> On 18 Mar 2021, at 13:02, nathan.art...@softhome.net wrote: > > Jan David Hauck via Bibdesk-users writes: >> This is a nice change. But I'm curious, what does "unique characters based >> on the item" mean, i.e., if it's not random what information from the item >> does it use? >> Also, it seems that %u0 still generates citekeys starting with "a" (for the >> first one that would clash), but when I change it to %u1 it generates a >> different letter. Just curious what the logic here is. > > Christiaan can explain the details; I will just note that the ticket related > to this change, with related discussion, is here: > https://sourceforge.net/p/bibdesk/feature-requests/878/ >
That’s right. Its is based on the title of the item (and when this is missing the DOI). It is a hash of the title interpreted in base-26 (i.e. using letters rather than decimal digits), and then the first few base-26 “digits”, in opposite order. We don’t do this with %u0 and related specifiers, because they won’t satisfy being deterministic anyway, because they default to nothing. Also, when you have more than one file, it makes more sense to have them numbered successively (authorYear.pdf, authorYear-1.pdf, authorYear-2.pdf, …) rather than "randomly". Christiaan
_______________________________________________ Bibdesk-users mailing list Bibdesk-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bibdesk-users
