Lists Home |
Date Index |
> Paul Prescod wrote:
> The "user pull" I'm speaking of is not end-user pull but developer
> The developers are the consumers of specifications and development
> technologies. There is a subset of developers who follow Microsoft's
> every move and another subset who are more interested in what takes
> in the open source and Unix worlds. The prevalence of Apache
> demonstrates that this latter group has power out of proportion with
> their numbers.
In some ways we may be in violent agreement. IMO it's distributed
application developers that are creating the "user pull" for things like
XML Schema, WSDL, and XQuery in business-oriented scenarios.
> One vision for the future of web services is that the vendors choose
> technologies, deploy them within businesses for EAI and then across
> business boundaries for B2B and non-EAI or non-B2B projects either
> the same technologies or go their own way.
> An alternate vision is that both groups will fumble around in the dark
> for a few years until a disruptive technology arrives that is the key
> that was missing. This disruptive technology is much more likely to
> arise outside of the vendor groupthink that initially resisted the
> Internet and then the Web.
Definitely a possibility.
> I have no evidence that this latter scenario is the one that will come
> about but my instinct says that reinventing SUN RPC as SOAP and IDL as
> WSDL and ASN.1 as XML and SQL as XQuery and ... will not get us to a
> fundamentally different place than the first go-round with those
> technologies did. I admit, I could be wrong. But something feels wrong
> about piling on layers of complexity hoping that they will provide us
> with a stairway to heaven.
Not a stairway to heaven but hopefully interoperability between
services. Achieving interop between distributed services is inherently
more complex than functionality provided by the Web up to this point.
Today, each distributed application platform provides a different set of
technologies for data representation, transport, description, and
discovery not to mention higher level services like security,
transactions, workflow, etc. I like to think of the layering you're
referring to as the new "Web Services Platform", which at least gives us
a common set of technologies, none of which are perfect, that have been
accepted by most vendors as a potential road to interoperability, in
this complex, competitive space.