Closed kosloot closed 4 years ago
No, this is not a bug. The default that is implemented is the set only. So if one has the following declaration:
<text-annotation />
Then it is equivalent to:
<text-annotation set=""https://raw.githubusercontent.com/proycon/folia/master/setdefinitions/text.foliaset.ttl" />
I'd ideally opt for the verbose form in any serialisation btw..
But if the there is no declaration altogether, then that simply means the document has no text (no <t>
). The same goes for phon
, these are the only two elements where FoLiA implements a default set.
Also this warning is quite confusing, as it mentions TextContent, and not Text. Wouldn't 'textcontent-annotation' have been a better idea, to express this relation?
The names can be a bit confusing here, I know, but for the end user it's probably clearer this way. TextContent (t) has annotationtype TEXT, the text body/root element is a special has no annotationtype and needs not be declared. Ideally we should have renamed the root element perhaps.
Yeah lets rename the root element :) Too late for that.
considering
<text-annotation />
I think this may be dangerous, suggesting users that ANY annotation may be defined in this way, omitting sets, processor etc. I assume this is thoroughly checked to be ONLY applicable for \<phon> and \<t>. Have to check this for de C++ library too.
For all other declarations except for text and phon that are defined that way it explicitly means: we use this annotation type without a set (i.e. no classes anywhere on the annotations).
Hmm. the I REALLY have to check the C++ library. Although is assume there are already testcases for all this.
The validator rejects documents without a text-annotation declaration.
I was under the impression that in such cases a default should be implied, like this:
Also this warning is quite confusing, as it mentions TextContent, and not Text. Wouldn't 'textcontent-annotation' have been a better idea, to express this relation? And also 'phoncontent-annotation'...