[tei-council] Should Roma be doing this?

[Syd Bauman]

That's about the result I expected, and is exactly the kind of thing
I think we should document. (Although, as with XSD, say of RELAX NG
that a basic processor will give you no knowledge of the default
value, and one that follows (the ISO equivalent of) _RELAX NG DTD
Compatibility Annotations_ will give you the default value.)

And while it may be unsatisfactory, it isn't of *our* doing, and
there isn't much we can do about it[1] except try to help our users
wade through it.

Notes
-----
[1] Yes, the entire TEI community could stand and cry out to the W3C
    in unison "schema-added information to an XML document is evil,
    please stop it". I wonder if they'd even waste the breath to
    laugh at us.

> > Shouldn't the Guidelines explain exactly what happens when an
> > element that has a default attribute value specified in its ODD is
> > processed by a DTD, RELAX NG, XML Schema, or non-validating
> > processor?
> 
> I did just look at this, and I see we do generate the default
> formally in XSD:
>      <xs:attribute name="type" default="simple">
> 
> so we can be in a number of different situations when we process a
> document containing <list>:
> 
>  1. we use no schema at all: we get no knowledge of "simple" 2. we
>  refer to an XSD but use a no-frills XSL processor: we get no
>  knowledge of "simple" 3. we refer to a XSD and use a full
>  schema-aware XSL parser: we can see type="simple" but its not
>  auto-interpolated (?) 4. we refer to a DTD and use a full XML
>  parser: we have type="simple" interpolated for us 5. we refer to a
>  DTD and use a minimal XML parser: we get no knowledge of "simple" 6.
>  we refere to a RELAX NG schema: we get no knowledge of "simple"
>  except in help during editing by some editors
> 
> pretty unsatisfactory, isn't it?