On Jun 22, 2007, at 2:22 PM, Mark Birbeck wrote:

I am strongly against using URIs to identify resources...

I think that's a discussion for another list. :)

all right, let's not discuss it further then.

In other words, if you believe that we shouldn't take anything from
the structure of the document, then we really should spell everything
out.

It's not what I meant, we should make use of the structure of the document, but we shouldn't infer too much, when you say that :

<span about="http://whatever.org";> food book</span>

corresponds to the triple

<http://whatever.org> rdfs:label "food book".

it is an interpretation, even worse an automatic interpretation which by definition will be wrong in many cases or if not wrong, will be different than what the developer/publisher meant. Moreover, the piece of information which semantics were not explicitly defined by the publisher does not seem to be of interest, it might be more important to automatically infer the triple :

<> rdf:li <http://whatever.org>

or  <> dcel:relation <http://whatever.org>

as the only thing that was said by the publisher is that this resource is related to whatever :-)

As it happens, in my view rdfs:label is *exactly* right for this. The
definition of rdfs:label says nothing more than that rdfs:label is a
human readable label for a resource. I really don't think a big leap
in logic is being if we were to assume that an author who tags some
text with @about is making exactly that statement.


The question seems not to be what is the minimum markup that should give a
triple but more :

 what is the minimum markup to generate the maximum triples...

They are two separate issues. The 'what is the minimum?' question
concerns trying to make things easy for authors to add useful
information. So if I have a blog and I mention a couple of books and a
company, should I need to specify more than an @about on a <span>? I
think it would be great for ease of use and adoption of RDFa if it was
not necessary. However, adding a @class value is not so much extra
work, and that's good enough for now.

But in terms of 'minimum mark-up, maximum triples' then all of your
examples that follow are perfectly fine. How far you go with packing
lots of attributes onto an element depends only on whether the mark-up
will ever be read by a human or not. :) Particularly your last
example, since a human reader might not immediately spot that some of
the attributes are attached to the bnode. But like I say, it's
certainly all correct. :)


Can we have :

<span about="http://amazon.com/ISBN:0091808189";
property="rdfs:label" class="bib:book">
 Canteen Cuisine
</span>

which would give:

<http://amazon.com/ISBN:0091808189> rdf:type bib:book.
<http://amazon.com/ISBN:0091808189> rdfs:label "Canteen
Cuisine".

and to go back to my example with Ben's solution:

<div class="foaf:Person" rel="rdf:li">
 some things about the person
</div>

which gives the triples :

<> rdf:li _:whatever.
_:whatever rdf:type foaf:Person.

Can we go further than that and do :

<div class="foaf:Person" rel="rdf:li" property="foaf:name">
 Cédric Mesnage
</div>

or even :

<div class="foaf:Person" rel="rdf:li" property="foaf:name" content="Cédric
Mesnage"/>

and get:

<> rdf:li _:whatever.
_:whatever rdf:type foaf:Person.
_:whatever foaf:name "Cédric Mesnage".



---
Cédric Mesnage
PhD Student
[EMAIL PROTECTED]
http://www.cedricmesnage.org
http://blog.cedricmesnage.org/



On Jun 21, 2007, at 12:10 PM, Mark Birbeck wrote:

Hi Ben/Cédric,

[I'm not proposing a resolution to this question in this version of
RDFa, but I think it's useful to collect use-cases.]

I had a use-case the other day that is related to the ones you are
describing. Essentially the question we all seem to be converging on
is what is the minimum amount of mark-up that should give us a triple?

So, the following feels quite natural, as a way of marking up the
mention of something like a book in my blog:

 Today I bought a copy of
 <span about="urn:ISBN:0091808189" class="bib:book">
   Canteen Cuisine
 </span>
 from my local bookshop.

Since my system uses the URI to retrieve some data about the book from
a book site like Amazon, I don't actually need any further triples
like title, price, publisher, author, or whatever. But there is a an
interesting question as to whether the following should be enough to
get an entry in the triple store:

 <span about="urn:ISBN:0091808189">
   Canteen Cuisine
 </span>

The system could still do the same thing, and retrieve additional
triples based on the resource, but the question is what are the
parsing rules that get from this mark-up to a triple?

The only way I can think of to achieve this from the mark-up I've
shown is be automatically generate labels from the content of
elements. The mark-up would therefore generate this:

 <urn:ISBN:0091808189> rdfs:label "Canteen Cuisine" .

and we would now have the URI for the book in our triple-store, and
can make use of it to retrieve further information.

Regards,

Mark

On 13/06/07, Ben Adida <[EMAIL PROTECTED]> wrote:


Cedric,

This is an interesting question. I had to deal with this with the RDFa
clipboard [1], and I chose to use the predicate rel="rdf:li" on any
bnode I wanted to appear on the page, effectively saying "this bnode is
an item of the current page." For example, in your code below:

<span class="foaf:Person" rel="rdf:li">
some things about the person
</span>

which yields:

<> rdf:li
     [a foaf:Person; ...things about the person...]

I'm pretty sure this is not a "best practice", but it's the work- around I came up with for precisely this issue, and it's not all that wrong in
terms of semantics: after all, that *is* an item on the page.

-Ben

Cédric Mesnage wrote:
> Hi all,
>
> I have a question regarding blank nodes in RDFa, I don't know if the > issue has been raised already and I apologize if it has. In the RDFa > Primer I saw that you can create unnamed blank nodes using the 'rel'
> attribute as in the example:
>
> <dl class="foaf:Person" about="#card" id="card">
> ...
>  <dt>Address</dt>
>  <dd rel="foaf:address">
>   <span property="foaf:address_line_1">77 Massachusetts
Ave.</span><br />
>     <span property="foaf:address_line_2">MIT Room
32-G524</span><br />
>   <span property="foaf:city">Cambridge</span> MA 02139<br
/>
>   <span property="foaf:country">USA</span>
>  </dd>
> ...
> </dl>
>
> This works for predicates layered in an instance definition, do you plan
> having a similar principle for classes? I'd like to have:
>
> <span class="foaf:Person" >
> some things about the person
> </span>
>
> to be considered as a blank node, currently in RDFa On
> Rails(http://rdfa.rubyforge.org/) I generate blank node names this way:
>
> <span class="foaf:Person" about="#BNode1">
> some things about the person
> </span>
>
> incrementing the number through the page generation, but this is ugly. > The other solution is that I can just forbidden the use of classes if no
> uri or explicit blank node name is given.
>
> Hope this does make some sense and look forward to get you point of view.
>
> Best Regards!
> ---
> Cédric Mesnage
> PhD Student
> [EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]>
> http://www.inf.unisi.ch/phd/mesnage/
> http://myunderstanding.wordpress.com/
>
>







--
 Mark Birbeck, formsPlayer

 [EMAIL PROTECTED] | +44 (0) 20 7689 9232
 http://www.formsPlayer.com |
http://internet-apps.blogspot.com

 standards. innovation.



--
 Mark Birbeck, formsPlayer

 [EMAIL PROTECTED] | +44 (0) 20 7689 9232
 http://www.formsPlayer.com | http://internet-apps.blogspot.com

 standards. innovation.


Reply via email to