OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

tag message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: test assertions - suggestion


Just a suggestion here:

While we ponder the public review period, how about we
write test assertions for the two specs (model and markup)
- at least make an exercise of doing so? To at least 'take
own medicine' as it were.

I guess we could decide that many aspects of the spec for
the markup do not require much more than the simple TA
predicate that the instance be as defined as valid by the
schema (though the recent W3C note on assertions for
'widgets' for mobiles seems to counter that practice). The
schema might be seen as effectively a set of optimised
TAs perhaps for details of the markup, negating the need
to duplicate such detail in the TAs. What other TAs might
we need for the model and perhaps especially for the markup
which cover areas not covered by the schema?

Also, something I've come across when writing TAs already:
could we think how best to distinguish 'universal' targets
(target 'types'?) from 'particular' targets? For example, with
TAs for a UBL invoice, a target might be 'invoice lines' in
general - where the predicate applies to each and every
invoice line in an invoice. On the other hand there might be
a more particular target which is the invoice tax point date,
of which it is a (perhaps implicit) prerequisite or assumption
that there is only one in an invoice. Are there clear ways to
cover these two use cases, which seem to me to be two
clearly typical categories of use cases?

Best regards

Steve
---
Stephen D Green


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]