RE: FW: draft findings on Unsafe Methods (whenToUseGet-7)

[replying only to www-ws-arch]

> -----Original Message-----
> From: Don Box [mailto:dbox@microsoft.com]
> Sent: Thursday, April 18, 2002 9:29 PM
> To: Hutchison, Nigel; jones@research.att.com; moore@cs.utk.edu;
> www-tag@w3.org
> Cc: dorchard@bea.com; www-ws-arch@w3.org; xml-dist-app@w3.org
> Subject: RE: FW: draft findings on Unsafe Methods (whenToUseGet-7)
> 
> 
 
> I'm sorry, but if I must chose between expressing most of my app in
> terms of HTTP GET vs. having the switch/router/gateway/cache 
> fabric get smarter, I'll take the latter any day. 

To quote my favorite passage from THE MYTHCIAL MAN MONTH once again, "The
real tiger is never a match for the paper one, unless actual use is wanted."
Sure, you'd choose the "paper tiger" ... unless you had to build a working,
scalable system today. "Then the virtues of reality have a satisfaction all
their own", as Brooks goes on to say.  

> As I recall, network infrastructure adapted to a new protocol 
> (HTTP) in the 1990's without the death or dismemberment of the Internet. 

Definitely. One can certainly envision new levels of the internet protocol
stack built on XML and SOAP, and smart switches and routers and caches that
do a much better job and getting the right messages to the right place in a
way that is much more effective and efficient than how we do it now.  Great!
We probably wouldn't be on this mailing list if we didn't appreciate this
vision, and most of the companies here are probably working on some piece of
that vision.  But right now the operative word is VISION -- that's we want
to be.  What's much less clear is HOW we get there, i.e. what we keep, what
we invent, and what we tweak.  There's a lot of baby in the HTTP bathwater,
and we need to be VERY careful about what we throw out.

> 
> It is happening again in this decade, since like it or not, SOAP seems
> to be where networked applications are headed. 

I vaguely remember some other "like it or not, this is the way the world
is headed" technologies that didn't quite have the impact that their
promoters
expected ... OSI networks, OS/2, Ada, CODASYL, X.400, X.500, come to mind.
Most of us probably wish we had made well-timed contrarian bets on companies
that invested in TCP/IP, Windows, C, RDBMS, SMTP/POP, etc. There are a lot
of good ideas for a smarter switch/router/gateway/cache fabric that tunnels,
and perhaps replaces HTTP, and I'm intrigued by them.  Nevertheless, they
are going to have to succeed the way HTTP succeeded -- by solving real
problems more cost effectively than the alternatives.

Received on Thursday, 18 April 2002 22:34:11 UTC