Lists Home |
Date Index |
> There isn't a whole there; just pieces to wire up
> into various systems with varying degrees of
> interoperation depending on the way these are
> hooked together,
And that's "just"?! 8-O
Dave wrote (if I've got the attributions correct);
> Fascinating question. Is the web identified by protocols, formats,
> identifiers, or some combination? Is the web:
> 1) HTML, URLs, HTTP?
> 2) XML and HTML, URLs, HTTP
> 3) XML, URLs, HTTP
> 4) XML, URIs
> 5) XML and HTML, URIs
> 6) URIs
> or some other combinations?
> I tend to personally think the web is #5 and it typically uses HTTP.
There was a Web before XML. I'd say it's closer to #1. URIs are the
core, of course. But running a close second is HTTP, or more precisely
the architectural style that it embodies (aka REST). HTTP may be
replaced in time (no time soon though, and not by SOAP - and we won't
need a new URI scheme either), but whatever that new protocol is, it
will look an awful lot like HTTP and will at least have the same basic
methods; GET, PUT, POST, DELETE.
As Gavin said, the Web is an *application*. Applications need
application semantics, and HTTP provides them.
> If I build an RDF application using URIs and I use gnutella/tcp to
> distribute it, is it part of the web?
> If I build a SOAP document and ship it via SMTP, is it part of the web?
Don't want to go there, except to say if the SOAP document is an RPC
call, it's definitely not the Web.
Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA. email@example.com