RE: [xml-dev] What are Web Services for? (Was RE: [xml-dev] lots of WS reading material)

Paul Prescod:
> I gave some examples in the article. You can't XInclude SOAP resources.
> you can't RDF-assert them. You can't download them with the document()
> function used in XPath, XPointer, XSLT and XQuery.
> 
> Also consider the following URI. It discusses a use-case involving
> mapping software:
>
>  * http://www.prescod.net/rest/rpc_for_get.html
>
> And this one discusses the *big issue*, integration:
> 
> * http://lists.w3.org/Archives/Public/www-tag/2002Apr/0286.html
> 
> This integration theme is also addressed here:
> 
>  * http://www.xml.com/pub/a/2002/02/06/rest.html

I read the first three sentences of your mail and finally got it. Amazing
what a good night's sleep can do for you. Sorry for being so dense, but
sometimes playing devil's advocate is the best way to understand something.

It's interesting that the thread about XML documents vs. infoset is running
in parallel. To me these discussions are very close. What makes XML magical
is actually extremely subtle, and tied up in the fact that it uses a text
representation, as some have pointed out. The same is true for REST vs.
SOAP. They accomplish basically the same things, but the nature of the
implementation actually makes REST far more powerful. The reason, as far as
I can see, is related to network effects. When you can do a lot of things
usefully with something, then its value goes up exponentially.

I also think that the case of RDF is analogous to that of the SOAP. The
underlying idea (providing metadata for resources) is very powerful. But RDF
doesn't let you leverage the "other" things you can do with XML as well as
alternative approaches could.

> In that particular case the resources happen to be read-only. They could
> be read-write, for instance, if it were possible to PUT a new
> representation of a cached page so that Google's cache and index would
> always be up-to-date. (of course there are security implications there
> but you get the idea!!!)

Still less convinced by this. I don't know many cases of PUT being used to
change resources. Maybe I am still in the dark ages, but most cases I've
seen used POST in a controlled way to change server data.

> I think it will depend too much on the technical competence of readers
> to make much money. Remember the blinking 12:00 on the VCRs. People will
> put up with a lot of hassle to avoid learning something technically new
> (unless it can be learned in a VERY incremental way).

Hmmm, maybe this is the answer to the question of how Yahoo can make money
in the post-advertising age. Offering a service that lets users consolidate
data from various sources in a user-friendly way could be a money spinner if
Yahoo (or AOL or MSN or whoever) pays the information providers and charges
users a flat monthly rate.

Matt