[Linking-open-data] Eurostat data about countries and regions available as L
Frederick Giasson writes on 2007-06-25
in response to
[Linking-open-data] Eurostat data about countries and regions available as L (Richard Cyganiak)
:
Hi Richard,
Thanks for this answer, it open up a full of new point of views,
discussions :)
> I'm a bit late to reply to this ... In short, Fred, I think your view
> contradicts the relevant standards. The assertion you disagree with
> follows logically is exactly in line with Web architecture as
> formulated by the W3C's Technical Architecture Group and TBL. See for
> example here:
>
> http://www.w3.org/2001/tag/doc/httpRange-14/2007-05-31/HttpRange-14
>
From this document:
HTTP itself provides one means of representing associations between
resources. Instead of returning a representation, when the URI of a
non-information resource is dereferenced, it is possible to return the
HTTP response code 303 (known as 'see other'). This indicates that there
is other, related information available concerning the URI that was
dereferenced. In addition to the code, *the response includes the URI of
the related resource*. This mechanism provides a way to draw attention
to the related information without the need to return a representation,
which might mislead the person or system making the request.
Importantly, the specification also states that "The new URI is not a
substitute reference for the originally requested resource." In other
words, responses containing this code *direct us to related material*.
This is exactly what is wrong in my vision of things: it shouldn't point
to *Another* resource.
And what introduce the difference between our two visions is the
"non-information resource"; that is a non sense in my vision. such
"non-information resource" is an abomination (okay, the word is possibly
a little bit forced); however, each resource has information. The only
"non-information" resource that could exist is an unlinked node and such
a node is sort of non-existant in the semweb since it can't be traversed.
Kingsley already said: Information is Data in Context. I agree. But the
thing is that each piece of data *has* some type, some sort, of
information, small or big, but it has. Informative things can be
processed out of *any* data. (this is why Kingsley say that: Information
is Data in Context). So, the same data can have different information
depending how it is processed, and in which *context*.
This is why we are starting to talk about *Data Spaces*. And URIs are
entry point (via HTTP or any other protocol/network) to resources of
this Data Spaces.
In fact, a single *Data Source* can be described, communicated, in many
ways: RDF, HTML, Video, Sound, etc.
Each of these mediums will convey some information, depending on the
context, the listener, etc. However, all the *information formats* will
come from the same source of *data*.
This said, it is for that exact reason why I see a *non-information
resource* as an abomination, and that one URI shouldn't be redirected to
another one, that act as *related material*. Both are the same, and
should be accessible via the same URI.
is this vision of things clearer now?
> (or, a bit further down in the Wikipedia article you quote:)
>
> http://en.wikipedia.org/wiki/Resource_%28Web%
> 29#Using_HTTP_URIs_to_identify_abstract_resources
>
> Or just read the abstract of this TBL piece:
>
> http://www.w3.org/DesignIssues/HTTP-URI2.html
>
> This debate comes up again and again, mostly because there's a lack
> of easily accessible material about how URIs and the Web work,
> especially with the changes brought about by the httpRange-14
> compromise. The “Cool URIs” paper you quote is an attempt by myself
> and some colleagues to explain these issues as clearly and accessible
> as possible. It has been reviewed by the TAG and has been confirmed
> to be in line with the relevant TAG recommendations. (The TAG has
> criticized some of our choice of terminology, which we will address
> in a revised edition.)
>
> In summary: URIs are not entry points to data sources. They are names
> for resources. There are two kinds of resources: information
> resources, which are document-like in nature and can be represented
> as a stream of bytes, such as an HTML page or an RDF document, that
> can be accessed by dereferencing the URI; an non-information
> resources, such as people, chemical compounds, and OWL classes, which
> cannot be dereferenced over the Web, but which can be described by a
> secondary information resource whose URI can be found by fragement
>
> That's the Web.
>
>
Well yeah, I agree. However here I am talking about the Semantic Web...
err... no... the Semantic Data Spaces; when you are talking about the
Semantic *Web*. This is possibly the main distinction between our visions.
Take care,
Fred
