Using dc:coverage in ArgotConsultations [was Re: RDFA for Consultations and foaf/vcard/replyByEmail]
Mark Birbeck
(I've snipped some paragraphs since they are dealt with in a separate thread.)
> Also -- if a consultation has a relevant area attached to it (ie,
> building extra runways) is ArgotLocation what we'll use to describe that
> (eg dc:coverage), and if so, is an area of affect similar enough to an
> address for responses that we can use the same argot for both?
My thinking here was that dc:coverage can be vague or specific,
depending on what data is available. So it could just be a string of
text, such as "London", since that is sometimes all we get in a job
vacancy.
But it could also be very explicit, such as a long/lat value or a full
postal address. I notice that in ArgotLocation I've mentioned that
addresses can be used, but I've not given an example, so I'll do that
when I get a chance.
So to answer your question, yes, you can use a full address with
dc:coverage if you have it.
Regards,
Mark
--
Mark Birbeck, webBackplane
mark.b...@webBackplane.com
http://webBackplane.com/mark-birbeck
webBackplane is a trading name of Backplane Ltd. (company number
05972288, registered office: 2nd Floor, 69/85 Tabernacle Street,
London, EC2A 4RR)
Harry Metcalfe
Sure -- it's certainly good that dc:coverage can be vague or specific,
as required. And that makes sense for jobs.
With a consultation, though, you're often talking about more than one
location, and about different types of location.
In a consultation without any particular geographic relevance, the only
location specified is the one to which responses should be snail-mailed.
dc:coverage is fine for that.
When a consultation does have geographic relevance -- for example, the
recent eco-villages or titan prisons consultations which included
potential locations for things to be built -- it's a bit trickier. In
that case, you have both a location for responses, and several locations
for things which are being consulted upon.
It's here that I think there's a problem with dc:coverage: it doesn't
distinguish between a location that's a part of the substance of the
consultation, and one which is merely a place to which submissions can
be posted. It also can't capture complex areas.
Almost all local government consultations have some kind of geographic
relevance, and it's crucial for location-aware apps that that
information is appropriately captured. This could be difficult, as the
information is often quite fuzzy, eg:
> the area around Robert Blair School near to Caledonian Road, N7
>From a recent consultation near me. I think we should make sure that the
vocabulary includes the ability to define boundaries around areas, and
to define multiple areas -- I can imagine some nice interface with a map
to let people define those areas and generate that rdf automagically.
These should be explicitly distinct from addresses for submissions etc
which are just administrative.
What do you think? I suspect it wouldn't be often used but we should
make it available for the forward-thinking types!
Harry
On Tue, 2009-05-05 at 11:57 +0100, Mark Birbeck wrote:
> Hi Harry,
>
> (I've snipped some paragraphs since they are dealt with in a separate thread.)
>
> > Also -- if a consultation has a relevant area attached to it (ie,
> > building extra runways) is ArgotLocation what we'll use to describe that
> > (eg dc:coverage), and if so, is an area of affect similar enough to an
> > address for responses that we can use the same argot for both?
>
> My thinking here was that dc:coverage can be vague or specific,
> depending on what data is available. So it could just be a string of
> text, such as "London", since that is sometimes all we get in a job
> vacancy.
>
> But it could also be very explicit, such as a long/lat value or a full
> postal address. I notice that in ArgotLocation I've mentioned that
> addresses can be used, but I've not given an example, so I'll do that
> when I get a chance.
>
> So to answer your question, yes, you can use a full address with
> dc:coverage if you have it.
>
> Regards,
>
> Mark
>
--
Harry Metcalfe
Tel: 07790 559 876
Web: http://thedextrousweb.com
Twitter: @harrym, @dextrousweb
Mark Birbeck
> Sure -- it's certainly good that dc:coverage can be vague or specific,
> as required. And that makes sense for jobs.
>
> With a consultation, though, you're often talking about more than one
> location, and about different types of location.
Right. But what's gone wrong here is that although dc:coverage was
only intended to be one of the ways that a location is specified, that
seems to have been lost from the wiki pages; on the ArgotLocation page
we also have long/lat values as a further illustration of how to
provide a location, but I realise now that I haven't added an address
example!
Sorry about that.
There is a complete example in the document linked to from the top of
ArgotConsultation, which shows how to code up 'reply to'. It uses
vCard properties, not dc:coverage, and I include it here for reference
(I'll put this into the wiki pages):
<p>Please send responses to:</p>
<p>
Post a comment online:
<a
rel="argot:replyOnline"
href="http://haveyoursay.culture.gov.uk/worldheritage"
>http://haveyoursay.culture.gov.uk/worldheritage</a>
</p>
<p>
Email:
<a
rel="argot:replyByEmail"
href="mailto:whre...@culture.gsi.gov.uk"
>whre...@culture.gsi.gov.uk</a>
</p>
<p rel="argot:replyTo">
Post:<br />
<div typeof="foaf:Person">
<div property="foaf:name">Caity Marsh</div>
<div>Heritage for the Nation Consultation</div>
<div rev="foaf:member">
<div typeof="v:VCard foaf:Organization">
<div property="v:fn foaf:name">Culture Team (World Heritage)</div>
<div rel="v:adr">
<div property="v:street-address">2-4 Cockspur Street</div>
<div property="v:locality">London</div>
<div property="v:postal-code">SW1Y 5DH</div>
</div>
</div>
</div>
</div>
</p>
> In a consultation without any particular geographic relevance, the only
> location specified is the one to which responses should be snail-mailed.
> dc:coverage is fine for that.
Although as you can see from the example above, dc:coverage shouldn't
actually be used for that.
> When a consultation does have geographic relevance -- for example, the
> recent eco-villages or titan prisons consultations which included
> potential locations for things to be built -- it's a bit trickier. In
> that case, you have both a location for responses, and several locations
> for things which are being consulted upon.
>
> It's here that I think there's a problem with dc:coverage: it doesn't
> distinguish between a location that's a part of the substance of the
> consultation, and one which is merely a place to which submissions can
> be posted. It also can't capture complex areas.
That's a problem in my explanation, I'm afraid. :)
The approach I've taken is as follows; first, a consultation (or
indeed a job vacancy, or anything else, for that matter) has someone
that you can send 'replies' to, and that person has an address. That
address is usually coded up as illustrated above, as the address of an
organisation, which in turn has a 'member' of a person, who in turn is
the contact for the consultation or vacancy.
However, there is also the notion of some kind of 'relevant area',
that applies to a consultation (the runway, the school, the factory)
and to a vacancy (where you will be working if you get the job). For
this, we use the dc:coverage predicate.
In the wiki pages I created an argot for 'location' which was intended
to cover both of these kinds of location, but things obviously need to
be explained much more clearly for that to be the case. I've obviously
made it look like all addresses use dc:coverage.
> Almost all local government consultations have some kind of geographic
> relevance, and it's crucial for location-aware apps that that
> information is appropriately captured. This could be difficult, as the
> information is often quite fuzzy, eg:
>
>> the area around Robert Blair School near to Caledonian Road, N7
>
> >From a recent consultation near me. I think we should make sure that the
> vocabulary includes the ability to define boundaries around areas, and
> to define multiple areas -- I can imagine some nice interface with a map
> to let people define those areas and generate that rdf automagically.
> These should be explicitly distinct from addresses for submissions etc
> which are just administrative.
>
> What do you think? I suspect it wouldn't be often used but we should
> make it available for the forward-thinking types!
Great idea, and I agree. The motivation for the argots is that
something like ArgotLocation can grow to include whatever techniques
for specifying location are available.
Harry Metcalfe
H