Comments on RDFa in XHTML Syntax draft

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

Comments on RDFa in XHTML Syntax draft

Laurens Holst-2
Hi,

I have two comments on the current RDFa in XHTML Syntax draft [1]:

First of all, I think the @cite attribute on blockquote and q elements
should have the same status as @href (be considered as @resource), so
that RDFa can pick up on existing cite attribute annotations in
documents, and to avoid authors having to write the citation URI twice,
one for RDFa and one for HTML. See the example in section 6.3.2.1.

Second, are you sure that it is a good idea to give @src different
semantics than @href? I think it would probably be better if they had
the same semantics in RDFa. The reason for this is that I find the
difference between overriding @src and @href values (as explained in the
Primer [2] section 3.6) confusing, and it is not really clear to me why
they are treated differently. After all, they both reference a resource
in a very similar manner, only differing in the way the resource is
displayed.

The usefulness of treating @src as @about also seems to be rather
limited, as <img> can not have child content and you can thus specify at
most three triples; one with a literal object, one with a resource
object, and one reverse relation of the resource. Or, maybe @href should
also be treated as @about?

I hope these comments (although a few days late for the comments
deadline, I understand) will be useful for you to improve the specification.


~Grauw

[1] http://www.w3.org/TR/2008/WD-rdfa-syntax-20080221/
[2] http://www.w3.org/TR/2008/WD-xhtml-rdfa-primer-20080317/

--
Ushiko-san! Kimi wa doushite, Ushiko-san nan da!!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Laurens Holst, student, university of Utrecht, the Netherlands.
Website: www.grauw.nl. Backbase employee; www.backbase.com.


lholst.vcf (196 bytes) Download Attachment
smime.p7s (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Johannes Koch-2

Hi Laurens et al.

Laurens Holst schrieb:
> First of all, I think the @cite attribute on blockquote and q

... and ins and del ...

> elements
> should have the same status as @href (be considered as @resource), so
> that RDFa can pick up on existing cite attribute annotations in
> documents, and to avoid authors having to write the citation URI twice,
> one for RDFa and one for HTML. See the example in section 6.3.2.1.

An (IMHO) important difference between @href and @cite (or @longdesc)
references is that with @cite there's an implicit RDF predicate ("is
quoted from") (@longdesc: "has a long description at"), while with @href
there must be an explicit RDF predicate specified.

--
Johannes Koch
In te domine speravi; non confundar in aeternum.
                             (Te Deum, 4th cent.)

Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Mark Birbeck
In reply to this post by Laurens Holst-2

Hi Laurens,

>  I have two comments on the current RDFa in XHTML Syntax draft [1]:
>
>  First of all, I think the @cite attribute on blockquote and q elements
>  should have the same status as @href (be considered as @resource), so
>  that RDFa can pick up on existing cite attribute annotations in
>  documents, and to avoid authors having to write the citation URI twice,
>  one for RDFa and one for HTML. See the example in section 6.3.2.1.

Speaking for myself, I absolutely agree. :) However, I have never
pushed this in the current version of RDFa, because in some research
that I did for the IPTC a while ago, it looked like you could create a
general rule where:

  any attribute + value

is equivalent to:

 bnode + predicate based on attribute name + value

You'll see that @instanceof follows this pattern:

  @instanceof="t"

=>

  _:a rdf:type <t> .

A similar example would be @role:

  @role="r"

  _:a xhv:role <r> .

And I think @cite should behave the same way, since it is the citation
of a quote:

  <blockquote
   property="ab:cd"
   cite="http://www.example.org/source"
  >
    Some text.
  </blockquote>

Would give:

  _:a ab:cd "Some text." .
  _:a xhv:cite <http://www.example.org/source> .

Where ab:cd was some appropriate predicate for a quote.

However...

I'm explaining all of this, only to show that there would be much to
discuss in resolving this issue. And I think it would be too much to
hope to use @cite in a way that everyone would agree on, in this
version of RDFa. My suggestion would be to leave this until the next
version.


>  Second, are you sure that it is a good idea to give @src different
>  semantics than @href? I think it would probably be better if they had
>  the same semantics in RDFa. The reason for this is that I find the
>  difference between overriding @src and @href values (as explained in the
>  Primer [2] section 3.6) confusing, and it is not really clear to me why
>  they are treated differently. After all, they both reference a resource
>  in a very similar manner, only differing in the way the resource is
>  displayed.
>
>  The usefulness of treating @src as @about also seems to be rather
>  limited, as <img> can not have child content and you can thus specify at
>  most three triples; one with a literal object, one with a resource
>  object, and one reverse relation of the resource. Or, maybe @href should
>  also be treated as @about?

You are right that there is not a great deal of difference between
using @src as a subject or an object in the presence of @rel. However,
by making @src act as a subject, you are able to make use of an image
in some statement, and then say something about the image.

For example, you could say that someone has a picture, and that the
picture is licensed under some CC license:

  <div about="#me">
    <span property="foaf:name">Mark Birbeck</span>
    <span rel="foaf:depiction">
      <img src="pic.png" rel="license" resource="http://cc..."/ >
    </span>
  </div>

If you let @src play the role of an object, then you'll see that you
can only make one or other of these statements in a compact way, and
then you have to repeat the URI for the image, in order to indicate
the second statement. For example:

  <link about="pic.png" rel="license" href="http://cc..." />

  <div about="#me">
    <span property="foaf:name">Mark Birbeck</span>
    <img rel="foaf:depiction" src="pic.png" />
  </div>

(This is not an official response from the group, I'm merely letting
you know the thinking that lay behind this.)


