Lists Home |
Date Index |
"Bullard, Claude L (Len)" wrote:
>... Once Gates understood it, from
> a standing start he bested most of his competitors
> in eighteen months.
Sure. He probably will again. I'm just pointing out that you may not
want to follow him down the blind alley before he refocuses on what will
work. He isn't down in the trenches working with the technology and he
may not have understood what went wrong the first time well enough to
avoid replicating the mistake.
> The question now is whether or not the base
> web service specs as they stand today and
> the code on the docks is ready for us to
> work with and get something happening. That
> is EXACTLY where HTML and HTTP started: not
> perfect but good enough to start the conversation.
The web started with one and only one really good idea: "Let's make a
universal namespace." HTML was crap (except for being easy to use and
implement). HTTP was crap (except for being easy to use and implement).
SOAP could be technically excellent or crap. Doesn't matter. It's
survival characteristics will depend on the extent to which it
acknowledges that the only thing that really, really, really matters is
that universal namespace. As Cairo and Blackbird did not. Oh yeah, and
it would help if it was easy to implement, which SOAP is not.
> As to getting "googled", it cuts in all directions
> including in the direction of the W3C if they
> ponder too long.
The W3C shipped the core technology of the Web Services revolution
fifteen years ago. Since then they have backed it up with some other
solid technologies including HTTP 1.1, XML and RDF.
What's missing? Just you. You need to model your business process in
terms of XML documents, hyperlinks between them and operations upon
them. You can do that
a) now, using HTTP, XML and URIs, or
b) a year from now, using SOAP, some proprietary set of SOAP methods,
XML bodies and some proprietary addressing scheme, or
c) two years from now, using HTTP, XML and URIs. ;)
Actually, I can't promise with 100% certainty that SOAP will go
away...we could rebuild something HTTP-like on top of it, just for the
thrill of reinvention.