I haven't started actually cataloging yet, other than assembling a batch of customized MARC records for a subset of our biblios, starting with a pretty extensive video collection. I intend to take as much advantage as possible of well-enforced standard Authority records, but am starting to appreciate how much extra work will be involved if I have to manually look up one-at-a-time all the relevant subjects and people via the public interfaces offered on the web - genre/forms don't seem such a problem as I'm planning on only actually using a small subset of them. So I'm exploring the possibility of importing the full set of OCLC/LoC authority records into my Koha database before I start importing my biblio records. The subjects aren't so bad, apparently under 300,000, but I believe the names side of things are several million. Will doing this greatly slow down my Koha? Our intended production server is 2GB Core Duo 2Ghz, and we're not likely to have more than a dozen concurrent users in the early stages - both OPAC and staff combined. Disk space isn't a problem - I've got a spare 15+ TB available on our filer's LVM 8-). Also my understanding is that the link_bibs script does actually work, as long as the authority-governed fields don't have existing/conflicting $9 fields, just the exactly-matching name or subject strings - and updated zebra indexes - and of course properly configured MARC frameworks. I've seen some references to using the regular biblio import script for authority records (accompanied by statements that the authority import script is deprecated), but I've seen just as many reports that people have had trouble getting that to work. It also seems that nothing will automatically handle the bib-to-authority linking at import time, that it *has* to be done via the link_bibs script. Confirmation of the above would be greatly appreciated, as would (even more so) any explanations on where I've gone off track.