Chunking review comments

["Kristen Eberlein" <kris@eberleinconsulting.com>]

Below is a list of some of the comments that were made in the DITAweb review of chunking material:

1) "Somewhere, we should capture the effects of chunk="combine" on topic <prolog> metadata. Specifically, all <prolog> metadata in subordinate "chunked" topics gets removed from generated output. Probably nothing that we can do about it, but users should be surprised when all their beautiful metadata disappears. :-)

2) "If the split was on a topicref for a single topic with sections, could it split the sections as individual docs?"

3)"What are the common use cases for using @chunk to "split"?"

4) "The fact that processors may create their own naming scheme has caused my clients quite a bit of grief because that naming scheme is typically smething like "chunk12345" and that number changes every time the content is processed, meaning there is no predictable file name to reference in a link. We have some customers therefore using @copy-to to provide a predictable name. But that attribute is going away, correct? So what is the recommendation for this situation?Obviously my experience is all on the combine side, so being able to specify the appropriate file name for the single output could be done with a single attribute. I don't know how the split situaiton would work, which is likely why there is no way to specify the names. But I'm still curious about recommendations. ...but once I got to the last example, I realized this issue is addressed. Could we add something to the bullet here that authors can specify a predictable file name using keys as shown in the last example?"