attempted fix for #2758 #2759
attempted fix for #2758 #2759
Conversation
P5/Source/Specs/sp.xml
Outdated
| <alternate> | ||
| <alternate minOccurs="1" maxOccurs="unbounded"> | ||
| <classRef key="model.global"/> | ||
| <elementRef key="speaker"/> |
There was a problem hiding this comment.
This would allow any number of speaker elements at any position inside sp, wouldn't it?
There was a problem hiding this comment.
Yes, this solution is overly permissive in that sense—particularly in allowing <speaker> anywhere. (Changing it to zero or one is probably not so big a problem to fix as limiting where it can occur.) We could add Schematron to constrain it further, and I figured it would hold if we can't find a better way quickly! (But I suspect we can.) I hope this is at least a starting point for a better solution.
…is not required and b) it is deterministic (aka non-ambiguous)
… content model a bit.
|
@ebeshero opened this PR, and thus apparently cannot be a reviewer. Nonetheless, if anyone can check this branch out and test to see if your In the meantime we should start setting up for a patch release soon. (E.g., establish Jenkins jobs.) |
|
Okay, I can confirm that @sydb 's fix not only passes tests, but is also working when I build tei_all.rng and validate my test file (attached).
So I think this will suffice for a patch. But I am wondering if we really want to continue with the complexity of this new content model for <content>
<sequence>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
<alternate>
<sequence>
<sequence minOccurs="1" maxOccurs="unbounded">
<classRef key="model.stageLike" minOccurs="1" maxOccurs="1"/>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
</sequence>
<alternate>
<sequence>
<elementRef key="speaker" minOccurs="1" maxOccurs="1"/>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
<sequence minOccurs="1" maxOccurs="unbounded">
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
<classRef key="model.stageLike"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
</sequence>
</sequence>
<sequence>
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
<sequence minOccurs="0" maxOccurs="unbounded">
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
<classRef key="model.stageLike"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
</sequence>
</sequence>
</alternate>
</sequence>
<sequence>
<elementRef key="speaker" minOccurs="1" maxOccurs="1"/>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
<sequence minOccurs="1" maxOccurs="unbounded">
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
<classRef key="model.stageLike"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
</sequence>
</sequence>
<sequence>
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
<sequence minOccurs="0" maxOccurs="unbounded">
<alternate minOccurs="1" maxOccurs="1">
<elementRef key="lg"/>
<classRef key="model.lLike"/>
<classRef key="model.pLike"/>
<classRef key="model.listLike"/>
<classRef key="model.attributable"/>
<elementRef key="q"/>
<classRef key="model.stageLike"/>
</alternate>
<classRef key="model.global" minOccurs="0" maxOccurs="unbounded"/>
</sequence>
</sequence>
</alternate>
</sequence>
</content>
Is it too much? Would introducing |
|
Also @sydb , whether you take that as a review or not, I think in the interests of time, any one of us should review your solution on this branch. The fact that I made the branch shouldn't matter at this point. I will argue that a simpler content model might be desirable for human readability. Whether mine is too simple depends on how imperative we find constraining the position of |
|
Keep in mind the content model is a lot simpler after we create the “spPart_con” and “spPart_sans” classes (or macros), whatever they may be named. |
|
BTW, I do not think this PR should be accepted as-is — the changes in "ebb-fix-#2758" should be merged into a new "release-4.10.2" branch, not "dev" (which is what this PR would do). |
|
Adding stage to model.global would permit retaining the constraint that speaker if present must precede what is spoken which i think is important. And is obviously much simpler. |
|
The proposed content model requires |
|
@sydb my point was not that the proposed content model doesn't do the jpob -- clearly it does. My point is that the same objective might be obtained with a much simpler content model , just by making stage a global element. And simplicity and consequent explainability are desirable. |
|
Closing this PR since we transferred to the work of it to the 4.10.2 release branch, made from 4.10.1. |
@sydb @martindholmes @raffazizzi : This branch isn't passing tests yet, but I think we need a few more eyes on it to figure out how to fix it.
It's an attempt to solve #2758 on the way to (most likely) another patch release.