In a note that was (accidentally) sent only to me:

I was hoping to write something in python to build marc records of our catalog, check books in and out and keep track of library card holders.

A quick note to say I agree with the others: don't do this.  You'll go mad.  Hacking on Evergreen would be more productive all around, I think. Or possibly there's some way you could use LibraryThing?

Bill [William Denton]

what will drive me mad?  managing marc records?  The smallness of our library means these records dont have to be very thurough.  Author, title, subject, etc.  I realise that each of those are not simple: corporate name or surname or family name,  theres several types of titles for some reason, and the subject should the authoritative one (ie, already recognized).

Checking in on some ILSs around, it seems that the interlibrary loan function is probably not well suited.  books arrive for one borrower only and do not sit in the catalog to be browsed.  Evergreen is already much more than I need, very complex, and written in a language I am not very familiar with.  hacking into it WOULD drive me mad.

Unless evergreen's ILL function does work the way I need... (anyone know?)

Hopefully someone has done something with pymarc to give me a head start on managing MARC records.  Mostly, I will just be downloading records from the library of congress and having a script check them for the absolutely necessary fields, prompt for missing data and then add in the local information (location, call number, etc). I am already 45% done with this and am cleaning it up (making it object oriented) so it can be adapted easily and reused.

As far as borrower and lender admin, that is not so difficult.