[tei-council] soft deprecation of @key

[Lou Burnard]

On 20/05/12 22:39, Kevin Hawkins wrote:
> On 5/20/2012 5:17 PM, Lou Burnard wrote:
>> On 20/05/12 20:47, Kevin Hawkins wrote:
>>> All,
>>>
>>> At last, I've implemented http://purl.org/TEI/fr/3437509.  See the "text
>>> changed" links at in case you're interested:
>>>
>>> http://tei.svn.sourceforge.net/viewvc/tei?view=revision&revision=10374
>>>
>>> Lou, over to you to review.
>>>
>>> --Kevin
>>
>>
>> Well, I think the proposed wording might be improved. The reference to
>> <taxonomy>   seems irrelevant -- if you use that, you'd use @ref to point
>> at entries in it.
>
> But couldn't someone use<taxonomy>  to define a controlled vocabulary of
> terms used in @key throughout the document?

I suppose they could, though properly speaking <taxonomy> is for 
classification, not just controlled vocabulary.

  Why can only a @ref point
> to<taxonomy>?

My point was that if you *have* gone to the trouble of defining all your 
terms in a taxonomy, there's nothing to be gained by not using the 
simpler @ref="#foo" type syntax.


>
>   >  And the term "magic token" is probably best not
>> eternalised (magic cookie I've never heard of before -- it invites
>> confusion with the other sorts of cookie.
>
> I was taking a cue from:
>
> http://wiki.tei-c.org/index.php/TEI-Council-FAQ#What_is_a_.22magic_token.22.3F
>
> in which Stuart helpfully added a link to the Wikipedia page on "magic
> cookie".  It appears that in computer science, people call it that.  It
> seems that we should acknowledge the terminology of this community.

I'd rather not since that's not really our target community, but if you 
insist, I'd use that term only, and maybe reference the Wikipedia page.

>>
>> No particular syntax is proposed for the values of the<att>key</att>
>> attribute, since its form will depend entirely on practice within a
>> given project. For the same reason, this attribute is not recommended in
>> data interchange, since there is no way of ensuring that the values used
>> by one project are distinct from those used by another. In such a
>> situation, a preferable ...
>
> I prefer the stronger guidance in my version.  In yours, you say "in
> such a situation" (of data interchange) as if the TEI isn't always meant
> to promote data interchange.

Well, this may be something we have to agree to differ on. Although TEI 
is of course primarily for data interchange, it's not exclusively so.

I think of @key values as being like the PUA in Unicode -- they're 
useful, but you need to know what the risks of using them are.

By the way, did we actually agree to translate *every single one* of the 
@key instances into @ref:tag:wibble,foo ones? I thought we'd agreed to 
keep some, to show that the old methods were still usable.