Testing reports

[Andrew Sullivan]

On Thu, Apr 19, 2012 at 11:33:00AM +0100, Colin Scott wrote:
> 
> > The plain fact about free software with volunteer maintainers is 
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^
 
> Yet somehow there are any number of "free software" open source projects out there - some of them accounting systems! - which do seem able to plan the intended direction of the product and produce road-maps for future development, often even with projected dates for the next few releases.

I think you missed some important distinctions here.

First, some of the free software (I'm old. That's what we used to call
it when I grew up, before Eric Raymond decided he knew better than
everyone else, and despite your condescending use of scare quotes I'm
going to continue using that term) accounting packages are not really
volunteer projects. They're businesses that release their packages
under Open Source Initiative-conforming licenses. Naturally, some of
the programmers working on those projects have motivations that are
not merely, "Gee, this is interesting, I'll work on it," or, "That bug
is driving me mad."  Note that it can be hard to be an external
developer and to contribute code to that kind of project, which means
that if you want a feature, you often have to pay for it.  This is
not a criticism of those projects; it's instead a distinction that
makes a difference for this discussion, so we need to attend to it.

Second, some of the accounting systems I evaluated when picking
gnucash had detailed roadmaps and so on; but also a spectacular record
of failure in meeting any of those milestones.  Announced plans that
don't actually happen are worse than no plan at all.  

Finally, it's worth noting that some of the larger software projects
with the road-maps and so on have that _because someone is
volunteering to do it_. PostgreSQL, for instance, has someone who
maintains the TODO file, and who goes through mailing list responses
and notes when people are working on something. He spends a lot of
time on each release working out what features are really ready,
negotiating deadlines with the developers, and so on. And he can do
this because the hacker community there is large and extremely active.
It seems to me that someone who apparently has a lot of time on his
hands to post long, detailed criticisms of project management might
usefully contribute some of that time to improving that project
management.  Of course, part of managing software projects involves
not alienating all the people who are writing the code.

This will have to be my last remark on this subject; I'm afraid I've
spent rather more time on it than I expected.

Regards,

A

-- 
Andrew Sullivan
ajs at anvilwalrusden.com