Dear List Poster,
The strategy I have in mind for doing the migration is to convert all our bbliographic records to UNIMARC using the MARC::Record tool and then import it into koha.
0. Does this strategy look correct to you ?
Yes.
1. According to perldoc MARC::Record, it seems that the module can handle only USMARC and MicroLIF styles. Is one of these two okay for handling UniMarc and, if yes, which one ?
USMARC ($record->as_usmarc) produces valid UNIMARC records. This way you can populate an ISO2709 MARC file that can be uploaded into Koha with bullkmarcimport.pl script.
2. After the initial import has been done, we will certainly want to refine it by using data from our national library, sch as leaders for instance (?), or perhaps or initial import won't be complete and we will want to add other things we did not at first to keep the first import possible. So, is the incremental modification of records possible and how ? Wil we have to export the records to marc, then modify them and re-inject them, or is it recommended to do things another way ?
It can work this way. You can also use Koha API to modify biblio records directly. Take a look at Biblio.pm Koha module. You will read all biblio records. For each biblio record, you will query your (our) national library. You will merge both records and write the result into your Koha DB. Cheers, -- Frédéric DEMIANS http://www.tamil.fr/u/fdemians.html