[tei-council] another High Noon proposal

[Martin Holmes]

I'm definitely against Lou and Syd here. I think deleting attributes is 
absolutely fine. The attribute classes are a convenience, not a typology.

Especially when you look at something like this:

> att.media would be a member of att.mediaURLless which provides all
> the attrs *except* @url, and att.media defines @url itself. Then
> <mediaObject> is a member of att.mediaURLless, and <binaryObject> of
> att.media.

Imagine being a regular user working on a customization, and trying to 
figure that out. It's far weirder than a deleted attribute.

Cheers,
Martin

On 13-01-11 05:42 AM, Syd Bauman wrote:
> I am leaning towards agreeing with Lou on this. I am reminded of the
> scene in _Raiders of the Lost Ark_ in which Indiana Jones and صلاح
> (Sallah) have asked the wise man Omar to read the markings on a
> medallion.
> ---------
> Omar: This were the old way, this says "six Kadan height - "
> Indiana: About seventy-two inches.
> Omar: Wait!
> [turns medallion over]
> Omar: "And take back one Kadan, to honor the Hebrew God whose ark this is."
> Indiana: Balloq's medallion only had writing on one side? You sure about that?
> Sallah: Positive!
> Indiana: Balloq's staff is too long.
> Indiana, Sallah: They're digging in the wrong place!
> ---------
> Consultant: This is P5, this says "member of att.typed"
> User: Has @type and @subtype
> Consultant: Wait!
> [scrolls down]
> Consultant: "but take back @subtype, we don't really want it."
> User: But *we* really want to use @subtype. You sure about that?
> Scholar: Positive!
> User: Then we have to put @subtype back.
> User, Scholar: We have to use our own namespace!
> ---------
>
> I'm leaning towards "as an editing convention, we generally don't use
> attDef mode="delete" in the source of the TEI Guidelines, and if we
> do, it's with great care"; but also to re-arranging the attribute
> classes to provide what we need. (Yes, it's a pain, but ...)
>
> att.typed becomes a class that is a member of att.typedGrossly which
> provides only @type, and also provided @subtype.
>
> att.media would be a member of att.mediaURLless which provides all
> the attrs *except* @url, and att.media defines @url itself. Then
> <mediaObject> is a member of att.mediaURLless, and <binaryObject> of
> att.media.
>
> I'm not sure that latter is needed. For @subytpe, it's perfectly
> reasonable that a project would want to add it back, so we should put
> in the effort to make sure that's not difficult. But for @url, if it
> really is non-sensical to have it, perhaps the fact that it would be
> hard for a user to add it back is not a strong argument, and we
> should in this case resort to deleting it from within the Guidelines.
>
> Hmmm ...
>

-- 
Martin Holmes
University of Victoria Humanities Computing and Media Centre
(mholmes at uvic.ca)