers ove eight life
cycle activities.  Salient aspects of DIF include integration
of documents within and across several projects, facilities for
browsing a hypertext of software documents, support for parallel
development of documents, interface to various software engineering
tools, and facilitation of documents reuse."

%A P. K. Garg
%A W. Scacchi
%T A Software Hypertext Environment for Configured Software Descriptions
%I University of Southern California, Computer Science Department
%C Los Angeles, CA 90089-0782
%D 1987
%R CRI-87-73
%O Available for $4.00 to USC Computer Science Department
%X "Large scale software systems are developed, used, and maintained
by teams of people over considerable periods of time.  The source
code of such systems, while useful for compilation and running of
the system, provides insufficient information for a proper understanding
of the system by humans.  Scaffolding information, such as design
documents, user's manual, and maintenance history are required to
efficiently use and maintain such systems.  In this paper we suggest
that this scaffolding information about a system should be tightly
coupled with the source code for proper evolution of the system.
Such a coupling can be achieved by maintaining all relevant information
about the system (including the source and object code) in a *software*
*hypertext*.  Versions and configurations of software systems then
reflect the differences in scaffolding information along with the
source code.  Various abstraction mechanisms on hypertext can be
effectively utilized to understand the software descriptions from
different perspectives.  In particular, we find the visual
representation of large software sytems a useful description that
complements the formal specification of a system configuration.

%A P. K. Garg
%A W. Scacchi
%T On Designing Intelligent Hypertext Systems for Information Management
in Software Engineering
%I University of Southern California, Computer Science Department
%C Los Angeles, CA 90089-0782
%D 1987
%R CRI-87-74
%O Available for $4.00 to USC Computer Science Department
%K DIF Intelligent Software Hypertext System Ishys
%X "Information management in large-scale software engineering is a
challenging problem.  Hypertext systems are best suited for this
purpose because of the diversity in information types that is
permitted in the nodes of a hypertext.  The integration of a hypertext
system with software engineering tools results in a software
hypertext system.  We describe the design of such a system called DIF.
Based on our experiences in using DIF, we recognized the need and
the potential for developing a hypertext system that could utilize
knowledge about its users and their software tasks and products.
Such a system might then be able to act as an active participant
in the software process, rather than being just a passive, albeit
useful storage facility.  As such, we define an Intelligent Software
Hypertext System (Ishys) as a software hypertext system which is
knowledgeable about its environment and can use such knowledge to
assist in the software process.  This knowledge is partly embedded
in the design of an Ishys (in terms of the `agents' that Ishys
supports) and partly defined during the use of Ishys (in terms of
taks that agents perform).  We present a framework for defining and
organizing this knowledge, describe potential uses of such knowledge,
identify limits of our approach, and suggest methods for
circumventing them."
  [dh:88-03-14.  USC report abstracts from ACM SIGSOFT Software
Engineering Notes 13, 1 (Jan. 1988), 73-74.]


--
	-- orcmid {uucp: ... !rochester!sjfc!deh0654
		   vanishing into a twisty little network of nodes all alike}