[
Lists Home |
Date Index |
Thread Index
]
[Simon St. Laurent]
>... as a location within the space defined by www.amazon.com as
>seen through the perspective of HTTP. What we find at that location
>varies over time.
Its more than just time - all attributes of the observer (headers + body of
GET/POST for
example) - effect the resource retrieved. The act of observing in effect
de-quantizining
the URL. We cannot know the state of a resource without observing it and
the act
of observing it can change the resource.
The fact that URLs can be and regularly are "hooked" to introduce
resolution trickery or
abritrary cleverness makes the more abstract layers above URL redundant in
practice
but perhaps (just perhaps) intellectually useful - like hadrons and other
abstract
base classes.
URLs are just points in an information space. You can send/recieve information
to/from these points. The effect of such sending/receiving on the information
space is unbounded and unknowable.
Information retrieved from aURL s a function of an unbounded
state space consisting of all information in the request (headers + body) +
time.
Thats plenty abstract and plenty powerful enough for me.
As for URIs that are not URLs - they make my food repeat on me something
terrible.
The GET verb is the great leveller of the Web information space. I think it
should be
applicable to all URIs so that the information space is homogenously
observable at
a minimum.
If using URI rather than URL looses that homogenous observability, it had
better
provide significant benefits to make it worthwhile.
Sean
|