[GNC] Please address broken QIF import

[Yann Salmon]

Hello,

Le 23/04/2024 à 23:48, David Carlson a écrit :
> Sir,
> 
> In my experience most financial institutions that offer QIF exports also 
> offer OFX format which will often go under a similar name.

Alas, a hundred times alas, not always. Not in France, at least. They 
should…

I was able, however, to use <https://github.com/georggrab/qif2ofx>. I 
had to reverse all transactions (which is easier to do in the QIF, btw) 
because my bank seems to directly produce a QIF for my point of view 
while GnuCash seems to expect a QIF with transactions from the bank's 
point of view, but that is another story.


> If you need the QIF format then you will probably be stuck either 
> waiting for the 5.7 windows release or reverting either to an earlier 5 
> series or maybe even 4 series if you have an older Linux based machine 
> as I do.

Even though the Gnucash team has been releasing new versions at a steady 
and formidable rythm of one every 3 to 4 months, that still makes 4 
months of non-working QIF import — downgrading is really not a 
straightforward route as software managers, for good reasons, make 
upgrading software easy and downgrading it harder (and one would have to 
get the idea of doing so).

While I admit I am not and have never been trained to be a software 
project manager, I did write some (extremely) modest pieces of software 
that are used by some other people, and the one thing I learned is : you 
do not break things, and if you do, you unbreak them as quickly as possible.

People can be patient with a desired functionnality not being present 
yet, or even a new functionnality being buggy from its start, but not 
when something that had been working, and that they are thus using, 
stops working. It disrupts their workflow, it may mean that something 
they had to do by some date, and that they were expecting to do with the 
software, suddenly cannot be done as planned anymore : all in all, it 
makes the software look unreliable — if faults appear today, and are not 
fixed, then more and bigger faults could appear tomorrow. And I think 
this is especially true for a software like GnuCash that is important to 
users because it is usually great at doing important things like 
accountancy.

I would be interested to know the bigger picture that I am obviously 
missing in the decision not to publish a .1 that fixes a regression 
(especially on a stable release) and allows Otto Normaluser and his 
Grandma to quickly get back on their feet by simpling doing what they 
know and have been educated to do : upgrading their software.

--
Cordialement,

Yann Salmon