[tei-council] the new <constraint> element

[James Cummings]

Sebastian Rahtz wrote:
> what are we trying to achieve with grouping? do you want the 
> constraintGrp to be addressable at the customization level? ie say
> 
> <constraintGrp ident="harsh" mode="delete"/>
> 
> in your ODD? currently the only point of <constraintGrp>
> is to group things by language/scheme; if you group
> by functionality, the Grp should certainly by identifiable,
> and we'd have to support @scheme on <constraint>.

I see what you mean, thank you for the explanation.  Playing devil's 
advocate, along with ODD customising ODD, might one want to be able to 
take on an ODD and then delete some category of contraints?  I honestly 
don't really think so.  It is far more useful from a documentary 
perspective to record deleting all the individual constraints rather 
than a Grp of them.

> if the Grps are to be semantic, then I suggest that we need
> to go back and add considerably more in the way of
> metadata to them (<desc>, or the full <biblStruct>). otherwise,
> whats the point of grouping?

Yes, I don't think we want to go that far, and I now agree that maybe 
semantic grouping isn't that useful.  However, even in non-semantically 
grouped constraints, where do I record a description of the rule? Let's 
say I want to have something different from the sch:report to describe 
the purpose of the rule. (Or a chosen language might not have something 
<desc> like.)  Should constraint specifically have <desc> inside it? 
I'm also thinking from a TEI Documentation generation perspective. 
Surely it is good to have somewhere to put information about why this 
rule exists?  Something like:

<constraint ident="alt">
<desc>This rule attempts to ensure that information is provided to 
encourage accessibility according to the <ref 
target="http://www.w3.org/TR/WCAG10/#gl-provide-equivalents">W3C Web 
Content Accessibility Guidelines on providing alternatives to auditory 
and visual content</ref>.  If HTML is not a target output this rule 
could be safely deleted, and it is not strictly part of the TEI Abstract 
Model.</desc>
  <sch:pattern name="Alt tags">
   <sch:rule context="tei:figure">
    <sch:report test="not(tei:figDesc or tei:head)"> You should
        provide information in a figure from which
        we can construct an alt attribute in HTML </sch:report>
   </sch:rule>
  </sch:pattern>
</constraint>

> my feeling at the moment is to keep it simple, and only
> use Grp for syntax (ie @scheme). if you want grouped
> constraints, we can extend <constraint> to model and
> attribute classes

I think that just makes it messier.  I'm happy enough to be forced to 
have a <constraintGrp> even if I have only one constraint.

-James
-- 
Dr James Cummings, Research Technologies Service, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk