[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [ubl] Discussion of substitution groups
I guess I expect that, with the design you are
articulating, the schema
filename might not have
the minor version as
a mandatory inclusion.
Also, if this same-namespace minor
version
design were adopted, so that software
receives
newer versions obliviously, to take
advantage
of this feature you'd want to call a schema
minor
'upgrade' by the same file name as the
previous
schema. There might not be much point
keeping
the 'space' name (namespace) but changing
the
filename, would there?
The schemaLocation is vague because it
can
be relative so you could make the
schema
minor version distinction by giving the
schemas
the same name but storing them in
different
absolute locations (a sort of external
polymorphism
without the audit trail). The relative location in
the
instance could be such that a different schema
is
pointed to depending on the location of
the
instance or some kind of caching, etc
(which
could change anyway as technology
changes).
That's what I mean by vague.
In short I wouldn't be so happy about
trading
under such an 'standard' architecture.
Nor,
I think, would my auditors or tax
authority!
(I'm not seeking to speak for these but
just
anticipating likely objections later which
could
hinder UBL adoption drastically.) I'd think
there
would be the addition of substantial
ambuiguity
by relying solely on the schemaLocation
and
schema filename rather than the
namespace.
In particular: The following doesn't look too good
-
the trading agreement gives the schema
filename
rather than the namespace (as normative).
??
That's what you get if you defer the
minor
version name to the filename but leave it
out
of the namespace.
All the best
Steve
----- Original Message -----
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]