Re: Uniform access to descriptions

Xiaoshu Wang wrote:
>> One can make many useful statements about an IR as identified by a URI.
>   
> I was asking you *exactly* what you mean by many useful statement, which
> you can do before and cannot do now?

I might record my opinion of it or the time when I first saw it. 
Archive.org might say when a representation of it was first archived by 
the service they provide.  I might simply say of it that it is a diagram, 
as I mentioned before, or that it is available in different formats via 
conneg.  I might bookmark it and associate it with certain tags, as social 
bookmarking sites already allow with some limitations.

I might even publish a list of all the Web pages I visited today, by 
making, about each of them, a similar assertion.  That would enable the 
kind of thing people are doing today with centralized services.

All of these would be perfectly reasonable things to say about an IR, 
would they not?

All of these seem like they can easily be done with simple browser 
widgets that simply publish RDF triples from my personal Web space.

Many of those triples would turn out to be nonsensical if the URIs 
is question in fact identify, say, a gene, or the moon.

> Did my reinterpretation of the web
> architecture prevents you from doing that?

I do not yet fully know.

It may not prevent it, but it at least changes the way I must go about 
it, and I think the change is a move away from the way things are now 
and away from (at least what I have seen as) some of the early promise 
of the Semantic Web.

> I have no idea what you have
> specific in mind.  Can you find one concrete example since you said "
> *this* (- what is it) has been well covered in the existing Semantic Web
> literature"?

I thought scenarios such as the above had been widely known as far back 
as 2001 or earlier, and that these kinds of use-cases are generally 
regarded as part of what the Semantic Web is meant to enable.

Michaeljohn

Received on Monday, 14 April 2008 01:45:49 UTC