Professionalism in the development version (was Re: Artwork shuffle?)

Ivan Krstic krstic at fas.harvard.edu
Sun Apr 16 23:15:17 BST 2006


Matthew East wrote:
> Everyone using it knows there is some
> element of risk in using it, but if we are going to be _this_ extreme
> about discouraging people from using it, then we are not going to get
> many testers, and the OS will suck more.

Extreme? I spoke about the operating system eating chocolate, for crying
out loud. My tongue was firmly in cheek, but my basic point stands;
while it'd be very bad for any OS release to do nasty things to one's
system, a strong expectation that it won't cannot be bestowed on
anything but a stable version. The same generally holds for other things
perceived as "release professionalism".

> The testing community is a fundamental part of the growth and quality
> assurance of the project, and some attention has to be paid to not
> alienating that community.

I've found that there are broadly two types of people who test
development releases. The first are hacker types, who are often on the
bleeding edge because their own development work requires it. These guys
understand the risks, conscientiously report problems, and perhaps most
importantly, have the know-how to dig themselves out of most any hole in
which living on the edge might bury them.

The second is the impatient user type, who just can't stand to wait
those 6 long months before the next software version arrives. The
thought of running anything older than 2 point releases away from the
trunk fills him with abject horror, and on account of this, he takes the
plunge towards a development version once it stabilizes enough that he
doesn't expect real problems with it. He projects most of the
expectations of the stable version onto the development one, and is
commonly irritated when something doesn't work properly.

No one wants to alienate the testing community -- they provide an
invaluable service to software development. But the second group has
unrealistic, mistaken expectations, and it's a waste of resources to
cater to them, alienation notwithstanding.

I don't want to continue this thread; it's adding noise to the list. For
the second time, I'm asking for followups to be directed to sounder.

-- 
Ivan Krstic <krstic at fas.harvard.edu> | GPG: 0x147C722D



More information about the ubuntu-devel mailing list