On 20 Jun 2007, at 06:04, Chris Travers wrote:
... I understand what you are saying, but I still disagree. Instead I think we are talking two different things. 1) Invoices need to be entirely self-contained. They need to store all information required to recreate and track them. We may want to add a lot more metadata (even if it looks like duplication) to the invoice to make that possible. That is the best way of tracking customer state information at time of invoice.
Hear, hear!!I find the way relational data is currently applied to old invoices to be *ledger*'s greatest shortcoming. Company changes address, I look up an old invoice & print a duplicate and its address is wrong.
I know others have posted suggesting ways to keep previous addresses associated with old invoices, but IMO invoices aren't suited to a relational structure. Invoices have a unique status as a legal document - once the invoice is published that address BELONGS to the invoice. Even if another invoice is printed to the same customer & address the same data belongs separately to both invoices and is not shared between them,
Having invoice records contain all the data shown on the invoice itself is the best way to ensure the invoice is set in stone, IMO.
Stroller.