Re: What is CONTENT?

[Barron Housel <housel@cisco.com>]

Hilarie raises an interesting question: Does the "Direction System" need to
parse the http headers to get more information in order to direct the request to
a correct surrogate?
Barron

Hilarie Orman wrote:

> Well, the issue is not exactly "what is content" but "what is the
> relationship between a URI and the content?"  And there is a further
> issue of the mapping between URI's and sets of related content.
> I believe that none of these questions are answerable at the
> current time.  Consider the following questions:
>
> It appears that a URI has two parts - a name and extra stuff (metadata).
> The combination yields a content response.  It's a many-to-one mapping
> at best.
>
> Sometimes, we think, there is a functional relationship between two
> sets of content:
>    f(content-A, metadata1) = content-B
>    f(content-A, metadata2) = content-C
>
> We can ask, is there a URI that produces content-B as the response?  Is it
> name(content-A)||metadata1?  Must there be?
>
> Alternatively, we can ask, if
>   URI1 = somestring||metadata1  and produces the response content-A
>   UR2 = somestring||metadata2  and produces the response content-B
> then, can we assume that there is some well-known function f, such that
>      f(content-A, metadata1) = content-B
>      f(content-A, metadata2) = content-C
>
> >>> "Iacovou, Danny" <danny.iacovou@ebenx.com> 11/10/00 04:46PM >>>
> > -----Original Message-----
> > From: Hilarie Orman [mailto:HORMAN@novell.com]
>
> > It's just an encoding of application-level data, isn't it?
> > Of course it's digital, this is the Internet; it's most
> > probably octets, but
> > that's not a particularly helpful part of the definition;
> > same for "ordered".
> > The type can be specified in the encoding; that's normal.
>
>   Well, at least one other person expressed concern that there doesn't seem
>   to be a way of associating meta-data with the "content" given some of the
>   definitions of content that have been floating around.
>
>   It could very well be that associating meta-data and preserving it during
>   "transport" is do'able as things stand right now, but I haven't figured it
> out,
>   and nobody has explained it to me.
>
>   It could also very well be that meta-data isn't of importance. But nobody
> has
>   said that either. So I'am assuming that Alex French, and myself, have some
>   point in raising this issue.
>
>   Right now we all seem to agree that content is something that can be
> pointed to
>   via a URI. I think we should figure out exactly what CONTENT is going to
> be, what
>   we are going to with the meta-data, and then, perhaps, ask the question of
> how to
>   treat "multiple-views" of an item (which is actually different CONTENT per
> view
>   since each view will have its own URI - but to know the grouping of the
> views and
>   which URIs can be grouped together would be very useful to know).