[
Lists Home |
Date Index |
Thread Index
]
Paul Prescod <paul@prescod.net> wrote: (in different posts)
>
> It was wrong to hijack a common term.
Well, yeah. OK. But the damn term is a meme out there in memespace and
the W3C, Microsoft, and IBM couldn't exterminate it at this point even if
they wanted to or had any motivation to work together and try. After all,
the whole point of the "Web Services" architecture WG being chartered was
to figure out what people really meant by "web services" in an
architectural sense and to try to find some order amidst the chaos.
To beat the "hijack" metaphor into the ground (endangering good taste in
the process!), a group of marketroids hijacked the plane, couldn't agree
where to fly it, so they appointed a committee of the passengers to decide
... retaining veto power over the decision. "Let's go back to the original
airport and start over" doesn't seem to be an option that they will allow
at this point <groan>.
> The question is whether the W3C should now ratify that decision so that people
> like my wife and your cousin will be permanently confused about what
> computer people mean by "Web Service" or whether the W3C should invent a
> new term so that elite, cutting edge computer people can
> use the term the same way everyone else has been using the term as common sense
> since the mid 90s.
You seem to greatly misunderstand what the W3C *is* and what it can
possibly do. It's a consortium of the vendors, not a legally constituted
authority in its own right. It can Recommend whatever it wants, but if the
Recommendations fall on deaf ears (as so many do), then what? And why
would the vendors even contribute to the development of a Recommendation
that essentially said "ignore all those billions of dollars of hype and
vision-ware, let's start over from scratch with this stuff." It's not
going to happen.
Sigh, if we want to get literal about it, XML is not really an "Extensible
Markup Language", it's a generic syntax and/or data model for defining
markup languages. Maybe we should discuss changing its name to GSDMDML or
something :-) Or maybe we should accept the reality that human languages
hijack terms all the time and learn to live with this one too.
But to address some of the questions in this thread: It is a VERY hard job
to figure out what common thread runs through all definitions of "web
services," and what the W3C can or should do to define a reference
architecture that points in the "right" direction. Is Google (the one we
know and love via HTML/HTTP) or the wretched canonical example of a stock
quote site a Web Service if I can write a program to generate its URLs and
screen-scrape the results? That's probably the lower bounds. The upper
bounds would be something that uses queryable metadata (RDF, WSDL, or
otherwise) to define the syntax and semantics of a collection of services
and something like the SOAP extension/processing model to allow secure,
reliable, choreographed, encrypted interaction between dynamically
determined endpoints using only generic software.
Does the term "web service" cover all these? Or maybe only the ones that
only use URI+HTTP+XML to do the job? Or maybe there has to be queryable
metadata defining message formats, and the stuff that involves sneakernet
communication of data formats and screen scraping doesn't qualify as being
sufficiently machine-oriented to be a "web service"? Or maybe they all are
"web services" but they get more and more specific adjectives associated
with them, e.g. "REST web services", "SOAP/WSDL/RPC web services" or
whatever. That makes "web services" a fairly meaningless term, but then
again, so is the "world wide web" when you try to deconstruct it too
rigorously.
|
