SOAP, RDF and Semantic Web screenscraping...

Just a quick thought after taking a look at one of the various SOAP
service registries that are out there, http://www.xmethods.net/

Their starter list of services all seem to be of a kind. Rather than
being generic distributed object / RPC apps, they're all simple
query-oriented services, like a richer version of the PICS-bureau-in-RDF
I sketched last november in http://www.w3.org/1999/11/02-RDFServices/

SOAP Service List from  http://www.xmethods.net/
	   Barnes and Noble Price Quote
                       Checks the price of a book at BN.com (Barnes and Noble) given an ISBN number. 
    
	   Domain Name Checker
                       Checks whether a domain name is available or not. 
   
	   Currency Exchange Rate
                       Exchange rate between any two currencies. 
   
	   California Traffic Conditions
                       California highway conditions. 
   
	   AreaCode / Prefix Locator
                       Location (city) of a given area code and exchange  prefix. 
   
	   Delayed Stock Quote
                       20 minute delayed stock quote. 
   
	   eBay Price Watcher
                       Checks current bid price of an eBay auction. 



Given the various threads on the similarities between the SOAP and RDF
information models, I'm intrigued. Quite likely I'm overgeneralising
from this one site, but if this is a trend, we might find many
(most?) SOAP services to be information-dispensing, non-state-changing
operations. The sort of thing that might be plugged into a query engine
to satisfy certain sorts of RDF query perhaps...? (FWIW I don't see the
RDF/SOAP overlaps as problematic in any way. As the various SW
screenscraping threads have shown, more structured data out there in
XML is a good thing for all of us...)

It's the oneway, information-grabbing focus of these example SOAP
services that suprised me slightly. Seems a far cry from the two-way-web
vision that Dave Winer and others have been (quite rightly!) arguing
for (eg. http://davenet.userland.com/2000/03/02/theTwowayweb ). 

Anyway, these all seem excellent candidates for wrapping up in an RDF
interface. One might imagine eg. the 'is this domain name free' checker
as a remote RDFish service that knew about factoids such as 
[?X]--rdf:type-->[foo:RegisteredDomain] and 
[?X]--rdf:type-->[foo:UnRegisteredDomain].  
assuming for simplicity a uri scheme for internet domains.

(or "rdf:type(?x, foo:RegisteredDomain)" in a less graphy syntax)


I've been playing around with wrapping Google backlink-searches and
Swish-E full text website indexes behind simple RDF APIs. I'm now
wondering whether it'd be more profitable to step back from the sort of
lowlevel screenscraping and think about... er... 'soap-scraping'
(or xml-rpc for that matter). Given say a Java implementation that
wrapped the above SOAP services, what extra conventions would we need
before these informationally oriented RPC-ish services could be represented as
RDF query servers?

That's quite enough thinking out loud. To compensate for taking up
bandwidth with half-baked speculation, here's a couple of interesting
links from yesterday's XML2000 in Boston:

	http://www.xml.com/pub/2000/09/06/distributed.html
	Distributed XML: the role played by XML in the next-generation Web, Edd Dumbill. 

	http://www.w3.org/2000/Talks/0906-xmlweb-tbl/
	XML and the Web, by Tim Berners-Lee 

Both available from the (recently XHTMLised) http://www.w3.org/RDF/

Dan


ps. I've a hunch Sergey's been building something similar to all this 
in GINF, though haven't quite figured out how GINF relates to XML-RPC,
SOAP etc...


pps. I'm going to make one more pass through the rdf issues list and
XGraph notes before/during the weekend, so suggestions/edits welcomed.

Received on Thursday, 7 September 2000 18:18:40 UTC