[DDI-SRG] question on comparability module

[Wendy Thomas]

See inserted responses

On Thu, 24 Jan 2008, Pascal Heus wrote:

> All:
> sorry but each time I look at the comparability module, I'm a bit
> puzzled by the limited types of elements and operations I can put in there.
> - How can I compare two studies to for example simply state that they
> should not be compared or only be compared for specific purposes?

We agreed to currently limit explicit comparisons to Univers, Concept, 
Question, Variable, Category, and CodeSchemes because we do not currently 
have sufficient detail in the description of methodologies and other 
upper level items to do machine actionable comparisons. These will be 
added when we get the methodology descriptions improvided (currently in 
the works and expected as a 3.x expansion). General comparative statements 
in grouped materials goes into the purpose of the group. That is why you 
grouped them which would include general comparative/non-comparative 
information.


> - What if I want to talk about a group of something (variables, surveys,
> ...)?

If you mean variable schemes are you saying you want to say that a full 
scheme is the equivilent of another full scheme? In an after the fact 
comparison? If its by design you'd simply use that other scheme by import 
or inclusion. Survey's as a whole....see note above.


> - Why can't I compare more than two items?

This is standard pair-wise comparison. You declare a source and target. 
Any two targets coparable with the same source are also comparable If A=B 
and A=C then B=C. We've gone over this several times and at this point 
pair-wise is how it is. I beleive Arofan has addressed this explicitly in 
the past.


> - Why does CodeMap gets  a special structure (going to a control construct?)
Because I, we, many of us, need to provide machine actionable translations 
of codeSchemes to create harmonized sets. We needed to be able to 
explicity state "Source values 1-4 map to Target value 1". We started to 
write all the pieces that would allow this and then discovered we were 
talking about IfThenElse, RepeatUnit etc and so made these usable by 
CodeMap.


> To alleviate the element type restriction, I remember suggesting a while
> ago to have a "generic" map element that allows for pointing to any two
> identifiable elements and talk about their comparability.

We intentionally limited comparison to the element types listed. This was 
agreed to at our SRG meeting at ICPSR when we had Reto on the phone for 
hours.

  Is this
> something that could be included? Given that the GenericMapType already
> exists, this should be pretty easy to add.
> thanks
> *P
> _______________________________________________
> DDI-SRG mailing list
> DDI-SRG at icpsr.umich.edu
> http://www.icpsr.umich.edu/mailman/listinfo/ddi-srg
>

Wendy L. Thomas                          Phone: +1 612.624.4389
Data Access Core Director		 Fax:   +1 612.626.8375
Minnesota Population Center              Email: wlt at pop.umn.edu
University of Minnesota
50 Willey Hall
225 19th Avenue South
Minneapolis, MN 55455