[foaf-dev] Proposal: deprecate pastProject and currentProject

Dan Brickley danbri at danbri.org
Mon Dec 14 14:55:08 CET 2009


On Mon, Dec 14, 2009 at 2:26 PM, Toby Inkster <tai at g5n.co.uk> wrote:
> On Mon, 2009-12-14 at 12:43 +0100, Dan Brickley wrote:
>> 3a. Richard's version - entity rather than page centric?
>>
>> Can you post RDFa and triples here. Non-scary HTML is a goal, ie.
>> stuff that people can write without thinking much about RDF.
>
> I am not Richard, but the most elegant RDFa I can write for
> entiy-centric properties is:

Thanks for taking the time! It's useful to have concrete examples.

> <p about="http://danbri.org/foaf.rdf#danbri" typeof="foaf:Person"
> xmlns:foaf="http://xmlns.com/foaf/0.1/">
> My name is <span property="foaf:name">Dan Brickley</span> and I spent
> much of the '80s at <span rel="foaf:school"><a
> typeof="foaf:Organization" property="foaf:name" rel="foaf:homepage"
> href="http://www.westergate.w-sussex.sch.uk/">Westergate School</a></a>.
> Before that I was at <span rel="foaf:school"><a
> typeof="foaf:Organization" property="foaf:name" rel="foaf:homepage"
> href="http://www.eastergate.w-sussex.sch.uk/" content="Eastergate
> School">Eastergate</a></span>, where we moved in 1978 from <a
> href="http://en.wikipedia.org/wiki/Hamsey_Green">Hamsey Green</a> near
> <a href="http://en.wikipedia.org/wiki/Croydon">Croydon</a> where I
> attended the <span rel="foaf:school"><a typeof="foaf:Organization"
> property="foaf:name" rel="foaf:homepage"
> href="http://hginfant.schoolwebbuilder.co.uk/" content="Hamsey Green
> Infants' School">infants school</a></span> I guess from late 1976 to
> summer 1978. Before that, it's all a blur :)
> </p>

So - this is bearable I think, but certainly a bit heavy for some of
our audience. Particularly if they're going to hear that the simpler
rel= version is an anti-pattern. It's good to have nice templates for
both idioms.

> Or, in summary, each school link becomes:
>
>  <span rel="foaf:school">
>    <a typeof="foaf:Organization"
>       property="foaf:name"
>       rel="foaf:homepage"
>       href="LINK">NAME</a></span>
>
> (Giving the <a> element a content attribute where you wish the link text
> and foaf:name of the school to differ.)

Yeah, it's a nice extra to have the name. Hmm how does that interact
with title= attributes in HTML? do RDFa parsers do anything with it?
What's our accessibility story here)

> This is certainly more verbose, adding an extra element for each school,
> but as a bonus, it gives us an explicit foaf:name and rdf:type for each.

Yup

> If each school had an established linked data URI, then it would be
> possible to do something like:
>
>  <a rel="foaf:school" href="LINK" resource="DATA_URI">...</a>
>
> And you win back a bit of simplicity, at the cost of losing the
> foaf:name and rdf:type triples.

That can sometimes be really nice, sometimes be too heavy. Semantic
Web means never being able to tell people exactly what they can say in
their docs, maybe? ;)

So for me one of the big reasons for pushing RDFWeb / FOAF was the
idea of an information linking 'division of labour', in which each
linked description might briefly describe some entities, but then
defers to the linked-to description for the full detail. For obvious
(at least to Kingsley :) reasons, this has some commonality with the
reasons people liked the relational model and SQL, you can avoid
redundancy and have a canonical shared description linked by reference
rather than wildly duplicated.

In this present scenario, we might ask "why is every attendee of the
school bothering to give it's name and type, when all they need give
is a single identifier and let the computers figure out the rest. The
homepage for the school - or maybe a govt database like
http://blogs.talis.com/n2/archives/818  - can have or link to the full
details of its name, org type, history etc etc. This is one of the
strengths of linking data items together. Now currently, there is
almost no machine-readable markup on school homepages. We can expect
this to change gradually, as tools like Drupal7 get adopted, as the
search engines start to consume RDFa, and as govt aggregations of this
data become available in Webby formats.

So the question here is whether the idioms we're using are already too
fragmented for the data to be put back together again. If my blog post
merely says I have a foaf:schoolHomepage of
http://www.westergate.w-sussex.sch.uk/ ... and if the govt data or
school site say that there is a :Organization with a :homepage of that
same URL, ... by what logic can machines put this together and figure
out that I attended that school?

Right now, it would have to be hard-coded. There is nothing in the
RDFS or OWL for FOAF that expresses this logic. We don't even name the
property 'foaf:school' that relates a person to one of their attended
schools.

Are there simple fragments of OWL2 we can use use to document these
'triangular' patterns?



More information about the foaf-dev mailing list