[DDI-SRG] TIC Meeting: Identifier issues

[Pascal Heus]

Achim:
Thanks for the updated. The identifier proposal looks good, very 
compact. I'm all for it.

For the data type discussion, keep me posted on progress. A mandatory 
@type attribute with a fixed controled vocabulary seem to me a bit 
dangerous (we can't miss anything).

Will this disussion only cover types or also include formatting? For the 
later, I think we'll have a hard time to have a generic approach to 
capture all possible variable formatting (particularly for dates and 
datetime). One thing that I'm not certain how to deal with is where we 
should capture proprietary formatting (for example, a specific date 
format in SPSS, Stata or SAS). Right now, this can only be done by 
creating a PhysicalDataProduct (which means we also need a logical 
record, segment, record layout, and all the related overhead). This is a 
bit heavy and I'm wonderign if we shuld provide the option to capture 
proprietary formatting at the logical product level? This could be 
embedded in the *Representation. For example, I could then specify 
something like:
<lp:Representation>
   <lp:DateTimeRepresentation @type='date'>
      <lp:ProprietaryFormat software="SPSS">DTime9.0</ProprietaryFormat>
   <lp:DateTimeRepresentation>
</lp:Representation>
Just something we may want to consider (this would be useful for the 
DExT tools).

Do you also plan a discussion on the 
RecordLayout/PhysicalDataProduct/LogicalRecord/VariableScheme 
referenceing issues? This one is pretty important to me, keep me in the 
loop.

best
*P

Joachim Wackerow wrote:
> Pascal, Chris, and others,
>
>
> We started today the meeting in Dagstuhl. We discussed again the 
> identifier issue. Please see the attached file for details. Currently 
> we are all opting for version number 4 (attributes) right now. We 
> would like to think on that. Do you have any opinions on that?
>
> Additional note:
> things like Name, VersionResponsibility and VersionRationale need to 
> be elements in base abstract Identifier ComplexType.
>
> Pascal:
> We will discuss data types. Do you have any additional opinion on the 
> issue beyond the recorded issues in Mantis?
>
> Achim