[tei-council] [Fwd: Re: recording multiword expression in lemma attribute]

[Syd Bauman]

> Making it xsd:token rather than data.word would help with the
> specific case Elena raises, at the expense of making this attribute
> inconsistent with all the other cases of "texty" attributes.

But I would shy away from using xsd:token (which itself imposes no
syntactic constraints at all -- any sequence of Unicode characters is
permitted except for those XML itself does not allow) in any case.
The more appropriate datatype would probably be one or more
occurrences of data.word, which already appears 30 times in the
Guidelines.

Personally, I think I'm in favor, but I don't consider myself an
expert in this arena whatsoever.