RE: [ubl-ndrsc] Groups - UBL-ExtraInfo-1.0-alpha-draft-12.zip upl oaded

[Chin Chee-Kai <cheekai@softml.net>]

Thanks again for pointing this out, Bill.
I should've explained what it is.

The "UBL-Reusable-1.0-alpha-draft-12-TypeUsage.xls" is a 
re-generated Reusable spreadsheet sourcing information directly
from the generated schema sets.  There was no human touch
throughout the process of generating this file.

The purpose of doing so, if one might ask, is to extract and
reformat the data from Reusable so that by comparing this with
the original Reusable model spreadsheet, we should obtain an
exact match in all the important fields.  Any mismatch would
indicate problem somewhere, perhaps in the original model,
in the tool's logic, in interpretation of NDR rules, etc.

For the Usage fields (the 8 columns having "Y/N"), the values
are also generated, or more correctly, computed from searching
all the 8 files and tallying usages.  This provides a more accurate
result than human look-ups.  So far, this usage computation
process has managed to detect properly unused types that were
in agreement with other independent means, such as UML diagrams.
That 8 columns of Y/N information would be new information and
meant for copying back into the original Reusable for annotation
purposes.

Now let's look at how the re-generated Reusable is being produced.
Although the filename is called "TypeUsage", the process of
re-generation actually first gathers all the globally declared
elements.  This is more a preference than necessity.  I've
preferred to contruct complexType names by attaching a trailing
"Type" to element names (that's what NDR says) than to obtain
the element names by reverse-engineering and removing trailing
"Type" from the complexType names (which NDR rules didn't say but
might have implied).

Because of this, due to referenced elements being also declared
as global (these are the elements listed in sorted ascending order
at the end of the schema), the elements also get processed.  
However, their defined types would be one of those
complexTypes defined earlier in the Reusable.

The contents of these global elements, however, are listed
just so that one could check whether the contents make sense
for that element.  This listing is also a set of new information
not found in (but can be derived from) the original Reusable model.

So one could ignore type usage columns for all the rows after
475 ("AccountsContact" onwards).

I'd take this feedback and perhaps make the demarcation more
noticeable in case it gives a "run-out-of-steam" impression.

Thanks.




Best Regards,
Chin Chee-Kai
SoftML
Tel: +65-6820-2979
Fax: +65-6743-7875
Email: cheekai@SoftML.Net
http://SoftML.Net/


On Fri, 17 Oct 2003, Burcham, Bill wrote:

>>Why is it that lines 476-850 of UBL-Reusable-1.0-alpha-draft-12-TypeUsage
>>are blank in columns AL-AS.  Does that indicate that the complex types and
>>elements shown on those rows are referenced in none of the UBL document
>>types (Order, Order Change, etc)?  Or did we just run out of steam in
>>filling in the check marks after 475 rows (understandable!)  Is this
>>spreadsheet generated automatically, or manually?