[
Lists Home |
Date Index |
Thread Index
]
On 12/9/05, Costello, Roger L. <costello@mitre.org> wrote:
> Hi Folks,
>
> A URI is used to "identify" a resource:
>
<snip>stuff which everyone has seen a dozen times by now</snnip>
Since I haven't seen it stated explicitly yet (I may have missed it),
let me throw another perspective at this: I find that I tend to treat
the portion of the URL prior to the query string as a class identifier
and the query string as an instance identifier. Thus:
org/stjude/cris/totxv/patient
might identify the class of patients enrolled on the TOTXV protocol
here at St. Jude. Whether this in itself returns something useful is
up to the implementer. In our case, if no specific instance is
identifiable, we might default to bringing up a search screen (or a
list) allowing one to find (or pick) an instance.
Given class identification, then:
org/stjude/cris/totxv/patient?patientId=123456
may return specifics on a given patient. In general, multiple query
parameters may be needed to completely qualify complex classes, but
the principle remains the same..
This viewpoint may be application specific, but I think it's pretty
generally applicable?
--
Peter Hunsberger
|