Re: [codelist] Fwd: ICG report from Rome

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

At 2009-05-13 19:15 +0100, Anthony B. Coates (DES) wrote:
>the genericode 2.0
>proposal that I recently sent to the CLRTC list included the changes that
>they need to support this.  So, if everyone is happy with what I sent, I
>now have to get cracking with updating the specification document to
>match.

I can understand <CodeListSet> might be useful for transporting a set 
of code lists which is then burst into independent code lists on 
receipt.  But I have questions regarding any use-case for reaching in 
to a <CodeListSet> to work with the code lists that are found.  I'm 
trying to determine what code changes and spec changes this might 
have on a CVA file when the user wants to point to a code list found 
inside a code list set.

Are there any uniqueness constraints on <Identification><ShortName> 
in a set of code lists?

If I am given a code list set and two code lists in that set satisfy 
my identification criteria, do both somehow apply?  Perhaps only the 
first?  Is it left up to users?

If it is left up to users, why bother standardizing the concept of a 
set and the implications it has on the markup?  If it is only for 
transporting a collection of code lists, then just ZIP up individual 
files so they can each have their own ID-value-space and can be 
easily burst on receipt.

I noted in the example all-cats-and-dogs-set-inline.gc how awkward it 
might become to guarantee uniqueness for all of the column 
identifiers.  One ends up divining a naming scheme to uniquefy the 
column identifiers.  And, should I make a typographical error on an 
IDREF that just happens to point to a column definition in another 
code list that this would pass XSD validation.  If the lists were 
independent as would be true after bursting, such a typo would be 
caught because the ID-value-spaces are independent (though, of 
course, it wouldn't catch pointing incorrectly to, say, a key 
definition instead of a column definition).  But if I'm supposed to 
access code lists sitting inside a set, user error could easily be 
pointing outside of one set into the information of another set.  My 
earlier post presented a situation where this is explicitly allowed 
and encouraged, but what I'm speaking of here is inadvertent.

Consider that there is a <CodeListSet> at a URL ... does it make 
sense to define a mechanism by which "#" would address one of the 
lists in the set?  In a CVA file, the genericode file of a single 
code list is pointed to by a URI.  Traditionally, I suppose a URI 
with a local identifier would be to some kind of ID value 
guaranteeing uniqueness (though of course that is never guaranteed in 
an HTML document).  Since ID values are used for column 
specification, would the short name be used instead?

    http://www.example.com/gc/lists/mySet.gc#Country

... where there is a list in the set with 
<ShortName>Country</ShortName>.  But this seems fragile.

Thanks for your guidance on these issues.

. . . . . . . . . . . Ken

--
XSLT/XQuery/XSL-FO hands-on training - Los Angeles, USA 2009-06-08
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/o/
Training tools: Comprehensive interactive XSLT/XPath 1.0/2.0 video
Video lesson:    http://www.youtube.com/watch?v=PrNjJCh7Ppg&fmt=18
Video overview:  http://www.youtube.com/watch?v=VTiodiij6gE&fmt=18
G. Ken Holman                 mailto:gkholman@CraneSoftwrights.com
Male Cancer Awareness Nov'07  http://www.CraneSoftwrights.com/o/bc
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal