[
Lists Home |
Date Index |
Thread Index
]
RDDL is fine at that level of integration. I just
don't see it mentioned as often outside XML-Dev in
articles on web services as the umpteen other web
service languages such as WSDL, UDDI, XLang, the MS proposal
for an Inspection language and so on. Part of negotiation
is picking the system or protocol of negotiation. I need
to be convinced RDDL is important. I know, for
example, WSDL and UDDI are given the vendor
support. I also understand there are problems
with WSDL if used as an IDL (see this weeks'
XML.COM article on the subject).
http://www.xml.com/pub/a/2002/01/16/endpoints.html
That aside, my intuition is that the really useful
web services for the organizational programmer will
be coarse-grained and at the level of reports because
that is something the organization has, it understands,
and which is easily discoverable and negotiable.
Below that level, intense and sometimes expensive
customization is required, so some web services will
turn out to be quite local or limited as the potential
numbers of clients and possibly transient.
Negotiation is typically dynamic, is done as a series, and
depends on acks to ensure both sides understand
each other or to put it another way, accept the
domination of one side's understanding. If you
negotiate contracts, you play a game of option
swapping (always this, this for that, this but
not that, never this or that) and so on.
If a web service is fine grained (the Amy Lewis
model), can I detect cases of deadly embrace
easily?
len
|
