[tei-council] New discussion document on 1.0 release priorities

[James Cummings]

Sebastian Rahtz wrote:
> > James Cummings wrote:
>> >>
>> >> While that is a nice useful feature, I think the 'right' way to
>> >> indicate an
>> >> invalid schema is at time of generation.  User generates schema,
>> >> schema is
>> >> validated (somehow) and user is presented with a list of the errors.
> > there's a big issue there with issuing an error report
> > which is of any use. just giving them the output from
> > trang may make them very frustrated...

More or less frustrated then when they take that schema and try to use it in
oXygen only to find it complains that it is invalid?  I realise the trang error
messages aren't very useful, but I prefer some indication that the generated
schema is invalid before I take it away from Roma.

>> >> Is there any way other than element dependency to have Roma make an
>> >> invalid
>> >> schema?  Lack of start element? What else?
>> >>
> > the punter has control over all the content models.
> > if they fool with them, all bets are off.

I suppose that could be a general rule.

> > lack of start element does not make a schema invalid, BTW

Ah... just not very usable from an average user's point of view.

-James

-- Dr James Cummings, Oxford Text Archive, University of Oxford James dot
Cummings at oucs dot ox dot ac dot uk