[
Lists Home |
Date Index |
Thread Index
]
Hi John,
John said:
> This has nothing to do with GET vs. POST. If you fetch a resource by
> GET, there is nothing to prevent the server from composing the entity
> body on the spot. You are merely guaranteed (architecturally speaking)
> that no matter how many times you repeat the request, you get the same
> *resource* (not to be confused with the same entity body, as in the
> case of a time-of-day server); the resource is not side-effected by the
> GET operation.
Didier replies:
You're right I wasn't clear enough. What I meant is that the XSLT workgroup
didn't considered what's happening on the development of new ways to access
documents, notably the fact that relational database manufacturers now offer
the capability to access XML document through SQL queries encoded in XML
POST requests. The workgroup also didn't considered new developments on the
side of distributed authoring and versioning (i.e. WebDAV) and internet
telephony (i.e. SIP). They missed, probably based on good reasons (if I
could only know what they are) the opportunity to propose a general way to
access internet messages (i.e. HTTP, SIP, WEBDAV and TuttiQuanti). The
workgroup seem that have limited itself to static or dynamically created
document accessible through the HTTP GET method (yes I know, with any valid
URI too).
You can off course have at the other end a script overriding the HTTP GET
method and dynamically create the document. Therefore, the document can be
dynamically created. Nonetheless, we still have the problem of not being
able to access an Oracle document (i.e. a valid XML document) from an XSLT
template since this request is using a POST method. If the same request
would be encoded in a URI then it would imply that we can include XML
documents in URIs or that we found a way as obvious and readable to encode
the same query in an other kind of encoding. If yes, how the URI would look
like?
cheers
Didier PH Martin
|
