Hi,
On Sun, 5 Dec 1999, Marc van Woerkom wrote:
> > I agree that we should press on with our simple XML over HTTP solution
> > rather than wait for an IETF standard to do RPC over the Internet.
>
> The SOAP i-d proposes a nice extension of the stuff we already do,
> and going that step further, interpreting the data exchange as RPC is
> quite elegant.
>
> What's the difference?
> Instead of sending arguments to distinct URLs that encode the methods,
> one states them explicitly in the HTTP message body.
>
> Plus you add an object, thus it makes adding different but related
> services (think: mp3, DVD) easier.
>
> Can't see much room for evil here.
Well, some points:
- - The current draft draft-box-http-soap-01.txt is of category
"Informational" which means it will never be any kind of internet
standard. See RFC 1796 for more details on this topic.
- - The draft makes use of the HTTP Extension Framework for the required
M-POST command which itself is currently an expired draft.
- - Finally, the current draft specifies soap.v1. IMHO this should have been
something below 1 (e.g. soap.v0.9) before the protocol has been
released as RFC. Else, implementation can't destinguish between
pre-release implementation and implementations of the final version.
IMHO, the draft is a nice way to do RFC (or even remote method calls
like CORBA does) over HTTP, but it's too early to jump on this technology
now. What we need is an API which can be immediately used right now. We
can always switch to a better network transport protocol later. As long as
the data model (and ID calculation) doesn't change, the transport protocol
should (almost?) be transparent to the application developer.
Cheers,
Jochen
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.0 (NetBSD)
Comment: For info see http://www.gnupg.org
iD8DBQE4SmS/0fhX0Y/ocz0RAja1AJ48CBtXXzI5F60M7wteJ+4MF7vwTwCePk5U
jUxXAFx1a6siBTI7jumXvt0=
=LFfI
-----END PGP SIGNATURE-----