[tei-council] Styling of TEI <ident> in HTML Guidelines

[Lou's Laptop]

Syd Bauman wrote:
>
>>>    1   <name type='class'>
>>>    5   <name type='datatype'>
>>>   38   <name type='xpscheme'>
>>>       
>> i wonder about these. not <ident>?
>>     
>
> Well, I claim that these (and many, many, things encoded as <ident>
> in the Guidelines) are more naturally described as <name>s than as
> <ident>s. 

Just for the record, I think this is mistaken. Leaving aside the 
slightly curious notion of "more natural", the point is that <name> is 
more general in its applicaion than <ident>. All <ident>s are <name>s, 
but not all <name>s are <ident>s. An <ident> is the name which is used 
specifically and solely to identify the entity concerned.

> I say this in part because in natural language we say that
> the encoded strings are the names of the entities being discussed.
> E.g., we say that "att.datable.w3c" is the name of a class which
> confers upon its members a set of dating attributes that are
> restricted to using W3C format dates in their values. 
We also say that that class has an identifier, namely "att.datable.w3c".

> Personally, I
> prefer to reserve <ident> for strings that more uniquely pick out an
> entity than a name does. Others use <ident> for any name from a
> formal language.
>
>   
I don't know how you could "more uniquely" pick out the att.datable.w3c 
class than by supplying its identifier -- the string used as the value 
for its @ident attribute, no less.
> However, I'm all in favor of consistency, so I think either
> a) they should all be changed to <ident>; or, better still,
> b) all of the <ident>s should be changed to <name>s. 
>
>   
As mentioned earlier I made the use of <ident> now consistent across P5 
last week. The only things still tagged as <name>s are (a) names of 
programs e.g. jing, (b) half a dozen or so "xpointer scheme" names which 
I suspect are not used as identifiers.