[xmlschemata] Re: Support of warnings by schema languages

On Fri, 2002-06-28 at 13:20, Rick Jelliffe wrote:

> The most likely candidate for such a feature is to be able to mark
> things as obsolete or deprecated, thereby allowing smoother
> migration.

Yes, this (versioning) is a good catch and something really missing
right now.
>
> Apart from phases, schema languages so far have not really reconstructed =
>
> the marked section feature of DTDs. WXS tacked on=
> unmanaged <redefine> at the last minute, and that recent
> stages idea would go a bit of the way.
>
> I wonder if we need to have a three layer model for invoking a =
> particular
> schema pipeline:
>
> 1) A phase selector switches on-or-off various piplines.
> 2) A version selector remaps parameters
> 3) the pipeline itself.
>
> This is because versioning is really a distinct issue from
> the phase of validation. To be more concrete:
>
> <schemachine>
> <phase name=3D"simple">
> <active process=3D"s1"/>
> </phase>
> <phase name=3D"extended">
> <active process=3D"s2"/>
> </phase>=20
>
> <validator name=3D"s1" engine=3D"relaxng" >
> <param name=3D"schema">xxx.rng</param>
> </validator>
> <validator name=3D"s2" engine=3D"relaxng" >
> <param name=3D"schema">xxx2.rng</param>
> </validator>
>
> <version name=3D"original" />
> <version name=3D"new">
> <override processor=3D"s1">
> <param name=3D"schema">xxx3.rng</param>
> </override>
> </version>
> </schemachine>
>
> In that case, the idea is that the version element allows
> parameters of processors to be overridden. So to=20
> validate I must specify a particular version and a particular
> version. =20
>
> The changed parameter could nominates a different schema,
> or the particular validation engine could itself know how to construct
> the schema (e.g. using xsvif or stages).=20

This may be OK for version management, but not necessarly for the more
general handling of warnings which deserves IMO to be an "inie" and does
impact the logic of the validation.

To come back to Relax NG Right, there is well defined composition tables
to compute what's happen when you have "emptiable" and "not allowed"
results in a group a choice or an interleave pattern and these
composition tables would need to be updated to deal with a third state.

It's probably not a big deal but that's the price to pay to make sure
that for instance when you have a choice between something which is
allowed and something which is allowed with a warning the result doesn't
carry a warning.

> Cheers
> Rick Jelliffe
>
>

-- 
See you in San Diego.
                               http://conferences.oreillynet.com/os2002/
------------------------------------------------------------------------
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------