[tei-council] Glosses, glosses, everywhere, and what do you all think?

[Lou Burnard]

Syd Bauman wrote:

  ---------
> 
> Should "gi" (as in <gi>, gi= of <tagUsage>, locus="gi" of <certainty>
> or <respons>) be glossed as "generic identifier" or as "element
> name"?
> 
> ---------

Interesting question. The problem is that we have another element 
(<ident>) which is presumably glossed as "identifier", so the clear 
implication for those who don't know the back history is that <gi> is 
somehow more general than <ident>. Whereas the reverse is actually the 
case -- <gi> is very specifically only for one kind of identifier: an 
element name. It's called <gi> because that's what it was called in SGML!

So my opinion is that the gloss should actually say "XML element name", 
but that there should be a sentence in the <remarks> explaining why it's 
called "gi". We do the same kind of thing for <g>

A more radical proposal which occurs to me this morning is that we 
should maybe merge this element with <tag> (and call it <tag>). Few have 
ever understood the distinction, and even in the text of the Guidelines 
I don't think it's applied consistently. Which suggests it's not a 
distinction really worth making.

> 
> data.outputMeasurement has the gloss "HTML dimension". See
> http://www.tei-c.org/release/doc/tei-p5-doc/html/ref-data.outputMeasurement.html.
> Some thoughts:
> 
> * I think there should be some examples in there: most people don't
>   casually read regular expressions the way many of us do. I plan to
>   add some soon.
> 
> * The gloss is clearly not right. This datatype is currently only
>   used for width= and height= of <graphic> and <binaryObject>. But
>   does it need a gloss? Should its name be changed instead (or in
>   addition)?

Changing its name might be a good idea, but it certainly needs no gloss 
in its current form, since its name is composed of dictionary words.

> 
> ---------
> 
> I found that most model.*Like classes did not have glosses, but a few
> did. As the "class made up of elements that are in some way like *"
> semantic is consistent, I removed those few glosses.
> 
> ---------

see rule stated above

> 
> As for the doctype= attribute of <refsDecl>, I gave it a gloss, but
> I'm not sure I understand this attribute. I *think* it might have
> been intended to deal with using different reference declarations for
> each of a variety of DTDs which might have been used at the same time
> ala CONCUR, but I'm not sure.

yes, I think so too: look at the Lachman example in the old version of NH.


> 
> ---------
> 
> We gave the ident= of <language> its name in order to parallel the
> ident= of <elementSpec>, <valItem>, etc. I don't wonder if a more
> semantically appropriate name would be better. RFC 3066 and its
> successor RFC 4646 call the thing they describe (which is the value
> of ident= of <language>) a "tag". Back when we were setting this up I
> thought that tag= was too confusing a name for an attribute. But now
> I'm not so sure. It seems better than ident=. Could use langTag= or
> languageIdentifier= or rfc4646= (ick), I suppose. 
> 
> Note that for <langKnown> we use tag= ('cause there's only one), and
> for <langKnowledge> we use tags= ('cause there's likely more than
> one). So I'm leaning towards tag=.
> 
> Thoughts?
> 

Notwithstanding what I said above (about the confusing <tag> element), I 
am half inclined to agree. The other half of my inclination is to apply 
a different general principle, keep this as @ident, and make the current 
@tag/s into @key/s

> ---------
> 
> I did not gloss "hand" as in @hand, <hand>, <handList>, <handShift>,
> etc. Speak up if you think it should be glossed (and say what the
> gloss should be).
> 
> ---------

It should not be glossed. It is a dictionary word.

> 
> Why doesn't <geo> have the military grid reference system as a value
> of datum=?

I am not sure what you mean. Are you saying that the default datum 
should be different from what it is, or that the datum should be 
specified on <geo>? If the former, we were assured by at least one 
expert at the Kings meeting that the WDS was the right choice and have 
seen no evidence to the contrary; if the latter, because that's what 
<geoDecl> is for.

> 
> Does anyone have strong opinions about whether proper nouns like
> "World Geodetic System" should be glossed with initial caps or not?
> 
> ---------

They shouldn't be glossed at all: they are dictionary words. But WGS 
gets capitals because it is a proper name.,

> 
> The most amusing gloss I deleted was that for 'model.recordingPart',
> which read "dates and date ranges". :-)
> 

Ah Kut and Paste, what strange paths they those comedians lead us into ...


> _______________________________________________
> tei-council mailing list
> tei-council at lists.village.Virginia.EDU
> http://lists.village.Virginia.EDU/mailman/listinfo/tei-council
>