Rachel Hamilton-Williams a écrit :
When you set up your Koha you can choose to have it running in "marc mode" or not - if you choose "not" then I believe it doesn't "show" the MARC and pretty much runs off the Koha db.
I think that you can search either the koha db (smaller and faster) or the MARC db (bigger and slower). The koha db is I think a subset of the MARC db - and is basically automatically "joined" to the marc db, so that the two are kept in sync. Additional info about an item is added to the Koha part - about it's life since entering the library, which MARC isn't interested in.
There is a fair amount of info in a MARC record which lots of libraries have no real use for, and give volunteer cataloguers the willies. So we have the most important fields (the must haves :-) in the Koha db. So the question here - is language a "must have". UNIDO thought it was, and it seems a few more might as well - basically it probably depends on where your library is, so it's worth having it maybe in the Koha db.
You get a 19/20 for this answer Rach, (should be put in a FAQ somewhere). (the point you loose is for saying "if you choose "not" then I believe it doesn't "show" the MARC and pretty much runs off the Koha db" : even if you set marc=off, searches are done on MARC part : the MARC search API is correctly optimized, that's not the case of the previous non-MARC search API (that was really slow => unindexed searches sometimes. will be improved in 2.4, with inverted list (marc_words table)) -- Paul POULAIN Consultant indépendant en logiciels libres responsable francophone de koha (SIGB libre http://www.koha-fr.org)