RE: HTTP-ng status

From: Mike Spreitzer (spreitze@parc.xerox.com)
Date: Thu, Oct 07 1999


From: "Mike Spreitzer" <spreitze@parc.xerox.com>
To: <martind@netfolder.com>, "'Simon St.Laurent'" <simonstl@simonstl.com>, <ietf-http-ng@w3.org>
Date: Thu, 7 Oct 1999 15:47:11 PDT
Message-ID: <002301bf1115$e4bb5190$27d1000d@deimos.parc.xerox.com>
Subject: RE: HTTP-ng status

> So what is happening now with the chair work?

Well, as I said, all *I'm* doing is working on publishing (well, maybe
cleaning up a *little*) what we've done, and working on the research
problems revealed (PARC is a research lab, after all).  If anyone else
wants to step forward to chair an IETF effort, that would be applauded by
me.

> >From an otehr perspective, being an ILU user myself, I found
> several things
> from usage...
>
> So, if HTTP-NG would use XML as a marshalling tool, we would have the
> occasion to:
> - merge potentially several world like Webdav, the usual HTTP and object
> middleware.
> - resolve the biggest problem experimented with the usage of XML is the
> impedance mismatch between this format and procedural languages.

I think you've represented one corner of the world I'd like to see HTTP-NG
address, but not the whole world.  There are others who would like to see
other things emphasized, and for whom XML-based marshalling, for instance,
is more trouble than it's worth.  I think the answer here is to be
radically modular (at least in design --- but not necessarily any
implementation, for performance's sake), letting different folks get what
they want, and decide for themselves what interoperability trade-offs they
want to make.

Another thing I'm interested in is the problem of significant evolution in
the sophistication of a largely deployed system.  Successful systems often
start out as small, simple things.  But as they succeed and grow in scope
and ambition, initial trade-offs between simplicity and power become
inappropriate, and a switch to more powerful underpinnings becomes
desirable.  But rarely practical.

Mike