Re: Datasets and contextual/temporal semantics

On Fri, Oct 14, 2011 at 7:08 AM, William Waites <wwaites@tardis.ed.ac.uk>wrote:

> On Thu, 13 Oct 2011 18:59:24 -0700, Jeremy Carroll <jeremy@topquadrant.com>
> said:
>
>    jeremy> On the WEB it is not possible to publish any information
>    jeremy> without also giving some time-information. The HTTP
>    jeremy> headers have it.
>
> I am very mildly uncomfortable tying up the interpretation of the
> document too tightly to the transmission protocol. This is partly the
> point of the range-14-esque,
>
>    <finger://river.styx.org/ww>
>    <finger://river.styx.org/ww.foaf>
>
> It might be a problem if code that doesn't "check the HTTP headers" is
> considered broken because more practically minded people might use USB
> sticks instead of the finger protcol to transmit big blobs of RDF and
> it should be perfectly adequate to embed all the necessary information
> in the document itself.
>
>
Sure -- if you embed all the necessary information in the document then your
data is properly decontextualized and the client doesn't need to look at the
HTTP headers.  I think most people here would agree that this is a good
thing.  But I think the point here is that if that context isn't part of the
data, webarch gives the client other means of discovering the context and
the client should take advantage of those if they're available.

-Alex



> Cheers,
> -w
>
>