Sune Karlson writes
I just stumbled across what might be a bug (or feature 8-) in how short-ids are generated. It is my understanding that the short-id is generated by concatenating 'p', the first two ascii characters in the last name and a sequence number. But there is at least one registered person with less than two ascii characters in the registered last name, http://econpapers.repec.org/RAS/pa1.htm, leading to an unexpected form for the short-id 'pa1'.
It is probably highly unusual but I guess possible for someone to have a "western" last name with no ascii characters. If such a person registers and does not provide an "english" version of the name we could end up with a short-id 'p1' and that causes problems given the way the files are organized in the per archive.
Ivan Kurmanov writes
i guess we could use "zz" as the default prefix, if the lastname does not provide us with enough letters. (and a similar string for non-personal short-ids.)
Could _ be used? I could try a patch, then test it. Cheers, Thomas Krichel http://openlib.org/home/krichel RePEc:per:1965-06-05:thomas_krichel skype: thomaskrichel