Re: A REST challenge

"Champion, Mike" wrote:
> 
>...
> 
> How does the buyer find the necessary information to put in the order
> document?  

Whether using REST or SOAP, they need to agree on a vocabulary a priori.
REST does not standardize nouns. That's why we have XML Schema, RDF,
XSLT, ebXML, registry/repositories etc. REST standardizes verbs. The
virtue of REST is that we don't need "verb schemas", "verb metadata",
"verb transformation languages", "verb registry/repositories", etc. We
can save ourselves building a massive redundant infrastructure.

> ... The "to-ind and fro-ing" presumably refers to various lookups of
> stock numbers, prices, inventory levels.  Of course, you could do that with
> HTTP queries, but this gets into some deep waters of REST, and the
> mechanical issue of formatting the URIs to do the queries will probably make
> the buyer long for WSDL.  

You're exaggerating. Any decent URI (=Web) toolkit does this formatting
for you.

> .... Of course, there's no profound conflict between
> REST and WSDL ...

Not entirely true but for the purposes of this discussion we can pretend
it is...

> So, one way or the other, you are going to have to expose some "methods"
> other than the REST/CRUD ones to perform these lookups.  No? 

All lookups are "GET"! Otherwise how do Google and Yahoo work? If you
believe that hidden in this use case there is a query that cannot be
turned into a simple GET then please describe it.

But I'll point out again that the substructure of a query is a noun. In
"SELECT X from Y" X and Y are nouns. REST does not standardize nouns. It
isn't a magic wand that makes *all* interoperability problems go away.
Just some: those involving unstandardized verbs and addressing.
-- 
Come discuss XML and REST web services at:
  Open Source Conference: July 22-26, 2002, conferences.oreillynet.com
  Extreme Markup: Aug 4-9, 2002,  www.extrememarkup.com/extreme/

Received on Tuesday, 16 July 2002 16:05:24 UTC