>  I hope these comments (although a few days late for the comments
>  deadline, I understand) will be useful for you to improve the specification.

They have been useful, thank-you.

Regards,

Mark

--
  Mark Birbeck

  [hidden email] | +44 (0) 20 7689 9232
  http://www.x-port.net | http://internet-apps.blogspot.com

  x-port.net Ltd. is registered in England and Wales, number 03730711
  The registered office is at:

    2nd Floor
    Titchfield House
    69-85 Tabernacle Street
    London
    EC2A 4RR

Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Mark Birbeck
In reply to this post by Johannes Koch-2

Hi Johannes,

> An (IMHO) important difference between @href and @cite (or @longdesc)
>  references is that with @cite there's an implicit RDF predicate ("is
>  quoted from") (@longdesc: "has a long description at"), while with @href
>  there must be an explicit RDF predicate specified.

That's right. I'd also say that you'll find in nearly all situations
that there is also an implied bnode; these attributes tend to have
more in common with @instanceof, than @href or @resource.

Regards,

Mark

--
  Mark Birbeck

  [hidden email] | +44 (0) 20 7689 9232
  http://www.x-port.net | http://internet-apps.blogspot.com

  x-port.net Ltd. is registered in England and Wales, number 03730711
  The registered office is at:

    2nd Floor
    Titchfield House
    69-85 Tabernacle Street
    London
    EC2A 4RR

Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Laurens Holst-2
In reply to this post by Mark Birbeck
Mark Birbeck schreef:
> If you let @src play the role of an object, then you'll see that you
> can only make one or other of these statements in a compact way, and
> then you have to repeat the URI for the image, in order to indicate
> the second statement. For example:

I understand. Is there any particular reason then that @href isn’t
treated in the same way? It seems to me it has the same benefits there.

Either way, thanks for the clarifications.


~Grauw

--
Ushiko-san! Kimi wa doushite, Ushiko-san nan da!!
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Laurens Holst, student, university of Utrecht, the Netherlands.
Website: www.grauw.nl. Backbase employee; www.backbase.com.


lholst.vcf (196 bytes) Download Attachment
smime.p7s (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Ben Adida-2


Laurens,

Thanks for your feedback on these issues, it's much appreciated.

> I understand. Is there any particular reason then that @href isn’t
> treated in the same way? It seems to me it has the same benefits there.

Typically, an element with @href can have child elements, so that allows
authors to attach additional triples:

   <a rel="dc:creator" href="http://ben.adida.net/#me">
     <span property="foaf:name">Ben</span>
   </a>

leads to

   <> dc:creator <http://ben.adida.net/#me> .

   <http://ben.adida.net/#me> foaf:name "Ben" .

Also, if you use @href without @about, then it can complete triples
"from above":

   <div rel="foaf:knows">
     <a href="http://example.com/markbirbeck">Mark</a>,
     <a href="http://example.com/ivanherman">Ivan</a>
   </div>

leads to:

   <> foaf:knows <http://example.com/markbirbeck>
   <> foaf:knows <http://example.com/ivanherman>

So I think the important use cases are fulfilled here. If we allowed
@rel to "slip in between" @href and @resource, that would get quite
complicated (and it would prevent @resource from overriding @href in
many cases.)

-Ben

Reply | Threaded
Open this post in threaded view
|

Re: Comments on RDFa in XHTML Syntax draft

Mark Birbeck
In reply to this post by Laurens Holst-2

HI Laurens,

On 24/03/2008, Laurens Holst <[hidden email]> wrote:
> Mark Birbeck schreef:
>
> > If you let @src play the role of an object, then you'll see that you
>  > can only make one or other of these statements in a compact way, and
>  > then you have to repeat the URI for the image, in order to indicate
>  > the second statement. For example:
>
> I understand. Is there any particular reason then that @href isn't
>  treated in the same way? It seems to me it has the same benefits there.

The main difference with @href is that we're constrained by the fact
that we want to incrementally build upon HTML/XHTML, and @rel/@rev
with @href already has a meaning.

It's not something that is widely used, but @rel and @rev are
perfectly valid on the <a> tag. So the following is correct HTML,
regardless of RDFa:

  <a rel="next" href="next.html">chapter 2</a>

One of the main contributions of RDFa is not actually the syntax, but
that it looks at HTML and XHTML documents through the prism of RDF. So
the example just given could be interpreted as:

  <> xh:next <next.html> .

The same would go for @rel="stylesheet", @rel="first" and so on. (RDFa
adds some new values, like "license", too.)

RDFa enhances this already existing feature of HTML/XHTML, by allowing
a wider range of predicates to be set (by using CURIEs with prefixes)
and enabling the subject to be something other than the 'current
document' (by using @about). But these features are clearly building
on the metadata features already in HTML/XHTML.

Which is a long ways round way of saying that I'm afraid we can't
change @href in the way you suggest, because @href was already an
'object' in HTML, long before RDFa came along. :)

Regards,

Mark

--
  Mark Birbeck

  [hidden email] | +44 (0) 20 7689 9232
  http://www.x-port.net | http://internet-apps.blogspot.com

  x-port.net Ltd. is registered in England and Wales, number 03730711
  The registered office is at:

    2nd Floor
    Titchfield House
    69-85 Tabernacle Street
    London
    EC2A 4RR