[tei-council] TEI Conformance

Conal Tuohy Conal.Tuohy at vuw.ac.nz
Wed Nov 22 09:38:47 EST 2006


Hi James

I can see what you're saying, but I still think a single "TEI Extension" namespace would be a mistake, and I'll explain why.

Firstly, if people want to define elements for "private use" only, i.e. ruling out interchange altogether, then it doesn't matter what namespace they use - even the main TEI one, or better yet, none at all. If the document is never interchanged then it doesn't really matter what they do.

However, if the document is ever to be interchanged then it's important that names don't collide, and if we encourage shared use of an extension namespace then we'd have to manage that namespace to avoid name collisions. I don't see what that gives us except extra administrative work. 

I think it would be very odd to say (even in big letters) "this is a special namespace in which name collisions are only to be expected", because the avoidance of namespace collisions is literally the only purpose of XML namespaces. The analogy with Unicode's Private Use Area is not valid, since Unicode only needs the PUA because it lacks namespaces. 

Where name collision is considered tolerable, I think it would be better to define extensions in no namespace at all, because at least there's a universal recognition that non-namespaced elements are prone to collision. If a user doesn't care to protect their customisation against name collision then let them not use namespaces.

Finally, can I ask: what's the actual benefit of defining a special "extension" namespace? It's not like by providing a "TEI Extension" namespace we would save anyone significant work since literally ALL they have to do is make up a unique URI for their namespace.

I very much like the idea of "incubating" vocabularies and importing them into the core later, but for this to work over the long term I think we will need to be able to incubate each vocabulary independently, in a namespace of its own. 

Cheers!

Con


-----Original Message-----
From: James Cummings [mailto:James.Cummings at computing-services.oxford.ac.uk]
Sent: Thu 23/11/06 0:41
To: Conal Tuohy
Cc: TEI Council
Subject: Re: [tei-council] TEI Conformance
 
Conal Tuohy wrote:
> James Cummings wrote:
> 
>> And if people were to add their entirely new elements in a different
>> namespace, then it also makes sense for TEI to suggest a namespace for this
>> (rather than every user come up with one).
> 
> I beg to differ. It seems to me that just because user A and user B having
> chosen the same name for a new element they have each defined, we are not
> entitled to assume that the 2 elements are related, which I think would be
> implied by qualifying them in a generic "TEI Extension" namespace. I think
> that new elements should always be qualified by namespace URIs belonging to
> the user defining the elements.

I recognise that worry and I'm not entirely sure it is misplaced.

However, I was thinking about the tei-ext namespace as similar to the Unicode
Private Use area.  If I create a new element tei-ext:article, then I use this
because it is not docbook:article, and I want to say this is a TEI-like element
rather than another one.  If you create an element tei-ext:article you are in no
way saying it is the same semantically to my tei-ext:article.  We would have to
say in big letters that tei-ext does not guarantee or in any way indicate
anything about the uniqueness of the elements defined in it.  While I think we
shouldn't pollute the TEI namespace, I don't think it is wrong to have a
separate namespace that we encourage people to pollute as long as we are very
clear that it shouldn't be understood by applications as anything more than that.

Another option is for a 3fold system:

- TEI namespace for TEI elements and their equivalents
- TEI Extension namespace for elements being 'incubated' as possible contenders
to be TEI elements submitted by members/SIGs/etc As part of a concerted and
documented proposal
- MyPersonalNamespace -- user defined namespaces for specific new elements which
they don't think should become part of TEI

-James


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




More information about the tei-council mailing list