On Mon, Aug 3, 2009 at 6:57 PM, Bruce Miller<[email protected]> wrote:
> On 08/03/2009 06:25 PM, Jerzy Jalocha N wrote:
>>
>> I think that you all have sent me enough evidence, that shows that
>> embedding a RELAX NG schema in an arbitrary document doesn't work, and
>> is probably not valid.
>
> But, it can be _made_ to "work", and it is "valid" if
> you say it is!
> Mainly, I think you need to refine (or define :> ) your terms.
Ooops, I guess, my vague and unfortunate wording created again more
confusion. Sorry!
I am already working successfully with my 'literate schemas'. The
'embedded schema' is extracted with xsltproc, and that is used in turn
to validate other documents with xmllint. I also have a
'literate-schema.rng', that defines (and validates) my 'literate
schemas'. And everything is nicely wrapped in a small shell script.
> However, you certainly could pursue your idea by
> either developing a version of xmllint that _did_ extract
> the schema from the enhanced schema, or alternatively
> developing separate tools like web/tangle/etc that extract it.
Fortunately, the described workflow is simple and does the job
('works'), without having to get my hands dirty. Semantically, the
documents express exactly what I want. 'Validity' is both defined and
guaranteed by the use of 'literate-schema.rng'. Performance isn't a
problem, either, on this small scale. (Anyway, I've never experienced
performance issues with xsltproc & xmllint --quite the contrary, they
always are lightning fast!)
Thanks for your answer, Bruce.
Regards,
Jerzy
_______________________________________________
xml mailing list, project page http://xmlsoft.org/
[email protected]
http://mail.gnome.org/mailman/listinfo/xml
