Re: Valid representations, canonical representations, and what the SW needs from the Web...

On Monday, February 3, 2003, at 01:05  AM, Miles Sabin wrote:
> Roy T. Fielding wrote,
>>> So if your thesis (which is a Resource which isn't on the web) is
>>> at http://.../thesis,
>>
>> If my thesis were at that URI, then it would be on the Web.
>
> How does that square with,
>
>> BTW, my dissertation doesn't exist within the Web any more than the
>> Sun exists within the Web.  Nevertheless, I could mint an http URI
>> to identify my dissertation if I wanted to and the system wouldn't
>> know the difference.
>
> given that a representation of your dissertation is accessible via,
>
>   http://www.ics.uci.edu/~fielding/pubs/dissertation/top.htm
>
> I'm not trying to pick nits, but these two statements strike me as
> inconsistent. Would you agree? In which case which of them would you
> reject?

I reject that the URI above directly identifies my dissertation.
It is specific to the top of the HTML edition, which means it is
good enough for indirect identification (since there is a N:1:1
relationship between that and my dissertation), but not direct.

> Or would you say that there isn't really an inconsistency here? I can
> see that you might argue that there's a difference between being "on
> the Web" and "existing within the Web", tho' I'd be interested to know
> what criteria such contrast would be based on.

I mean that I can mint a URI (I haven't yet) that does have a N:1
relationship with my dissertation (not being representation specific)
and then provide representations of it.  A URI with accessible
representations is how I define "on the Web".

Meanwhile, the HTML and PDF editions of my dissertation are on the Web.
I don't know if this is splitting hairs or not, but it is consistent.

....Roy

Received on Monday, 3 February 2003 15:08:28 UTC