I think we might be nit-picking over what constitutes at 'TEI'
document. Yes, in my mind TEI P4 documents are 'TEI' because they
followed the TEI abstract model and the TEI-C supported them for
many years and then 5 years after the launch of TEI P5. If people
created good TEI P4 documents and are happy with the
infrastructure they have to support them, then fair enough. I
would still say they are good TEI citizens and we should relish
our long history; there is no reason for them necessarily to
change (but on their own head be it).
However, the point of this, surely, is that in giving examples to
people who are facing encoding problems at the moment we want to
give the best possible advice we can. At the moment that is TEI
P5 version 2.5.0 (with 2.6.0alpha in preparation). This doesn't
mean that we shouldn't also have examples of TEI P4 around (TEI
P4 is available in the Vault after all), especially in a context
where someone wants a corpus to test tools against which they
want to cope with more than one TEI generation. However, if
someone is asking for advice we certainly wouldn't be pointing
them to TEI P4 as a solution in any way. To do so would be
actively bad practice, IMHO.
If I add to the silly analogies, if someone was asking your
advice to buy a car and their criteria was good fuel economy,
wanted to use it daily for commuting, reliability, and with a
load of useful features that just worked, then you (most likely)
wouldn't suggest they buy an ancient car where they couldn't get
the parts any more. No, if budget allowed, you'd suggest they get
a newer car, maybe customising what features it had. (Why are
electric windows, power steering, central locking and air bags
still considered 'features'? Surely they should be standard by
now!) and knowing that model will be supported for some time by
garages. If asking for an encoding scheme we like to think you'd
suggest a recent version of the TEI, helping them go to a website
to customise what features it has, and knowing that even their
customised version will be supported for some time by the
available tools. While new versions of the same model of car
come out with improved sat-nav maps and electric seat warmers are
introduced as a feature there is nothing that stops you
continuing to use your version or, if you choose, upgrading your
sat-nav to the latest maps. Encoders can continue to use an ODD
based on an earlier version of TEI P5 and still be 'TEI P5' --
they can modify and regenerate their schemas and documentation
having their ODD point at that version of TEI P5 in the Vault.
They can even add third-party extensions and features to their
car. Part of the problem is that for some bits of this they may
need the help of a mechanic.
It is good to remember that TEI isn't and shouldn't be viewed as
an orthodoxy but an inclusive community trying to help people do
whatever it is they need to do.
On 22/11/13 09:41, Lou Burnard wrote:
> Given the way the technical landscape has changed, we would not
> win many friends by recommending people to use the same
> technologies as they did 20 years ago. There is nothing wrong
> with using TEI P4 and associated technologies if you want to,
> though they are (and will become increasingly) hard to find. For
> that reason, if for no other, the TEI announced that it would be
> withdrawing support for P4 well over three yours ago, and
> actually did so last November. Hence Sebastian's point that a TEI
> P4 document is not strictly speaking a TEI document. And, at the
> risk of further confusing the confusible hogs of whom Martin
> Mueller speaks, may I reiterate that the very notion of "TEI
> conformance" did not enter into the TEI world view until
> publication of TEI P5. As Sebastian also notes, conversion tools
> for P4 documents are available, and have been for some time. The
> conversion process is also a LOT simpler than it was the last
> time we had this discussion, when we went from P3 to P4.
> On 22/11/13 09:00, Peter Boot wrote:
>> I would say that from a PR point of view this is an undesirable
>> position. We can't on the one hand argue that people should use
>> TEI to be prepared for changes in the technology landscape and
>> on the other hand relegate their documents to the dustbin of
>> history if some changes actually happen.
>> The idea that only those documents that are technically TEI P5
>> conformant can count as TEI documents seems very restrictive to
>> Van: TEI (Text Encoding Initiative) public discussion list
>> [[log in to unmask]] namens Lou Burnard
>> [[log in to unmask]]
>> Verzonden: donderdag 21 november 2013 23:15
>> Aan: [log in to unmask]
>> Onderwerp: Re: [TEI-L] TEI Examples
>> The concept of "valid TEI document" was defined in TEI P5. P4
>> documents, though certainly "TEI" in some rather vague sense,
>> cannot be considered valid by that definition.
>> They do of course remain valid TEI P4 documents -- except that
>> I don't think there's any definition in TEI P4 of what exactly
>> that might mean.
>> What's disappointing about that?
>> On 21/11/13 22:10, Stuart A. Yeates wrote:
>> Is it your contention that all valid P4 TEI documents became
>> 'not TEI' overnight when P5 was released? That all documents
>> that were once valid against the then-standard but are not
>> valid against the current standard are 'not TEI' ?
>> I would be disappointed to see this being adopted as an
>> official position.
Dr James Cummings, [log in to unmask]
Academic IT Services, University of Oxford