[
Lists Home |
Date Index |
Thread Index
]
Hi Paul,
Paul said:
> The point is that query templates is not really a first-class part of
> the Web model. The Web model would encourage you to use XSLT templates
> on the client side as your "query templates." We have to pop up a few
> levels to the *real-world problem* you are trying to solve. Perhaps
> query templates is not the way to solve it on the Web.
Didier replies:
But its just another and more powerful way to get a document. You send a
template with placeholders included and the server sends back a document
with the placeholder replace with the required content. Its just another way
to obtain a document. The problem is that the actual web paradigm, at least
the one presented by REST, cannot include that. The real web, the web
million of developers are actually building allows such versatility. It
seems that GET and POST are perceived by these million of developers as two
way of getting a document. Maybe its the GET function which is badly adapted
to the needs. If that one would support a body to express more elaborate
queries (as long as there are no side effects on the server side) we would
have no problems with the POST function to be badly used. But you right,
maybe the web is far more limited than we thought it would be (at least for
its capacity to absorb the diversity of query languages). Relying on
everything on the server with dumb interpreter is not necessary the solution
and relying on simple URLs to query documents neither.
More and more I observe what's happening the more I see a chasm building
between the actual web reality and the web theory. Its the battle between
Bacon and Descartes between a priori and a posterori reasonning.
Cheers
Didier PH Martin
|
