Editor's Draft of ADMS spec

Hi all,

We've spoken at length about ADMS but one thing that has been missing is 
a draft GLD version of the spec. This I have now created at [1].

The original ISA Programme version [2] includes several things I have 
removed for this version:

1. controlled vocabularies as values for various properties;
2. cardinality constraints;
3. text relating specifically to the ISA Programme etc.

I assume 3 is straightforward an uncontentious. Points 1 and 2 are 
related to the discussion we had on last week's call about conformance 
and that I wrote about briefly on the Team blog [3] (to which several 
folk in the WG were kind enough to respond).

Does the WG agree with this approach? Based on the discussion with Rufus 
Pollock about DCAT last week I'm pretty sure the answer's yes but it 
would be good to record this explicitly (and, of course, any 
discussion/views to the contrary).

The specific implementation of ADMS that the European Commission is 
building, Joinup [4], *does* have cardinality constraints on various 
properties and *does* specify controlled vocabularies as allowed values 
- and that strikes me as entirely reasonable. But I don't think they 
belong in the spec.

The Conformance section of the draft tries to capture this idea. I hope 
it, or something like it, will also apply to DCAT and other vocabs. If 
we do this right it might become a boilerplate for future vocabularies 
(or even a separate Note).

The current draft refers to the RADion vocabulary I've raised several 
times [5]. That is, ADMS subclasses the RADion structure and directly 
uses a small number of its actual properties. However - I think a better 
solution is probably to use the proposed schema.org extension [6] which 
more or less does the same job. DanBri is keen to get feedback on this 
extension and I believe that if we took a look at it in the light of 
both DCAT and ADMS before providing feedback we'd be able to help see 
that proposed schema.org extension move into the schema.org core. RADion 
would then disappear back into the ether from whence it came (I might 
have to put something in the namespace to that effect).

The examples in the doc are all written in RDF/Turtle. We really ought 
to have examples in XML as well I think since we're trying to present 
this as technology neutral. That will take a bit of work but it's not 
impossible. The XML schema that accompanies the EC release of ADMS 
includes all the controlled vocabularies and cardinality constraints so 
it's not directly usable (it's also incredibly complex to my eyes). I 
hope that the PwC folk I'm working with will be willing and able to 
create a version that matches the document we're looking at here.

Dave R made comments on the RDF schema for ADMS to which I will reply 
separately.

Thanks

Phil.

P.S. I'll be offline from Tue 7th and not fully back in the saddle until 
Tue 28th so I won't be able to take part in any active discussion at 
that time.


[1] http://dvcs.w3.org/hg/gld/raw-file/default/adms/index.html
[2] http://philarcher.org/isa/ADMS_Specification_v1.00.pdf
[3] http://www.w3.org/QA/2012/07/conformance_for_vocabularies.html
[4] http://joinup.ec.europa.eu/
[5] http://philarcher.org/isa/radion_v1.1.html
[6] http://www.w3.org/wiki/WebSchemas/Datasets

-- 


Phil Archer
W3C eGovernment
http://www.w3.org/egov/

http://philarcher.org
+44 (0)7887 767755
@philarcher1

Received on Friday, 3 August 2012 12:59:45 UTC