DDR Simple API Draft 1c

Hello Everyone

Here [1] is Draft 1c of the API document. A bit rough and ready but I
wanted to get it out in time for everyone to review before Monday's
call. There is no diff from previous versions as the changes too
extensive for a diff to make any sense.

[1] http://www.w3.org/2005/MWI/DDWG/Drafts/api/080215


Note that the discussions of methods are given under Java rather than
IDL signatures for now.

I'm not sure we considered the possibility of adding Aspect to
PropertyName. Doing so reduces the complexity of the API by removing the
needs for a number of methods of the SimpleService. This would also
solve a problem we have with SimplePropertyValue which while it says
what property the value is associated with does not say which aspect the
value applies to.

This would also mean that aspects are part of the vocabularies rather
than standing alone, which on reflection makes sense to me.

PROPOSED RESOLUTION: We normatively define a Vocabulary, from the point
of view of the Simple API, as a set of names of properties, a set of
names of aspects that the properties refer to, a binding between
PropertyNames and Aspects and an IRI declaring the Namespace for the
Vocabulary. We normatively define there to be one value in all such
Vocabularies, with the meaning "Unspecified", represented by a null
value, empty string or similar as appropriate to the language of the
implementation.

I don't see a need to define any structure or serialization
representative of a vocabulary, but I do think we need to define what
data types values of properties can attain. The suggested list is
contained in SimplePropertyValue getDouble etc. I don't have strong
views about this myself but am unsure that we need double float long
integer etc.

We suggested at the Editors' meeting that property names and namespaces
are defined like they are in XML - i.e. that the names of properties are
valid NCNames. We should take a resolution on this:

In the current draft API it is suggested that Apsects are regarded as
IRIs. I am not personally in favour of this if they are to be considered
as part of the vocabulary. I think that NCName is fine for this too.

DRAFT PROPOSED RESOLUTION: Property Names are valid XML NCNames.
Vocabulary Namespaces are IRIs. Aspects are likewise NCNames. Property
values my be chosen from the following types: Boolean .... each with the
interpretation defined in XML Schema ...

Jo

Received on Friday, 15 February 2008 14:11:25 UTC