WWW Summary from IEEE MM 1994 Conference

Taken from Mail Summary given by

sg04%kesser@gte.com (Yechezkal-Shimon Gutfreund),

Tue, 24 May 1994 12:33:07 +0200

Hierachical caching, multi-part MIME messages

briefly discussed as a good thing.

Continous (as apposed to stateless HTTP)

A general consensus for need, but a feeling that it should not be part of the HTTP protocol

Multi-Threaded WWW clients

Sure. When can we get them?

Embedded audio, movies, and other viewers

Considered part of a new WWW client archicteture

New client architecture

General consensus that one large "do-everything" client is an outdated concept. Clients should be composed out of a hetergenous mix of pieces - all cooperating to produce one logical WWW client. However, there was much to still be resolved as to how these pieces communicate. Some were inclined to extend HTTP, so want a new protocol. I am partial to shared distributed memories along the tuple-space model of Linda, etc.


lots of concern about the cost of firewalls (e.g. at Bell Labs). and also lots of concern about new protocols for continuous interaction, and the holes they might open up. Basically, we face competing needs (1) to open up the protocol to allow more live embedded and interactive apps to work in conjuction with the client (2) the need for restricited protocols that can be verified for rigor and security.


Ed. Henrik Frystyk