Closed raffazizzi closed 1 year ago
No, I think that (RNG does not have @subtype
) is an old problem, which already has a ticket. (Will look for ticket # later.)
So while the Stylesheets may be doing exactly what you want (which, as I said, I am quite impressed with), I am not sure it is what we want them to do. Do we want the output of a cherry-pick like this to look the same as a <memberOf>
and "change"
(or "replace"
)? Maybe so. But I am not sure. Give me a night to sleep on it, as it were.
We may want to find a way to distinguish inheritance via membership from inheritance via cherry-picking, somehow, but where the cherry-picked attribute comes from is important and should be shown.
In the spec pages for attribute classes, this is already shown; e.g. in att.textCritical you can already see:
att.written (
@hand
) att.typed (type,@subtype
)
So I reproduced the same in element spec pages. I would say that for now this is fine so that this information is made available. We can then discuss how to distinguish these two forms of inheritance.
E.g. the new display on lem is:
For membership inheritance (NB lem is not a direct member of att.global.rendition)
att.global (
@xml:id
,@n
,@xml:lang
,@xml:base
,@xml:space
) (att.global.rendition (@rend
,@style
,@rendition
))
For both membership and cherry-picked:
[att.textCritical]() (
@type
,@cause
,@varSeq
,@require
) ([att.typed]() (type,@subtype
)) ([att.written]() (@hand
))
@peterstadler the space problem should be fixed! Please double check when you get a chance and please merge if it looks good!
Merging per Stylesheets mtg 2022-01-05
Thanks for this feedback, @sydb! The plot thickens:
lem
(which was the element I looked at for testing) is member ofatt.textCritical
just as much asrdg
is: they should both get@subtype
fromatt.typed
becauseatt.textCritical
usesattRef
to cherry-pick it. So the stylesheet is doing exactly what I told it to do, as far as I can tell.In short:
lem
andrdg
are both members ofatt.textCritical
, which cherry-picks with<attRef>
@subtype
fromatt.typed
.But actually the latest
tei_all.rng
does not allow@subtype
on either of them.@subtype
does not make it into the RNG definition ofatt.textCritical.attributes
. So it looks like we have a new problem, right?