Re: FOAF spec revised - addtion of foaf:focus, a skos extension linking topical and factual information

+Cc: Leigh

On Mon, Aug 9, 2010 at 4:03 PM, Leo Sauermann <leo.sauermann@gnowsis.com> wrote:
> uh, didn't Dan's mail begin with
>
> "we just published the new version"
>
> that implies - its fixed. comments to change it aren't helpful, I guess.

We have a few mechanisms in hand here, to allow a little flexibility.

Each term in the FOAF namespace has a vs:status property, whose values
were originally borrowed from Debian: 'unstable', 'testing', 'stable'.
To these we recently added a new option, 'archaic', which is a kinder
gentler and more Webby variant on deprecation. In a global system it
strikes me as a bit rude to deprecate RDF vocabulary unless there is
some strong reason to do so. People may have used it in good faith,
and if the meaning was clear, likely it still remains clear. But there
is of course a need to avoid vocabularies accumulating clutter over
the years, hence 'archaic' which is intended to indicate the term is
somehow old-fashioned or falling from widespread usage. I added retro
CSS sepia colouring to the spec too, and these terms are currently
sorted to show up last in the documentation.

(see http://www.w3.org/2003/06/sw-vocab-status/note for draft note on this)

A second mechanism is the SKOS spec itself. FOAF has always served a
dual role: as a language for describing, and as an experimental
testbed for Web-scale informational-linking. Unlike a W3C
Recommendation, the overhead for getting something into FOAF is fairly
low and informal. However if our implementation experience from
foaf:focus leads to a clearer idea of the design options, we could use
that to prepare a proposal for a new W3C property in SKOS itself.

This is always a pragmatic balancing act. If, say, today someone came
up with a new property name suggestion which was universally applauded
for being clear, not already in use, etc etc., we could probably
switch foaf:focus with little cost and it might not even be worth
keeping the old name in the spec as 'archaic'. If however we leave it
there for say 3 months, and people start to publish data that uses it,
and write code and SPARQL queries and OWL axioms etc., then I think it
would be somewhat anti-social to completely delete the term from FOAF.
If we left it for two weeks and then had a great idea for a new name,
probably we could talk to all the relevant implementors, and see if
they were happy to switch.

> you usually discuss before and then publish an ontology. Also, afaik its not
> very common and not good to change the property identifier after publishing.

Yep, as I say it's a balancing act. When we started FOAF, the spec was
created by documenting terms that people used 'in the wild', which is
how we ended up with mess like firstName vs last_name. It is hard to
get implementation feedback on things that haven't - to some extent -
been documented. So that's how we ended up with the current rough
process, which is that things flagged 'unstable' have some risk of
being changed under your feet. That said of course it's best to avoid
needless change.

With foaf:focus I think it's still unclear, but the discussion is more
than welcome, since even if foaf:focus stays as-is, debate here can
inform future plans for SKOS. If we get a bright idea for a new and
better property name in the very near future, it isn't too late to
switch and either mark 'focus' as archaic, or if nobody screams, even
consider deleting it in favour of the (currently theoretical)
better-named successor.

But I would also like to see actual data and usage too, rather than be
stuck forever on the naming problem. The larger question of how
SKOS-like modelling and OWL-like modelling connect won't be addressed
by a single property, so we'll have plenty of future opportunities to
reconsider things. My main hope is to make sure those discussions are
grounded in real data published in the Web; having a reasonably-well
documented property out there and usable is an essential ingredient
for that. So I hope Simon's concern that 'focus' is already used
doesn't prove too much of a showstopper, since I don't yet since
another usable naming suggestion.

