Expressing position in RDF

[Bernard Vatant]

Hi all

Have been lurking at this conversation so far. Some thoughts below, 
about this very generic use case in science : how do you attach 
measurement results to an object? Seems to me the scientific community 
(at least its members involved in Semantic Web) should try and 
standardize this at some point. I've been looking for relevant pointers 
to people working on this in other domains (say e.g., Biology, Earth 
sciences ...) without much success so far I'm afraid. Of course hoping 
astronomers would be leaders in such a process :-)

Matthew wrote
> Although this works, it is very verbose for such a simple (!) 
> statement. To sell this sort of stuff to end users, we really need a 
> practical and succinct set of object properties that would allow me 
> just to do:
>
> :MyObject a:rightAscension 134.556^^xsd:float
Two remarks

1 : "object properties" might be confusing in this conversation context. 
I guess you mean "properties of objects", not "ObjectProperty" in the 
OWL sense ... since actually your a:rightAscension would be a datatype 
property.

2 : With such shortcuts, you don't assert the coordinate reference J2000 
and the unit, so it is implicit, which is not a good idea ... or you 
have to built-in the coordinate system and unit in the property ...
:MyObject a:rightAscensionJ2000In DecimalDegrees 134.556^^xsd:float

The problem with this is that it leads to a proliferation of properties, 
and you can't in this case assert "properties of the property", such as 
accuracy of the value, method of acquisition, and whatever relevant 
description of the measure process ... Direct assertion is blurring the 
fact that attaching a property-value pair to an object, in science, is 
called a measurement, of which description must come along with the data 
if those data want to be qualified as scientific data.
So it is not only a technical issue of representation, what is at stake 
is what qualifies scientic data from simple property-value pairs. I 
don't know which "end users" you have in mind, but there is a huge trend 
in science outreach to hide all the data context to "simplify" the 
presentation of data. See what I mean :-)

This is the local illustration of all the issues debated at length in 
the Semantic Web community at large about trust, data quality etc.

Ed, for a simpler expression of your example, one can use nested blank 
nodes for coordinates (it's indeed a typical use case of blank nodes)
and I don't think either asserting the class of those nodes is really 
useful, nor specifying a URI. Merging or matching of such information 
would in any case drill down to the values. OTOH, I added the precision 
of the data, and some Dublin Core metadata on the measurement. Which is 
not possible to add using the direct property on the object.

:MyStar a a:star;
 ivoa:hasPosition _:b1

_:b1 geo:hasCoordinate _:myRA.
_:b1 dc:source   :MyCatalog
_:b1 dc:date   "2008-10-14:00.00"

_:myRA a a:rightAscension;
  q:hasFloat 34.5343^^xsd:float;
  q:hasUnits ui:degree;
  q:hasPrecision  0.0002^^xsd:float.

My 0.02

Bernard

>>
>> On Oct 13, 2008, at 7:24 PM, Ed Shaya wrote:
>>
>>> Here is "MyStar has a J2000 RA of 34.5343 degrees, it would be 
>>> something like this in N3:
>>>
>>> :MyStar a a:star;
>>>    ivoa:hasPosition :MyCoordinates.
>>>
>>> :MyCoordinates a 
>>> a:J2000Coordinates;                                                       
>>> geo:hasCoordinate :myRA.
>>>
>>> :MyRA a a:rightAscension;
>>>   q:hasFloat 34.5343^^xsd:float;
>>>   q:hasUnits ui:degree.


-- 

*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel:       +33 (0) 971 488 459
Mail:     bernard.vatant at mondeca.com <mailto:bernard.vatant at mondeca.com>
----------------------------------------------------
*Mondeca**
*3, cité Nollez 75018 Paris France
Web:    www.mondeca.com <http://www.mondeca.com>
Blog:    Leçons de Choses <http://mondeca.wordpress.com/>
----------------------------------------------------**