On Thu, 05 May 2011, Samuele Kaplun wrote:
> just to know, why for arXiv would be better the above policy? 

It emulates the current arXiv workings better, and it offers an easy way
to display, cite, preserve, and otherwise access or manipulate every
user-submitted version of the paper without having to dig inside
technical record history.  For example, it strictly separates
user-submitted changes (that would create new records for each revision)
from theoretical technical curation changes and fixes (that would keep
the same record ID but create new internal revision).

> Wouldn't be better to simply keep one record, and use the file
> revisioning?

Note also that this concerns author-submitted metadata revisions as
well, not only author-submitted file revisions.  See for example
<http://arxiv.org/abs/0810.0012v1>.  All the revisions should be
independently citeable, etc.

Best regards
-- 
Tibor Simko

Reply via email to