One of the things that really annoys me about bibtex is the lack of
ability to distinguish between magazines and research journal articles.
I admit to not having looked closely enough at the XLST, it scares me
a little, to know whether citeproc handles this better? I realize
that one person’s magazines might be another’s journal, so it can be
a judgement call…
–J
Internally, these should get assigned types like “article-journal”,
“article-magazine”, etc. E.g. there’s a function that just strings
together types for the different levels/relations, so you can also in
theory end up with “paper-conference”, “chapter-book”, and so forth.*
In the absense of reftype definitions for those specific types,
citeproc will use the generic fallback: “article” in this case. I have
never found a need to use anything but that generic definition.
So, yeah, I think it works better than bibtex 
Bruce
Yes, that’s the thing, they most likely ought to be ‘laid out’ the
same way, but stored as different types. I like the approach.
–J