Last update: 24 March, 2001 by Jacob Palme E-mail: jpalme@dsv.su.se. at the Department of Computer and Systems Sciences, Stockholm University/KTH
This paper is written in response to queries for information about the new version of the KOM computer conferencing system being developed in 1994-1998.
This paper will mainly describe differences compared to our previous conferencing system, SuperKOM. You can read the paper SuperKOM -- Design considerations for a distributed, highly structured computer conferencing system in Microsoft Word format by Jacob Palme and Torgny Tholerus, published in Published in Computer Communications vol. 15, no. 8, October 1992 pp 509- 518, for information about SuperKOM.
Two user interfaces will probably be developed for KOM. One is the Web4Groups interface, which will provide access from ordinary web browsers to special HTTP servers using HTML forms to access the KOM facilities. The other is a fully graphical user interface, using menus, windows and buttons according to the present fashion for personal computer software. This will probably be programmed in Java. No decision has yet been taken to implement this user interface.
The HTTP based user interface has been carefully tested on users within the Web4Groups project. The Java user interface has been carefully tested by developing a series of prototypes which have been tested in several stages with real users at DSV and at Celsius Information Systems AB.
Because of the use of HTML and Java, Web4Groups will be platform independent and we hope that KOM will be usable on most major platforms (Windows, Macintosh, Unix Motif, etc.).
KOM will initially support:
Every user has a personal address book. The user has full control
of this address book, there is no automatic purging of it.
Every server has a directory of names of users, conferences and
external e-mail addresses which have passed that server. There is no
central directory and no directory queries via the server-server
protocol.
The server-server protocols is at present only available for
TCP/IP nets.
Like SuperKOM, KOM can be distributed on several sites. Like
SuperKOM, replication is automatic: As soon as any user at any server
becomes a member of a conference, that conference will be distributed
to that server. No manual set up of replication is thus necessary.
All information is not replicated, messages are only replicated to
servers where there are users who are recipients of them, as e-mail
or as conference members.
Different from SuperKOM is that if a user asks for a message which
has been purged from the data base of that server, KOM can get the
message from its original server. Messages can however also be cached
locally at servers where they are frequently accessed.
Conferences are normally presented with presentation messages in one
or more other conferences. To make information about a conference
known at other servers, so that their users can join the conference,
you just need to send the conference presentation to a conference
which is replicated to that server.
The client-server protocol is at present only available for TCP/IP
connections. Dial-up phone connections are only possible if TCP/IP is
used, via e.g. SLIP, PPP or ARA.
The client-server protocol at present requires on-line connection
during the whole session. All information is stored in the server,
and only cached temporarily in the client during execution.
The following advanced functionality will be implemented:
In August, 1996, there was a working line-oriented user interface. The HTTP server was in alpha testing stage. The Java client had not yet been developed..
KOM will probably be available free of cost to universities, and sold commercially to non-universities. No prices have yet been fixed. Source code will probably be available subject to special agreements in each case.