[tei-council] conformance draft

[James Cummings]

Lou Burnard wrote:
> 1. The definition of "recommended practice" used to have a third point:
> 
> "all textual features which the guidelines recommend be captured are in 
> fact encoded. "
> which James removed presumably because he could think of no case where 
> the Guidelines actually  makes recommendations about the level of 
> encoding to be applied (i.e. where it suggests that you really ought to 
> mark up paragraphs, say, or lines in verse)

To be honest, I think Sebastian commented that one out, not me, but I agreed 
with him that I couldn't think of a place where the TEI Guidelines (strongly at 
least) recommend to what level encoding should be applied.  I.e. if I'm doing a 
poem it is fine if I do <lb/> rather than <l>, never mind that I'm not marking 
up <w>.

> 2. The original draft distinguishes two formats for a conformant document:
> "A document is /TEI-conformant/ if it is either in /TEI local processing 
> format/ or in /TEI interchange format/. A full description of the 
> document should specify which format it is in."
> 
> The new draft does not make this distinction up front, although it is 
> discussed later on at 1.3.3.2.  Was this a  deliberate decision or an 
> oversight?   Should we not make the distinction earlier? I can see that
> the case for a third possible format originally discussed  ("TEI packed 
> interchange format") is hard to maintain with the availability of XML, 
> but maybe we ought to preserve the other two?

It was deliberate.  My TEI is in TEI Interchange Format and is TEI Local 
Processing Format.  I thought we should keep the distinction to make those who 
are not-fully-Conformant because of local processing needs, realise that this is 
an acceptable thing to do even if not strictly Conformant.

> 3. The discussion of a "TEI derived schema" (1.2.2 in the current draft) 
> talks about ODD-derived schemas and about exemplars, but does not 
> mention the feasibility of constructing a TEI schema by combining either 
> DTD fragments or RelaxNG fragments. Are we specifically outlawing that 
> approach, or are we just passing over it in silence? I don't think 
> either approach will win us any friends.

I was wondering who would raise that first -- that is before Wendell sees it -- 
I was intentionally being hard-line dictatorial and stating that TEI Conformance 
requires there to have been an ODD at some point to generate the schema. I think 
we should outlaw that approach for TEI Conformance.  I'm happy for people to do 
it, but think we should have a No-ODD, No-Conformance, approach.  I don't feel 
that people who will be doing that will be as concerned about Conformance in any 
case.

> 4. The suggestion that a sourceDesc element is no longer required seems 
> unwise to me.  If a document is "born digital" the fact should be 
> stated, and the sourceDesc is the place to do it.

Yes, and that is the suggestion made.  I just wondered about making it a 
requirement for *any* TEI document (which is what the old description did).

> 5. 1.2.5 introduces a new term "TEI compliant" but does not seem to 
> define it anywhere: is it meant to be synonymous with "TEI conformant"?

Typo.

> 6. As others have suggested,  the list of exemplars should probably move 
> somewhere else. Except that the definition of tei_all is important, 
> since it's used in the following section to define varieties of 
> conformant schemas.

Since Modification is now a section of the Using TEI chapter, as long as that 
goes before the Conformance section, then that would seem an appropriate place 
to discuss these in depth.

> 7. I am a bit unclear about  the purpose of a "TEI Supported Extension 
> Schema". Can anyone suggest one? what does "is supported (to some 
> degree) by the TEI" mean?

I.e. TEI with XInclude, TEI with SVG in graphic, etc.  Customisations that are 
not pure TEI which the TEI produces.

> 8. The last bullet point in the list at the start of 1.3.3 (requiring 
> the existence of an ODD to document the schema used by a non conformant 
> customization) while I agree with it in principle is (a) not a property 
> of a document (b) somewhat at variance with my point 3 above.

This list is meant to be identical as the one at 1.2.


> 9. I don't like section 1.7 much: it makes me feel uneasy. We are not in 
> a position to tell funding bodies what they should or shouldn't think, 
> and even if we were it shouldn't be a topic for this chapter. We ought 
> to be clear first about what *we* mean by TEI conformance, and although 
> I think this discussion is clarifying that quite a lot, we ain't there 
> yet. Issues of "superior quality" and "greater academic scrutiny" surely 
> relate to matters which are out of scope here (see for example my point 
> 1 above)

Agreed. But I still feel it is important to stress (to them and others) that 
Conformance != Quality.

> 
> Let me repeat that I think this draft is definitely going in the right 
> direction, despite these somewhat picky comments! I think it's really 
> important to get a good version of this chapter into draft 1.0 so I hope 
> we can focus on that for a while..

I think all these comments are the kind I was hoping to get.


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