Lists Home |
Date Index |
Miles Sabin wrote:
> Paul Prescod wrote,
> > Miles Sabin wrote:
> > > HTTP just wasn't designed for that kind of communication model.
> > HTTP is perfectly acceptable for that kind of transaction. Hint: the
> > "client" needs to be running a micro-HTTP server.
> How does that help with disconnected operation (ie. the client might
> be offline when the response comes back)?
You need either retry or a forwarding intermediary, just as with any
other protocol. The question is merely whether the intermediary should
"speak" the same protocol as the rest of the Web, using the same naming
system (URIs), or should have its own naming system (e.g. email
addresses or Jabber IDs) and its own protocol.
By the way, if you are concerned about developing conventions for using
HTTP in manners it has not been used, then how do you feel about
SOAP-over-HTTP wherein people knowingly and consciously do things that
violate the SOAP model such as using POST for idempotent queries and
"success" result codes for failures?