[Twisted-web] HTTP-AUTH for web2 / Kudos on web2's operation

Clark C. Evans cce at clarkevans.com
Sat Nov 19 20:01:57 MST 2005

On Sat, Nov 19, 2005 at 07:47:23PM -0500, Stephen Waterbury wrote:
| >On Fri, Nov 18, 2005 at 05:54:19PM -0600, L. Daniel Burr wrote:
| >| Every Resource *should* be identified by a URL, and that URL *should*  
| >| refer to one logical Resource.  This is how the web works today; I for 
| >| one, hope that it continues to do so.
| >
| >It should be noted that Felding and most others do not require that the
| >relation from URIs onto Resources be injective -- that is, two different
| >URIs may (and often do) refer to the same Resource. The object
| >publishing methodology seems to imply (or assume?) an injective relation
| >between URIs and Resources.  This is not a necessary assumption.
| Actually, what Clark says here does not contradict what L. Daniel said.

I certainly was not contradicting L. Daniel Burr, in fact, I agree with
his position right up-to where he seems to assert that object publication 
systems support REST, while a request handling mechanism doesn't.  

| L. Daniel said "Every Resource should be identified by a URL" -- he
| did not say "one and only one URL"; therefore, his words did not say
| injectivity was a requirement. 

L. Daniel Burr did not say "one or more URLs" either.  He was vague in
his declaration; and I was hoping to respectfully clarify the definition
of Resource without causing a ruckus.

| Hence, it is needless for Clark to say that no one requires it.  There
| was no such claim.

Actually, there is a need to make this clarification.  Most object
publication systems assume that each Resource object has a single parent
(that is, that Resources form a tree).   This is a very tempting
assumption to make, but there is nothing in the web architecture which
requires that a Resource have exactly one URI. 

Not that I think this restriction is unreasonable in many contexts. I
don't mind if a slightly-higher-level object publishing system is built
on-top of low-level primitives. However, I don't think that a system
like Twisted should be making this assumption at its most basic level.
The trunk does not yet make this assumption: I'd like to keep it that way.

Kind Regards,


More information about the Twisted-web mailing list