Re: sketch of an exposition (time)

On Mon, 2010-05-24 at 08:39 -0500, Dan Connolly wrote:
> On Thu, 2010-05-20 at 01:15 -0500, Pat Hayes wrote:
> > On May 17, 2010, at 4:40 PM, Jonathan Rees wrote:
> [...]
> > > Note on time
> > >
> > >   Although W is time-sensitive, we'll ignore time as it is not
> > >   helpful to account for it right now.  later we'll redo the
> > >   treatment to take time into account.
> > >
> > >   So W is OK as a binary relation for now.  Later it might be
> > >   W(X,R,t).
> > 
> > Or it can still be binary, but X can be timesliced:  W(s(X, t), R). I  
> > would recommend this as a stronger (more expressive) way to deal with  
> > time.
> 
> Nifty. Yes, I prefer that approach too.

This looks good for representing time, but I have two comments:

 - A URI-resource binding can change occasionally over time.  For
example, the domain name may be sold and the URI now denotes a
completely different resource than it did before.  Although this *could*
be modeled as above, I think it is important not to confuse this
situation from the situation in which a URI denotes a dynamically
updated web page on the current weather in Oaxaca, as they are
qualitatively different.  For the case where a domain name is sold and
URIs are re-used for completely different things, I think it makes the
most sense to model that as the URI being bound to a different resource,
rather than modeling it as one big resource that is the union of all of
the things over time that the URI is ever used to denote.

 - The timesliced approach still does not yet taking into account the
GET request, which is the other parameter that is used to determine
which w:Representation should be associated with the resource at a
particular time.  Once you add that parameter, it changes W from merely
being a relation to being functional:

  W: Resource x Time x Request -> w:Representation

which makes W equivalent to the function that takes any ftrr:IR 
http://lists.w3.org/Archives/Public/public-awwsw/2008Apr/0046.html 
and applies it:

  For any Time t, Request req and ftrr:IR ir,
    W(ir, t, req) = ir(t, req)



-- 
David Booth, Ph.D.
Cleveland Clinic (contractor)

Opinions expressed herein are those of the author and do not necessarily
reflect those of Cleveland Clinic.

Received on Monday, 24 May 2010 17:06:04 UTC