[
Lists Home |
Date Index |
Thread Index
]
Didier,
On Tue, Apr 23, 2002 at 02:22:07PM -0400, Didier PH Martin wrote:
> Didier replies:
> Good point. Thus, the whole issue though is to create a new URI that allows
> us to encode the queries.
More or less, yes. But I like to focus on "identifying resources"
rather than "encoding queries". Think of a query as identifying
something. For example, this query identifies all resources that
Google knows about that includes your name.
http://www.google.com/search?q=didier+martin
> If we say that
> only the GET method can be used to fetch documents, then we have to create a
> new URI.
Yes!! This is a Good Thing.
> This is why millions of document fetching or dynamically created
> documents are using the HTTP POST method. Simply because it is limited and
> suffer from several usage breakdowns. Abstract URI concepts do not resolve
> the issue. We have to propose a concrete one. Suggestions?
I think the creation of new URIs is a wonderful side effect (no pun
intended 8-) of having a single "retrieve" method. Looking at
PROPFIND again, if the "properties for a resource" were identified by a
separate URI rather than retrieved via a non-GET method, then we could
make assertions about the properties, cache them, use XSLT on them, etc.
All the goodies you get by having a URI.
The "U" in URI stands for "uniform" (previously "universal"), which
means that a URI identifies the same "thing" in all contexts. Having
a single method meaning "show we what you are"/"resolve"/etc.. helps
ensure that uniformity/universality is maintained.
MB
--
Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA. mbaker@planetfred.com
http://www.markbaker.ca http://www.planetfred.com
|
