Hi Andrea and welcome :-)Hi Piotr, thank you!
I would suggest to start small, that is to create a minimally complexYes, it helps. Thank you very much. As I said, we'll keep you posted!
dictionary that you could later extend in various ways. That will give you
(a) a usable product, (b) sense of completion, (c) it will not bury you
under the issues of lexicographic description.
I'd say, select one of the existing databases with light grammatical
markup (try to restrict yourselves to part-of-speech, if you can), and go
ahead. I don't know what your background is, so I'm not sure how trivial
the following is going to sound: begin with a word/lemma list. You could
take it from a frequency list of some corpus of Italian/Sardinian (with
appropriate cosmetics), and then cut according to a frequency threshold
plus some grammatical conditions.
Or maybe you plan to take an existing dictionary, like ita-eng (I haven't
checked that one while writing this, so it's not an informed
recommendation), and handle "the other side", by more-or-less substituting
Sardinian for English. Then you could reverse it and clean/enrich the
reversed database.
HTH.
Good luck,Thanks again!
Piotr