- From: Paul Ramsey <pramsey@makitso.com>
- Date: Sat, 24 Sep 94 11:51:52 PDT
- To: www-html@www0.cern.ch
While I can imagine that being able to send files might be useful the method you describe sounds like a dramatic shift in the architecture of the WWW. Currently the HTTP protocol is a client/server model, a client establishes a connection, sends a request, gets a response, and shuts down the connection. Each time you select a link this process is repeated. You're proposing that clients start behaving like servers to allow remote WWW sites make requests for files. On a unix system this is security problem you're unlikely to get through very many firewalls with such a request. On many other systems this is also impractical because the software isn't really set up to handle transactions that are initiated by another system. Any real solution will need to stay within the stateless environment of the current HTTP protocol. paul ramsey
Received on Saturday, 24 September 1994 21:02:49 UTC