Re: "tdb" and "duri" URI schemes...

On Mon, Feb 7, 2011 at 9:56 PM, Larry Masinter <masinter@adobe.com> wrote:
> I don't think I meant anything odd by" use cases. In almost all of the
> examples I can think of,

It would help me if you described, in engineering terms, some of the
examples or use cases you are thinking of. As it is I can't tell where
I would use one of these URIs.

> tdb needs a date to be meaningful because the
> lifetime of the thing described is different than the lifetime of the
> description. Maybe this is the design principle? That a well_designed uri
> scheme has the lifetime of utility of the Uri to resource mapping should
> match the lifetime of the resource....

No, it makes perfect sense to refer to things that no longer exist or
that don't yet exist. The idea only applies to the documentation the
URI reader will need to glean "meaning". And the scheme designer can't
be solely responsible for preserving this documentation; the URI
writer bears some of the burden as well, e.g. in checking to make sure
that the documentation is archived somewhere.

Jonathan

> Connected by DROID on Verizon Wireless
>
> -----Original message-----
>
> From: Jonathan Rees <jar@creativecommons.org>
> To: Larry Masinter <masinter@adobe.com>
> Cc: "www-tag@w3.org" <www-tag@w3.org>
> Sent: Thu, Feb 3, 2011 13:14:20 GMT+00:00
> Subject: Re: "tdb" and "duri" URI schemes...
>
> These fixes all look fine to me.  Just one question...
>
> On Wed, Feb 2, 2011 at 10:25 PM, Larry Masinter <masinter@adobe.com> wrote:
> ...
>> But I think it's a false economy to leave the date out, since most
>> of the use cases for tdb that I can think of need a timestamp, because
>> the lifetime of the resource described has nothing in common with the
>> lifetime of the description.
>>
>> Anyway, I'm resisting. I admit that leaving the date out of tdb
>> is feasible, because one could combine 'tdb' and 'duri', but I think
>> of the two choices, I like the one I have since it matches better
>> what I think the common use cases would be.
>
> Can you tell me what a "common use case" might be for each of the two
> schemes?  That is, what kind of agent would want to write down one of
> these URIs, as part of what extant data format, and what would another
> agent do when encountering one?
>
> I have been assuming you aren't talking about anything related to RDF
> or HTML or the needs of their user communities, and that you're mainly
> presenting the specification mostly as a discussion aid so that when
> someone says "duri" everyone knows what's being discussed. But now
> that you're talking about use cases I'm not so sure.
>
> Thanks
> Jonathan
>

Received on Tuesday, 8 February 2011 13:40:05 UTC