Re: [ubl-ssc] Schema comparison

["Stephen Green" <stephen_green@bristol-city.gov.uk>]

Folks

I've done more work looking at the Despatch Advice Schema and found the attached issues/evidence of bugs. Sorry I haven't been able to completely track down what appear to be bugs but I've found some pointers as mentioned in the attached. I guess there may well be further problems to be discovered once the initial ones are removed. I still have XML Spy (no time yet to install a xerces based validator) but the validation side is at least evidence of some differences between the new spreadsheet-generated Schemas and the published ones.
I really suggest after this initial testing the use of text-based diff. With XML Spy you do get the added bonus of canonicalization before the diff. But Anne's system for command-line text diff might be the sort of thing it looks like we need. Then I'd personally suggest we look for a text-to-text match rather than xml-to-xml since the inconsistencies seem to be logically complex and it looks like it would be better to ensure textual eqivalence rather than xml/logical equivalence. The latter might be the longer route.

All the best

Steve

>>> "Stephen Green" <stephen_green@bristol-city.gov.uk> 07/12/04 14:11:59 >>>
Hi all

Many thanks indeed to David for completing this stage
in good time for the Thursday meeting.

As a first pass this looks very encouraging to me.

I've been very short of time at present (I hope for more 
by Thursday) but I did manage to look at the DespatchAdvice,
at backwards instance compatibility, and found only a few issues -

1. Currency Code issues as David describes 

2. Some issues with the UDT components which I don't yet
understand so I'll work on some more hopefully by Thursday

3. Again, as David identifies, the trailing space

The second issue is a bit peculiar and look into it some more later
but otherwise this looks like we have great progress being made
in EF by David et al. Many thanks :-) Excellent work!

I haven't looked into aspects of documentation/annotation as
this was not in my action item scope but it will of course be in
need of examination and may throw up some interesting matters :-)

I hope we have time to look too at Chee-Kai's issues on the subject
soon.

All the best

Steve


>>> "David Kruppke" <dill2@gefeg.com> 06/12/04 18:27:53 >>>
Hello SSC,

I did the comparison between the schemas exported from the old EDIFIX models
and the new ones imported from ss.


In the attachement you can find the following files:

1. xsd and xsdrt schemas from o l d EDIFIX models
(these files are not the originals from 2004-08-16. They were created after
we had completely recasted our UBL export.)

2. Changes to schemas from 2004-08-16.doc
(this file I sent already in October to Stephen. I think it should be known
to all SSC members. It describes the differences between the original
schemas from 2004-08-16 and the one in 1. .)

3. Used spreadsheets and used code list text files for importing

4. xsd and xsdrt schemas from new EDIFIX models imported from Excel

5. Report.doc
(this file lists the differences between 1. and 4. and the problems I
encountered during the import of the ss.)


I'm not be able to attend the TC meeting at Wednesday, but I can join the
SSC meeting on Thursday.



All the best,



David



Issues and Possible Bugs in EF alignment 2004-12-07 (sdg)

1. CurrencyCode values - at least one previously illegal code value is found in the currency code list

2. Errors in UBLAmount mean code not enumerated - same design needed as in published 1.0-cd

3. In CodeList Schemas simpleContent has been changed to complexContent (due to finding that XML Spy didn't like
   the simpleContent ?)
   - same in corresponding place in Currency in SDT Schema 
   but I think this means that a new EF SDT Schema will not be compatible with the 1.0 Currency CodeList Schema

4. possible errors in common lead to a LineItemCountNumeric error message

5. possible errors in maindocs lead to a CopyIndicator error message

6. codelsist OK except two trailing space errors - one the lack of escaping of `SPACE`, the other a new
     trailing space in codeListURI attribute in AllowanceChargeReasonCode