OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] Meta-somethingorother (was the semantic webmega-permathrea

[ Lists Home | Date Index | Thread Index ]

Hi Joshua

> Well, for purposes of "semantic web", we are not talking of "data about
> data", we are talking about "data about resources".  In fact, even if
> you define some RDF data about a particular URL, and the URL is GETable,
> the RDF data is *not* "about" the data returned by the GET call.  It is
> about the resource for which the data serves as a representation.  So
> it's just fundamentally wrong, in a really painful way, to say that RDF
> is about data which can be retrieved through a GET.

I agree and this is not what said or meant.

> I noticed you postulated a controversy earlier in the thread ("some
> people mean 'data about data' and others mean 'data about resources').
> I wonder where you got that controversy.  There is no ambiguity;
> "semantic web is 'data about resources'.  If you have found someone who
> thinks that semantic web is 'data about data', they are completely
> confused.  There is no point in arguing about it.

Yes because I have a hard time to know when people are talking about
meta-data meaning in fact that they are talking about a resource that may be
existing or not as long as it referred by a URI and they really mean
meta-data. If you tell me that people are confused about the term meta-data
and data or are making distraction mistake when they mention the word
meta-data but not really meaning meta-data. I simply need to know the
context and not put a de facto filter to what they say and infer my own
meanings on what they say doesn't it?

> If you are instead arguing that we shouldn't use the phrase "metadata"
> to talk about 'data about resources', then maybe you have a point.  It
> doesn't bother me, but if people get so sidetracked than maybe it's the
> fault of imprecise language.

Yes this is precisely the problem, when we talk about the RDF framework
(let's get explicit here) we are facing two cases:
a) RDF is used to add some meta-data to some existing data (also called a
resource). The data may not necessarily be accessible thought an HTTP GET
but nonetheless, the RDF description is about some data. This world has its
own set of problems and opportunities. For instance, if WinFS publishes some
RDF statements about files I have on my computer, then the context is about
data that I can obtain through different means not only and not necessarily
through HTTP. The important point here is that what is said about that data
is related to that data.
b) RDF is used to provide some data about a real or virtal object. For
example a machine or the aleph concept (Cantorian concept about the infinite
and cardinality) as long as both get a URI to name them, characterize or
locate them (URN, URC or URL). Again this context will to different problems
and opportunities.

At the very abstract level RDF allows us to define a set of properties (call
them triple as you whish). But the operational problems may be different if
I manipulate data or meta-data even if they are all encoded using the same
syntax. Sharing the same syntax doesn't mean they share the same space or
that they share the same problem space. Don't get me wrong, I do not say
that "cannot" share the same space, only that they won't "necessarily" share
the same space. Or said differently, most of the time, it is "unlikely" that
they will share the same space. The operational issues may be different. So
knowing if we talk about meta-data or data has some importance. If there are
no ways to make such distinctions, then... do you call that progress doc?
(with a bugs bunny voice).

Didier PH Martin


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS