No Subject

I wrote most of the internet chess server (which can currently be
reached with "telnet ics.uoknor.edu 5000").  A number of other
programmers have written interfaces that run on your local
PC/MacIntosh/Xterminal/Next.  The interface parses the output from the
chess server and displays a pretty chess board on your screen.  It
also allows the user to make moves with the mouse.  These mouse
strokes are converted into algebraic chess move notation.  Some of the
interfaces show a dynamically changing clock, indicating the amount of
time remaining for each player.

Each of these interfaces is a custom-made special purpose-program for
connecting to this chess server.  A similar set of interfaces has been
created for use with the go server.

This whole set-up is really cumbersome for a number of reasons.
(1) It's very difficult to set up other game servers with graphics,
because a whole set of special purpose interface programs must be
written, one for each major platform. (2) Any change in the chess
server must be backward compatible with all the interfaces out there.
This limits the the way the server can develop.  (3) When a separate
new platform emerges, a new interface must be written for each of the
internet games with graphics.

All of these problems could be solved by a standard language for
expressing the simple types of graphics needed for these games,
combined with a set of interfaces to interpret this language on a
variety of platforms.  Postscript suffers from at least two drawbacks.
(1) It's extremely verbose -- since many users connect through slow
lines, this a problem. (2) full postscript is a HUGE language that is
very difficult to implement, and 99% of it is unnecessary for these
purposes.

I believe that the development of such a set of tools would spur a
tremendous explosion of new simple graphics games and applications
that could run through telnet the way the chess server does now.  It
would be possible for one person to write a new game (such as double
bughouse chess) without having to write a half dozen graphics
interfaces.  Many really cool things change from being impossible to
being quite feasible.  (The PLATO system developed in the 70s at the
University of Illinois had some of these properties: simple graphics
available to all users, fast interaction among a large pool of users.
The result was the development of a number of very popular and
engrossing interactive games.)

So, my question is: does a language with these properties already
exist?  If not, how do we go about creating it?  This whole idea seems
to fit quit well into the philosophy of Mosaic, which is a standard
interface to the net that runs on all platforms.  If it emerges, would
this new type of network interaction be built into Mosaic?

Please forward this message to anybody else who you think would have
some useful insight on this problem.  Thanks.


Daniel Sleator

Professor of Computer Science

Carnegie Mellon University