- From: Tim Berners-Lee <timbl@w3.org>
- Date: Mon, 10 Feb 2003 10:23:46 -0800
- To: Dan Connolly <connolly@w3.org>
- Cc: www-archive@w3.org
On Wednesday, Jan 29, 2003, at 10:22 US/Eastern, Dan Connolly wrote:
> On Mon, 2003-01-27 at 17:23, Tim Berners-Lee wrote:
>> On Friday, Jan 24, 2003, at 09:45 US/Eastern, Dan Connolly wrote:
>>
>>> offlist, copy to www-archive...
>>>
>>> On Fri, 2003-01-24 at 08:23, Tim Berners-Lee wrote:
>>> [...]
>>>> If you like, it is as though there is an axiom
>>>>
>>>> { ?x log:uri ?u. ?u string:match "^http://[^#]$" } => { ?x
>>>> rdf:type
>>>> doc:Work }.
>>>>
>>>> (where string:match is a regexp matcher)
>>>> This axioms comes from the URI spec and the specs it references.
>>>
>>> ???
>>>
>>> Which section of the URI spec does that come from?
>>>
>>
>> The URI spec calls out the MIEM type registry,
>> The MIME type registry calls out the HTTP spec.
>
> Huh? MIME type registry calls out the HTTP spec?
> Perhaps you meant that the URI spec calls out
> the URI scheme registry, which calls out the
> HTTP spec.
Yes.
>
> But actually, the URI spec, RFC 2396, says
> nothing about IANA nor a registry.
>
> Perhaps you meant to include RFC2717 in "the URI spec".
> http://www.ietf.org/rfc/rfc2717.txt
>
Yes, I guess so.
> But you're asking a lot of your readers when you
> leave all this implicit.
>
Yes. Actually the fact that RFC2717 isn't referenced from RFC2396
could be a bug.
I should certianly mention it.
> Anyway, all this is minor compared to...
>
>> The HTTP spec says it identifies a network information object.
>
> ... which begs the question again: which part of the HTTP
> spec says that? Ah; that has a reasonably straightforward
> answer:
>
> The "http" scheme is used to locate network
> resources via the HTTP protocol.
> -- http://www.w3.org/Protocols/rfc2616/rfc2616-sec3.html#sec3.2.2
>
> But it's not at all clear that 'network resource'
> is a synonym for doc:Work.
>
No, that is the weak point over which we are all arguing.
> OK, I think I see your argument now, but you're asking
> a lot of your readers when you leave all this
> implicit.
>
Very true.
>
>>> This sort of thing totally undermines your argument,
>>> I think.
>>>
>>
>> It is difficult to make the point about the URI spec constraining
>> models in OWL without being concrete, and its difficult to
>> be concrete when there is no formalization in the web specs.
>>
>> What would you have suggested as an alternative?
>
> I suggest you slow down and explain the chain of reasoning.
>
> I have a lot of shared context with you, and I read your
> message as saying
>
> <rfc2396> log:includes {
> { ?x log:uri ?u.
> ?u string:match "^http://[^#]$" } =>
> { ?x rdf:type doc:Work } }.
>
> which is plainly false.
>
> The actual argument is much more subtle.
>
> Hmm... might be fun to formalize the actual argument
> and have Euler spit out the proof.
>
>> Tim
>>
>>>> Any semantic web engine can conclude it. It is not authorized
>>>> by the OWL spec, it is authorized by the URI spec.
>>>
>>> --
>>> Dan Connolly, W3C http://www.w3.org/People/Connolly/
>>>
> --
> Dan Connolly, W3C http://www.w3.org/People/Connolly/
>
Received on Monday, 10 February 2003 16:47:46 UTC