Re: Linotype & rdf/a: textual properties introduction

From: Paolo Ciccarese <paolo.ciccarese_at_gmail.com>
Date: Mon, 14 Nov 2005 22:35:25 +0100

On 11/13/05, Stefano Mazzocchi <stefanom_at_mit.edu> wrote:
> Paolo Ciccarese wrote:
> > The first step for semantic annotation of content in Linotype (if you
> > want to know
> > more about Linotype see the note below) has been the introduction of
> > "qualifying links":
> > <a
> > href="http://creativecommons.org/licenses/by-nd/1.0/"
> > rel="cc-license">Creative Common License</a>
> >
> > That leads to the following triplet:
> > <> <cc-license> <http://creativecommons.org/licenses/by-nd/1.0/>
> >
> > Through a prompt,it is possible to define the "qualifying links" and of
> > course the
> > "simple links".
> >
> > Current prompt accepts:
> > - predicate (link type --> rel)
> > - object (link --> href)
> >
> > But rdf/a allows also to define "textual properties":
> >
> > <span property="foaf:name">Joe Lambda</span>
> >
> > This leads to the following triplet:
> > <> <foaf:name> <"Joe Lambda">
>
> wait, do you mean
>
> <> <foaf:name> "Joe Lambda"
>
> ?

Yes <> <foaf:name> "Joe Lambda"

>
> > and qualifying Links plus property
> > <a
> > href="http://www.acme.com"
> > rel="foaf:workplaceHomepage"
> > property="foaf:work">ACME Inc.</a>.
> >
> > that leads to:
> > <> <foaf:workplaceHomepage> <http://www.acme.com>
> > <> <http://www.acme.com> <"ACME Inc.">
>
> Hmmm, not sure I like this, I mean what is the benefit of having
>
> <> <http://www.acme.com> "ACME Inc."
>
No it was another mistake the triplets should be
<> <foaf:workplaceHomepage> <http://www.acme.com>
<> <foaf:work> "ACME Inc."

Think about a paragraph like

<p> The first chairman was Gustav Klimt. </p>

I could annotate "Gustav Klimt" as:

<p> The first chairman was <a
  href="http://www.artisthomepage.com/klimt"
  rel="art:artistHomepage"
  property="art:artistFullName">Gustav Klimt</a>. </p>

This leads to
<> <art:artistHomepage> <http://www.artisthomepage.com/klimt>
<> <art:artistFullName> "Gustav Klimt"

And honestly, I find it useful to highlight the fact that the artist
full name has been used to generate the link to is homepage. In fact,
you can have different other examples. One can be (ontology is just an example):

<p> My favourite painting is 'the kiss' </p>

<p> My favourite painting is <a
  href="http://www.artisthomepage.com/klimt"
  rel="art:artistHomepage"
  property="art:workTitle">'the kiss'</a>. </p>

This leads to
<> <art:artistHomepage> <http://www.artisthomepage.com/klimt>
<> <art:workTitle> "the kiss"

With both above examples i reach the Klimt homepage but starting from pieces
of text with a different meaning.

Given the above considerations, conceptually, to jump from a simple
link to a qualifying link is quite straight (and intuitive I think)
and requires to add to the link (which corresponds to 'href' in XHTML) a
link type (which corresponds to 'rel' in XHTML).

But I think that also textual properties have a big value. Textual
properties are expressed in two different ways:

1) <span property="art:workTitle">'the kiss'</span>

2) <a
  href="http://www.artisthomepage.com/klimt"
  rel="art:artistHomepage"
  property="art:workTitle">'the kiss'</a>

Thinking about the second one, the "property" is referred to the text
enclosed in the "a" tag. So, the idea was to enrich the link prompt
with a field "property".
This is a completely new concept for HTML people, I think less intuitive than
qualifying links.

I was wondering if there's a way to make it easy.. for example the
term "property" doesn't sounds clear to me for an UI.
>
> > For adding these features it is necessary to make the prompt more complex.
> > In fact it should accept:
> > - predicate (link type --> rel)
> > - object (link --> href)
> > - property (property of the enclosed literal --> property)

--
Paolo Ciccarese
Received on Mon Nov 14 2005 - 21:29:24 EST

This archive was generated by hypermail 2.3.0 : Thu Aug 09 2012 - 16:39:18 EDT