Re: [jats-list] Versions of an article

[Kevin Hawkins <kevin.s.hawkins@xxxxxxxxxxxxxxxxxx>]

We need to distinguish between DOIs created through a membership in 
CrossRef and other DOIs.  Since many of us only use CrossRef for DOIs, 
we come to think of their rules as being the only rules.

The International DOI Foundation explicitly has no policy on whether you 
should assign a new DOI when the content has changed slightly:

http://www.doi.org/faq.html

On the other hand, CrossRef's publisher rules ( 
http://www.crossref.org/02publishers/59pub_rules.html ) says:

12. CrossRef only registers DOIs for Definitive Works (or Versions of 
Record, if not formally published) but not for Duplicative Works, as 
defined in the CrossRef Glossary. This means that only original 
scholarly material, for which there is no actual DOI at the time of 
submission, and no expected duplication in future, is admissible for 
CrossRef DOI registration. CrossRef does not permit multiple DOIs to be 
assigned to  certain closely related versions of a work, and hence does 
not support assignment of DOIs to Pre-prints or Post-prints of 
Definitive Works or to the Personal Version or a Self-archived Copy of a 
Definitive Work. For the same reasons, materials for which DOI 
duplication can be reasonably anticipated, such as an Authors Original 
Draft of a work being prepared for publication, are not admissible for 
CrossRef DOI registration.

However, it appears that if you use CrossMark metadata, you can create 
more than one DOI for the same published work as long as the proper 
CrossMark metadata is attached to each of those DOIs.  That appears to 
be the mechanism used by F1000 Research.

None of this, of course, answers Alf's original question.  :)

--Kevin