On 2008-06-13, Tobias Brox <tobias@stud.cs.uit.no> wrote:
> [jim]
>> Most FOSS projects die within 5 years just because people have to feed their
>> families. Just look at the FOSS Graveyard (aka sourceforge.net) and that's
>> all the proof you need of the unsustainability of most FOSS projects.
>
> Quite many commercial software projects are canned as well. I've
> heard failure rates of 90%. Anyway, does the failure rate matter? I
> think the quality and quantity of the successful projects matters more.

True. I think it is more relevant to see how often peope are hooked into
a system that dies. I've been running public software for many many
years and I can't recall much. I liked qps (Qt process monitor) a lot
and I think is was better than most things I see around now, but that
won't be mission critical. My scanner (Epson) died when moving to the
libusb stuff. That is annoying, but not much different from the
situation around major Windows upgrades. Other than that, I can't recall
I ever had to discontinue using something because the developer stopped.

The figures indeed say little. Many so called open source projects are
just personal or student projects. They are easy enough to avoid if you
want. Would be interesting to see how often big Linux distributions drop
packages due to stopped maintenance/development.

Finally, in the OS community there is a lot more attention to open
format standards, so even if the package goes down it isn't a nightmare.

--- Jan