A Dilluns, 29 d'abril de 2013 11:13:07, Marc Rechté va escriure:
> Sorry to come back on the subject, but I think this "company"
> denormalisation should be applied to even more tables. It would also
> benefit to the KIS principle and would avoid mistakes of forgetting to
> include not obvious where clauses.
> 
> I checked OpenERP database: they have included "company" field in many
> tables.
> 
> In a real case application, either company field will be simply not used
> (viz one company / database) or if used, will give headache to the
> programmer.
> 
> Accounting module being the core module, I think a lot of attention should
> be given to it, to make it reliable and comfortable to use. IMO it is not
> currently the case.

I haven't seen any arguments that make me think that denormalization is a good 
thing. Why do you think it will give headaches to the programmer? Why does 
denormalization avoid those headaches? Why do you think there's not enough 
care with accounting?

I'd say accounting in Tryton is better than in OpenERP:

- Tryton manages taxes with a one2many from move lines instead a many2one. 
This avoids stupid move lines with debit == credit == 0.0 when a move line has 
more than one tax.
- Dates and periods in lines and moves are consistent. They are normalized and 
avoids things hard to justify to users such as lines with dates different from 
the move (this happens in OpenERP).
- The date of the move it is guaranteed to be in the period of the move. We 
had to develop a module in OpenERP for that because otherwise it makes things 
inconsistent.
- Partial reconciliation in OpenERP is broken: it will show you the wrong 
pending amount in invoice lines.

And one more thing:

OpenERP cannot handle searches like 'period.fiscalyear.company' in a domain as 
it can only manage one step in the dot notation. That would be at most 
"period.company". So it is possible that what you argue as a feature of 
OpenERP it is probably a limitation of the framework.


-- 
Albert Cervera i Areny
http://www.NaN-tic.com
Tel: +34 93 553 18 03

http://twitter.com/albertnan 
http://www.nan-tic.com/blog

Reply via email to