On 26 juin, 18:17, Rom1 <[email protected]> wrote: > On Tue, Jun 26, 2012 at 5:14 PM, sampac <*[email protected]> wrote: > > > I did something here:https://github.com/sampaccoud/tryton-bank > > > So if we want to allow storing different numbers for a same bank > > account I can refactor to make a separate object for the bank number. > > Then we can add several account numbers to a bank account. > > > But I don't think it is necessary because a given bank account only > > works with a given number. RIB and IBAN are just the same thing with a > > different checksum... > > > The way I did the account number is very generic and we can write as > > many methods to validate and split the account number depending on the > > country case... > > > Looking forward to your comments / pull requests. > > Very nice job, much better than what I did.
Thanks, I built on yours ;-) > Just a few comments : > BankAccount > Who is "party" ? The owner or the bank agency (because I don't > understand the domain on the address). For me, you have the party > (the owner), the bank agency (party.bank), the address (with domain on > the bank agency). > Ok let me refactor and propose something like this. > The RIB key could be calculated with stdnum.iso7064.mod_97_10 A bit overkill ? It's just 15 lines of code so maybe we keep it like this ? > Bank > Why is it an inherit of address and not of party? > The bank code shouldn't it be stored in parent of the bank? > I think there is not one BIC per agency, but you have one BIC for > several agency within the same region. bank code: I will propose a parent on bank which will be a inherit of party (same scheme as "company.company" in fact). Bank code will be available with required = false. BIC : the field is not marked as unique so should be ok. -- -- [email protected] mailing list
