[DotGNU]Why DotGNU should go Jabber (was Re: Microsoft guru: Stamp out HTTP)

[Norbert Bollow]

> > http://zdnet.com.com/2100-1105-845220.html
> 
> I wouldn't be too concerned.  HTTP does publication
> extreamly well, and the www browser is ubiquitous. 
> RPC servers other purposes.  

Yes... even though I tend to be very suspicious of Microsoft's
business ambitions, I don't sense any evil in what Don Box has
actually said -- he is simply right that there is a need for a
protocol that allows to send to send the requested results
immediately or send them later in essentially the same way.

(Of course an intention to "Stamp out HTTP" would be evil, but
that's just the title of the article, it's not what Don Box was
actually speaking about.)

With widely-deployed standard protocols, it's of course possible
to send back an XML response to the HTTP request which means
"we'll work on your request and send the results by email".
This is not a good solution though.  Email systems are often
administered by different people using different tools from what
it used for administering "web stuff".  This kind of system
would simply not be robust enough.  A system is needed which,
when tested in a simple situation (that results in an immediate
response) will also work in complex situations (where results
are not available immediately).

So Don Box is right, a replacement for HTTP is needed for the
case of requests where "full results immediately" would be nice,
but where getting the results later is acceptable also.

In fact this has been discussed before (I'm not sure whether
that was on this list or on one of the Jabber meetings that Adam
Theo has organised), and the suggestion was made to use Jabber,
specifically to endorse (and thereby hopefully revive) the
"Jabber-as-Middleware" ideas, see

http://oid.jabber.org/index.php?oid=547

If/when Microsoft comes out with a new protocol of their own
(possibly endorsed by other industry giants as well) DotGNU
will have to support that also.  But if we push strongly enough
in the direction of establishing jabber:middleware as a
standard, I think that there is still a hope that it could
become the de-facto standard.

What are our alternatives to going for jabber:middleware ?

- We could wait and see if W3C comes up with something usable.

- We could wait and see what Microsoft does in .NET and do the
  same.

- We could try to get together with the major corporate players
  (Microsoft, Sun, IBM) and agree on a protocol.

The big advantage of jabber:middleware over all of these ideas
(even assuming that they're even workable, which is by no means
sure... it could be that Microsoft goes for a patent-encumbered
protocol, W3C doesn't come up with a good alternative, and the
major corporate players aren't interested in talking with us)
is that Jabber has significant developer mindshare already.
I think that DotGNU probably just doesn't have any real
alternative to going for a protocol which has such a following
among developers already, as we don't have the resources to pay
developers for implementing things with a new protocol that
hardly anyone is really interested in yet.

Greetings, Norbert.

-- 
A founder of the http://DotGNU.org project and Steering Committee member
Norbert Bollow, Weidlistr.18, CH-8624 Gruet   (near Zurich, Switzerland)
Tel +41 1 972 20 59       Fax +41 1 972 20 69      http://thinkcoach.com
List hosting with GNU Mailman on your own domain name   http://cisto.com