Article on wine development strategy

[Dan Kegel]

On Sat, Apr 18, 2009 at 1:23 AM, Kai Blin <kai.blin at gmail.com> wrote:
>> The one that worked out best was to pick some random
>> user who's almost happy, fix the last few bugs that
>> are keeping his apps from working, and then once
>> he's happy, move on to the next such user.
>
> The problem seems to be identifying these people. The model assumes that you
> can tell which piece of software almost works, and that you know the almost
> happy users. In reality, you only seem to hear from the pretty unhappy users
> and the occasional really happy user.

I suspect that one can come pretty close by fixing one bug in
each application that a particular complaining user uses.
If that makes the app work (or if it seems clear that the
app is really, really close to gold), bingo, you've found one.  If not,
you can move on without having wasted too much time.

Photoshop CS2/CS3 was kind of a special case.  It was the #1
requested application, it was a convenient rallying point, and
I think having it work as well as it does now is a significant point
in Wine's favor.   Plus a number of the fixes benefitted other apps
(especially Adobe apps).  There were other big thrusts during the
same time period (MSI and gdiplus, for example) that were more
broadly helpful.   Plus all the while I was looking for individual
users with complaints, triaging their bugs, and trying to make them
happy.   So we were really following three strategies at once:
1) fix the most popular app
2) fix key components that many apps need
3) find individual users and make them happy
I think it was a good mix.
- Dan