I have to agree it's an interested idea. It shouldn't be a big problem to implement it for all three formats.
Right now I'm not sure only about how to mark ORMD texts in the whole comment. ORMD works in way that it removes known annotations and replaces them with updated ones and keeps any other content (another annotations, user comments,....) as it is.
So if we need to export (and continuously update) descriptions, it's necessary to create some marks to distinguish between ORMD descriptions and user's texts. For example to use some special annotations like:
or some marks, for example colon as first char:
* : afsfsafs
* : asfafsdf
In xml format there are also some ways how to implement it. One possible way is to use default XML comments. But this should be a limitation if anyone needs to proceed descriptions from exported XML
<!-- description ..... -->
Another way is to define some ORMD namespace and export these texts in own element in this namespace:
<ormd:description> asdfafsaf </ormd:descriptions>
In YML the situation is very similar as in XML. One way is to use default comments, second way is to define the own elements. I'm not sure if there is a way how to use namespaces, so for example some
I will be glad to open a discussion about this topic. Take this as suggestions how we can implement it. But I will be more than happy to hear how you already export these texts in your script, so we don't have to reinvent the wheel ;-)