Systems and Applications

Here is a list of existing systems, past present and future, which we have come across. These include both academic research systems and commercially available systems. The order is random. Some Information retrieval systems have crept in which are not hypertext.

Based partly on the ECHT90 tutorial by Paul Kahn , IRIS . See also: Sources of data . See also IR systems .

MIME

Multimedia Internet Mail Extensions are relevant thouigh they are neither IR nor hypertext.

Library 2000

MIT project by Michael Charity. Full-text index and bitmap graphic retrieval system for MIT technical reports. They don't want to use W3 because they want to be free to change the protocol any moment.

Interes

Developed by Milos Kravcik of University of Bratislava for PCs. Portable Acedemic Encyclopaedia -- in progress. We have a ZIP file of the system in hypertext/Products/Interes.

InfoMesh

Karen Rosin Sollins . See file://allspice.lcs.mit.edu/pub/wp/wp.tex,

Mediaview

Note from Dick Phillips. See also README . A graphics "publishing" system.

AMIX

Commercial PC/Modem "Information Marketplace"

Internet Gopher

PC/Mac client for CWIS form Minnesota. Supports index search and document retrieval. Many other clients.

Hyper-G

CWIS for technical university of Graz, Austria. Now fully deployed. Home of the Hacker's Jargon. A gateway from the Web exists. Their browsers have access to HTTP servers as well. We need a link here to their technical notes. Seen a nice character-based browser at JENC'92.

Hyperbole

Next in a line of research hypertext systems from Brown University.

HyperBase, EHTS

Experimental system. EHTS = Emacs-based HyperText System Graphical browser runs under X.

WWW

WorldWideWeb initiative. A wide-area system using hypertext and information retrieval concepts. Originated at CERN .

W.A.I.S

Wide-Area Information Service. The protocol is based on an ANSI standard Z39.50-1988. See also discussion group, mailing list wais-discussion@think.com and release notes . Putting a wais access protocol onto WWW is a very interesting possibility. JFG to look at it. A getway exists.

Bootstrap Initiative

Don't know much about this it may be tackling the same problem as us. See Doug's seminar . A reference is:

Engelbart, Douglas C. "Knowledge-Domain interoperability and an Open Hyperdocument System", "Proceedings of the ACM CSCW90 Conference on Computer Supported Cooperative Work", October 1990, pp 143--156

I have the book (Tim). Doug has an industry-based viewpoint in that paper. He aims at a fairly complete CSCW system, but suggests starting with several pilot projects based on Augment . This would furnish experience needed for the next phase.

NLS/Augment, (1969!)

structured documents, kept as outline, which allows zoom-in and zoom-out. Link anchors are full pathname and therefore can be across machines.

Xanadu

Ted Nelson's dream since early `60s: all the world literature in one publicly accessible global online system (analogy: you can today get a telephone link from anywhere to anywhere, so why not from any text to any other?). Every reference to a text will lead to royalties being paid automatically to the author. Autodesk, (the makers of AutoCAD) will produce a product "real soon now". Includes the use of full versioning (claimed to be horrifyingly complex), "hot links" (called transclusions) and zippered texts (eg. parallel texts like for translations or annotations.)

NoteCards (Xerox PARC)

Written in Interlisp D, single user, uses the card metaphor. There are two camps: the card sharks, who propone the idea that information comes in card-size chunks, and the holy scrollers, who propone that a hypertext node can essentially be any size and therefore has to be scrollable on the screen.

NoteCard links can be labelled with keywords for filtering. The most useful part of the system is its graphical browser, wherein the user can pick documents or edit links between them. There is one composition methaphor, the Filebox.

KMS (marketed by Knowledge systems)

A card system (claimed to have the Hypercard model but in 1972), wherein graphics and text can be mixed. The com- mercial version allows two card windows side by side (for copy-and-paste operations). It was described as "MacDraw with links". Its main feature is that sub-second response time is one of the design criteria. Scripts were first used in KMS: the traversal of a link can result in the execution of a piece of code of arbitrary complexity and effect (eg. Unix shell scripts). There is no graphical browser. KMS is multi-user but does not warn anyone of simultaneous access, so during updating, if two people edit the same node, one loses...

Hyperties (university of Maryland)

Designed for browsing (idea that many users will look at information, only a few will author it), so it is asymmetric and the authoring system is not nearly as easy to use as the browser. Sinlge user, runs on PCs with a research version for the Sun. Anchors are created with markup in the precompiled text.

Intermedia

A research tool developed at Brown university (van Dam (yes, Andy,) et al). The demo was impressive. There is a data base system parallel to the texts, that holds the links. A web is a context of links, several webs can be created over the same set of documents. Several links can emanate from the same anchor. Web views can be presented, and they also give the history of viewing for each user. It runs under Apple's old Unix version for the Macintosh (and will not run on the current version). It seems to be dead, since there is no funding to continue the research.

Guide

A product from Owl , a UK company. Guide is based on Peter Brown's work (U. of Kent, 1984). Runs on MS-Windows and Macintosh with a development version for Unix. Documents may contain text and graphics, struc- ture is given in application-specific files called "guidelines". Access to video and documents from other applications possible. There is no web visualisation.

Hypercard

The system for the Macintosh (Apple, 1987). Many claimed this not to be a hypertext system, but then the majority of real applications presented used it. This is the product that made hypertext take off. It is the BASIC of the 1990s. It follows the card model (but has scrolling fields...) Incorporates graphics and text. Other media are accessible through extensions. Powerful scripting language. It is the only one with which I personally have a lot of experience. The problem with Hypercard is that any fancy application requires writing scripts, but its advantage is that the scripting language allows you to do most of the things that are absent from other systems.

DynaText

SGML-oriented, X/posix based, portable. From EBT .

Toolbook

Hypercard like but has an object type drawing package rather than canvas type, and an object oriented scripting lan- guage. From what little was said about it, it seems to be better than Hypercard.

Analyst

Runs over ParcPlace SmallTalk, available for Sun, Mac, Dos. The document model is of an information centre where different types of documents are assembled from forms, databases, spreadsheets, maps, graphics and folders (which are like directories). Each document type has its own editor. Queries may result in virtual folders, somewhat like SQL views. Some support for artificial intelligence. (Who made it? -Tim)

HyperNews for sun workstations

Try contacting the developers on newsdev@turing.ac.uk.... See news article .

For the PC:

Microcosm

Multimedia for MSDOS from Southampton University. Neat.

Doorway

A hypertext system with annotation, graphics mixed with text. Commercial.