[Auscope-geosciml] RE : CGI Value abomination [SEC=UNCLASSIFIED]

Stephen M Richard steve.richard at azgs.az.gov
Fri Sep 4 11:34:38 EDT 2009


Alex--What I'm curious about is what the semantics of the quality 
category is supposed to be. Schema wise, it just associates 0 or 1 
scopedName with a quantity. In geology we deal with quantifications for 
things that vary quite a bit (think of trying to describe a typical 
dog!). A color description of a rock unit may run something like 
'usually forms reddish orange outcrops, sometimes white or gray'. Or a 
rock description might say 'typically strongly indurated, but 
occasionally very friable'. In GeoSciML, we have a CGI_Value schema that 
allows a qualifier on a quantity (numeric or category) to express the 
'usually, typically, occasionally, sometimes' aspects of these 
descriptions. What do you think of using the quality category on an 
swe:quantity to express these kind of qualifications?

there are some other questions as well

how might swe account for
1)  observation values of the sort 'value is greater than x' or 'value
is less than x'
2)  'the value ranges from a to b, median is c, mean is d'
3) categorical ranges -- e.g. 'the sandstone is medium-grained to
fine-grained'

Is a soft typed swe:dataArray approach the only option?

steve


Robin, Alexandre wrote:
> Hello Steve,
>
> I am not sure I understand your requirements but I can nevertheless tell you a little bit more about the intent of the (scalar) component quality property in SWE Common.
>
> A category component inside the quality property is used to carry a quality measure expressed as a categorical value. It can especially be used to carry a value taken from a codeSpace (defined in a remote dictionary of course)
>
> For example, a well defined list of possible quality assessment could be
> {bad, good, excellent} for codeSpace urn:myOrg:quality:simple
>
> The category component can then list one of these values like:
>
> <Quantity definition="urn:x-ogc:def:property:OGC:chemicalConcentration">
>   <uom code="mg/l"/>
>   <quality>
>      <Category definition="urn:x-ogc:def:property:OGC:qualityToken">
>         <codeSpace xlink:href="urn:myOrg:quality:simple"/>
>         <value>good</value>
>      </Category>
>   </quality>
>   <value>223</value>
> </Quantity>
>
> On the contrary a quantitative quality measure could be specified like:
>
> <Quantity definition="urn:x-ogc:def:property:OGC:chemicalConcentration">
>   <uom code="mg/l"/>
>   <quality>
>      <Quantity definition="urn:x-ogc:def:property:OGC:relativeAccuracy">
>         <uom code="%"/>
>         <value>10</value>
>      </Quantity>
>   </quality>
>   <value>223</value>
> </Quantity>
>
> Note that reuse of the same components for quality allows defining variable quality (i.e. quality measures are in the data stream itself).
>
> Hope it helps.
>
> Regards,
>
> -------------------------------------------------
> Alexandre Robin
> Spot Image, Web and E-Business
> Tel: +33 (0)5 62 19 43 62
> Fax: +33 (0)5 62 19 43 43
> http://www.spotimage.com
> Before printing, think about the environment
>
>
>
>   
>> -----Message d'origine-----
>> De : Stephen M Richard [mailto:steve.richard at azgs.az.gov]
>> Envoyé : jeudi 3 septembre 2009 20:26
>> À : Robin, Alexandre
>> Cc : auscope-geosciml at lists.arcs.org.au
>> Objet : Re: [Auscope-geosciml] RE : CGI Value abomination
>> [SEC=UNCLASSIFIED]
>>
>> Alex --
>> Can you clarify the Quality property categoryQuality on swe:quantity?
>>
>> Also, how might swe account for
>> 1)  observation values of the sort 'value is greater than x' or 'value
>> is less than x'
>> 2)  'the value ranges from a to b, median is c, mean is d'
>> 3) categorical ranges -- e.g. 'the sandstone is medium-grained to
>> fine-grained'
>>
>> thanks
>> steve
>>
>>     
>>> 	The swe abstract data components all have a Quality property, which
>>> is a union that includes a categoryQuality and textQuality. Can the kind
>>>       
>> of
>>     
>>> qualifiers we're using (always, common, never, sometimes, rare, equalTo,
>>> greaterThan, lessThan, approximate, quadratic mean, ...mode, median) be
>>> accounted for using categoryQuality? CategoryQuality appears in the
>>>       
>> sensorML
>>     
>>> v.1.0 uml in HollowWorld, but a search of the pdf for SensorML (OGC 07-
>>>       
>> 000)
>>     
>>> gets no hits on categoryQuality, so there's not any discussion of the
>>> intention of the quality category.
>>>
>>> 	can swe:singleConstraint account for greaterThan, lessThan type
>>> bounding value declarations?
>>> 	the quadratic mean, harmonic mean, geometric mean, arithmetic mean,
>>> mode, median have to do with the observation procedure, but how does swe
>>> attach those to data?  Using abstractDataComponent.definition URI?
>>> 	always, common, sometimes seem like possible quality categories.
>>> 	'Never' is only necessary for defining descriptions where the
>>> presence of some property precludes membership in a category -- I don't
>>> think it would appear in any kind of occurrence description. This kind
>>>       
>> of
>>     
>>> information should be encoded with OWL or something like that anyway, so
>>> maybe we can deprecate it.
>>>
>>> 	So maybe don't need to change swe?
>>>
>>> 	BUT...
>>> 	SWE does not appear to have a CategoryRange that would account for
>>> CGI_TermRange.
>>> 	We still have the case brought up by Bruce or Ollie of geophysical
>>> data for which there is a value range and a typical or preferred value
>>>
>>> 	Maybe these could be accounted for with some elements derived from
>>> swe:dataArray?
>>>
>>> 	steve
>>>       
Stephen M. Richard
Section Chief, Geoinformatics
Arizona Geological Survey
416 W. Congress St., #100
Tucson, Arizona, 85701 USA

Phone: 
Office: (520) 209-4127
Reception: (520) 770-3500 
FAX: (520) 770-3505

email: steve.richard at azgs.az.gov




More information about the GeoSciML mailing list