cheers,

Dan

> the remarks are intereesting for the "rdf:comment" section though.
>
> my 2c
> Leo
>
> It was Bernard Vatant who said at the right time 09.08.2010 15:55 the
> following words:
>
> Hi Dan
>
> Although I loved foaf:focus at first sight because of the convergence
> metaphor (see http://blog.hubjects.com/2010/03/foaf-focus.html) I would tend
> to follow Simon's argument.
> Maybe I miss something, but are not we speaking about referent here, in the
> semiotic sense.
>
> So why not foaf:referent ?
>
> Bernard
>
> 2010/8/9 Simon Spero <ses@unc.edu>
>>
>> Dan-
>>
>> can i suggest using a different word  than focus, as this is term of art
>> in controlled vocabularies. It is used when referring to
>> modified/specialized "terms".
>>
>> Possible labels that might work could be  isReferredToBy ; SKOS concepts
>> are intentional-with-a-t, so reference is a natural label;
>> isFoafProxyForIntentionReferencedBySKOSConcept is awful ComputerDeutch.
>>
>> Foaf person "Paul The Octopus" isReferredTo by SKOS Concept "#PTO1".
>>
>> Where "#PTO1" isSubjectOf "#document" "Decideabity and tractablity of
>> logical inference with binary serial octacles".
>>
>> (The halting problem has time complexity PTO(1) but other tasks may
>> require an infinite series of questions.)
>>
>> On Aug 9, 2010 8:19 AM, "Dan Brickley" <danbri@danbri.org> wrote:
>> Hi SKOS folks
>>
>> http://xmlns.com/foaf/spec/
>> http://xmlns.com/foaf/spec/20100809.html#term_focus
>>
>> Just to let you know, there's a revision of the FOAF specification
>> today. It includes a new term, foaf:focus that links a skos:Concept to
>> the thing that the concept stands for.
>>
>> This notion has been discussed many times here over the years,
>> sometimes as "skos:it", but never made it into W3C's REC-track SKOS
>> spec. FOAF has long contained a cluster of topic-oriented properties
>> (topic/page, primaryTopic), and in FOAF we have a long-standing
>> concern with describing the areas of interest and expertise for people
>> and other agents (eg. organizations, groups, projects). The addition
>> of foaf:topic is intended as a modest and pragmatic bridge between
>> SKOS-based descriptions of topics, and other more entity-centric RDF
>> descriptions. When a SKOS Concept stands for a person or agent, FOAF
>> and its extensions are directly applicable; however we expect
>> foaf:focus to also be used with places, events and other identifiable
>> entities that are covered both by SKOS vocabularies as well as by
>> factual datasets like wikipedia/dbpedia and Freebase.
>>
>> Other relevant changes: the overview of FOAF at the top of the spec
>> now more clearly separates two informal sub-sets of FOAF terms: "Core
>> FOAF" terms and "Social Web" terms. The distinction is made with
>> regard to whether a term is useful in describing someone or something
>> who lived before the Web / internet. Only the more universal
>> characteristics of groups, people etc are considered 'core FOAF';
>> things like 'homepage', 'openid', 'weblog' are in the "Social Web"
>> layer. Previously, we mistakenly gave the impression that FOAF was
>> only for describing modern-day online accounts; hopefully the new
>> formulation more accurately conveys an interest in capturing
>> historical information too. There have also been some other textual
>> changes that attempt to indicate more clearly what we're attempting
>> with FOAF - essentially the combination of social and informational
>> networks.
>>
>> Re the "Core" subset, brief excerpt: "Core - These classes and
>> properties form the core of FOAF. They describe characteristics of
>> people and social groups that are independent of time and technology;
>> as such they can be used to describe basic information about people in
>> present day, historical, cultural heritage and digital library
>> contexts. In addition to various characteristics of people, FOAF
>> defines classes for Project, Organization and Group as other kinds of
>> agent."
>>
>> Also, various older terms (used in early demonstrations and
>> prototypes, plus some spelling variations) are now marked 'archaic',
>> both in human and machine-readable documentation.
>>
>> Feedback on the current design and description are welcome, either
>> here or on the foaf-dev list. My hope is that with foaf:focus we can
>> begin today gathering real-world implentation experience and data that
>> could inform any future revisions to SKOS itself. If W3C were to
>> eventually charter and complete an effort to update SKOS with matching
>> functionality to foaf:focus, we would of course update FOAF
>> accordingly to indicate the new mechanism. In the meantime, foaf:focus
>> is available for use, experimentation and collaboration. I hope it
>> proves useful when linking topically structured and factually based
>> RDF information.
>>
>> cheers,
>>
>> Dan
>>
>> ps. one thing the spec currently lacks is an example of the new
>> property. I'm waiting on this point, as several people are working on
>> related datasets, and I hope soon we'll have real-world examples to
>> illustrate foaf:focus's usage.
>>
>
>
>
> --
> Bernard Vatant
> Senior Consultant
> Vocabulary & Data Engineering
> Tel:       +33 (0) 971 488 459
> Mail:     bernard.vatant@mondeca.com
> ----------------------------------------------------
> Mondeca
> 3, cité Nollez 75018 Paris France
> Web:    http://www.mondeca.com
> Blog:    http://mondeca.wordpress.com
> ----------------------------------------------------
>
>
> --
> Leo Sauermann, Dr.
> CEO and Founder
>
> mail: leo.sauermann@gnowsis.com
> mobile: +43 6991 gnowsis
> http://www.gnowsis.com
>
> helping people remember,
>
> so join our newsletter
> http://www.gnowsis.com/about/content/newsletter
> ____________________________________________________
>

Received on Tuesday, 10 August 2010 08:34:03 UTC