search

raffazizzi / TEI-TEST

This repo is for TESTING purposes only. Changes pushed here will be lost
0 stars 0 forks source link

inheritance of @xml:lang #103

Closed raffazizzi closed 9 years ago

raffazizzi commented 13 years ago

http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.global.html says, "If no value is specified for xml:lang, the xml:lang value for the immediately enclosing element is inherited; for this reason, a value should always be specified on the outermost element (TEI)."

I dislike this because it assumes that the primary language of the content of of the header are the same. While this is often the case, there are entire collections where this is not the case, and many collections involve a mix of documents in which this is only sometimes the case.

I think it would promote clearer thinking and more straightforward interoperability if this were revised to say: "If no value is specified for xml:lang, the xml:lang value for the immediately enclosing element is inherited; for this reason, a value should always be specified on both teiHeader for the primary language of the metadata and on text (or fsDecl or facsimile) for the primary language of the content."

Original comment by: kshawkin

raffazizzi commented 13 years ago

The proposed change seems quite plausible, though I think I would phrase it more loosely, since the revision doesnt really help in cases where you have say a <group> of texts using different languages.

How about "... the xml:lang value will be inherited from the immediately enclosing element, or from its parent, and so on up the document hierarchy. It is generally good practice to specify xml:lang at the highest appropriate level, noticing that a different default may be needed for the teiHeader from that needed for the associated resource element or elements, and that a single TEI document may contain texts in many languages."

Original comment by: lb42

raffazizzi commented 13 years ago

Original comment by: lb42

raffazizzi commented 13 years ago

Original comment by: lb42

raffazizzi commented 13 years ago

Original comment by: lb42

raffazizzi commented 13 years ago

I had forgotten about <group>. While I am at first inclined to simply change my proposed text to say

and on text (or group, fsDecl, or facsimile)

I can live with Lou's proposed text. It's simpler and easier to maintain in the long run (in case we implement the pending genetic editions thing). Unfortunately, it means that people processing TEI documents from various sources will need to look at TEI@xml:lang in case <text>, <group>, <fsDecl>, or <facsimile> is missing an @xml:lang. That is, I was trying to promote clearer thinking about the difference between metadata and content and to facilitate straightforward processing.

Original comment by: kshawkin

raffazizzi commented 13 years ago

I prefer Lou's wording, in part because many documents (perhaps even the majority) will still be fine with @xml:lang on the TEI element. I don't think enumerating approprite high-level elements will help much, since sometimes languages change at the level of divs, or differ between front and body, for example.

Original comment by: gabrielbodard

raffazizzi commented 13 years ago

I agree with Lou's wording, it makes it looser, I didn't like the "should always".

Original comment by: jamescummings

raffazizzi commented 13 years ago

Added new wording at rev 9636

Original comment by: lb42

raffazizzi commented 13 years ago

Original comment by: lb42