Paul POULAIN said:
Just to lower your enthusiasm, let me just remind that the solution must work with english as well as french & other languages. It's really easy to do it just for french (for example), it's more complex to do it for any language.
Actually, the solution NPL uses works with MARC21 but not UNIMARC. MARC21 uses the second indicator of the 245a subfield to tell how many characters to skip before "indexing" the title. So if a title starts with "the " the second indicator is 4 (skip three letters and a space to get to the correct "filing" character). From what I can see, UNIMARC has no such handy indicator. Certainly a language issue, as you suggest. [This is probably more appropriate for the developers list, but I sometimes think it might be a good idea to start considering "flavours" of Koha, just as there are flavours of MARC. We loose lots of capability that is built into MARC when we try to engineer a solution that will work for everbody all of the time.]
<stephen> I think the best solution could be to use the "stopword" table. Add 1 field 'ignore also in unititle', and automatically compute the unititle from the title, by removing 'ignore also in unititle' words at the beginning of the complete title. I've no time to do it for instance.
Yeah, that sounds like it has possibilities. Now, who does have the time... Stephen -- Stephen Hedges Skemotah Solutions, USA www.skemotah.com -- shedges@skemotah.com