[no subject]

Noting that UBL's 8 main documents are themselves ABIEs, one could also
start from, say, an invoice type, if the UBL's InvoiceType already has a
major portion of content that fits your purposes.  But instead of
removing those elements that you don't want, you'll need to define flags
outside of UBL's InvoiceType to mark which element isn't used.

One physical process of doing the above will be to start from UBL's
source spreadsheet and copying and pasting the entire rows of ABIE
definions onto your new document spreadsheet (stored in the mod/maindoc/
subdirectory), which itself could be a copy of one of the 8 main
document spreadsheets found there.
Then run the schema-assembler, such as UBLish, to assemble the newly
created document from spreadsheet into schema.

Another approach is to manually assemble the ABIE schema types using XML
Spy or even just text editor (yes, it's possible).
But you'll probably need to constantly check various XML Schema syntaxes
(e.g. imports), and also whether some of the global design guidelines
have been consistently maintained.

Note that such a new document schema will necessarily interoperate in a
complete manner only within your designated usage parties.  But the
approach of assembly will enable the other parties to quickly interpret
the big blocks of ABIEs within your new document based on published
semantics from UBL TC.

Note also that UBL TC is still discussing customization, compliance and
conformance.  What I mentioned above is only to share with you what
works for me so far, and may or may not end up being what UBL TC
eventually concludes.



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



On Thu, 17 Jun 2004, Steven Herod wrote:

>>How do people typically go about creating a new document type from
scratch?
>>
>>Is there a magic tool?
>>
>>I'm looking to assemble a new document (xsd) out of the ABIEs and Data
>>Types, but I'm at a loss of how to start (do I just try and pull it
together
>>in XML spy or something of the like?)
>>
>>Regards
>>Steven Herod
>>TEDIS
>>
>>