Re: [tlug] Not always about scratching an itch

["Stephen J. Turnbull" <stephen@example.com>]

Darren Cook writes:

 >>>  2. The examples on the project website cover just about every possible
 >>> use case, allowing you to quickly confirm it will work in each device
 >>> you have to support. Contrast this to more complex open source libraries
 >>> where evaluating if something does what you need can take longer than
 >>> the rest of the coding on the project.

 >> I don't understand why this comparison is interesting.

 > Interesting for the observation you gave: open source projects that
 > focus on one thing, the narrower the better, have a higher chance of
 > being successful. Those that try to implement more general solutions
 > tend to fail, and because success becomes more tied to the management of
 > the development process than the skills of the lead developer.

My point is that that theory fails in practice.  It's not a question
of what you *try* to implement.  What you *need* to implement is what
determines the interactions.  If a feature involves interactions,
you're in "management is not a luxury" territory.  This is why "herd
of cats with no catherd" projects sooner rather than later produce
such sucky software, and why people used to HOCWiNC[1] projects think
that the Linux kernel, the BSDs, Python, etc, etc are "hard to join".

An example is GNU Emacs, which has a nearly HOCWiNC process.  But
there's no question in anybody's mind that XEmacs development is
technically easier to join (despite the implementation arcana that
sprang full-blown from the mind of our intermittently resident Evil
Genius Ben Wing -- or maybe that should be "intermittently Evil
resident genius" ;-).  And the worst successful HOCWiNC project I know
of is Org-mode, which produced the world's ugliest markup language. :-P

So my conclusion is "if you want to succeed as an open source project
choose undergrad student homework assignments to implement" ... and
sometimes you may get lucky or have the good taste to come up with a
"killer nano-app". :-)

Footnotes: 
[1]  The fact that this acronym sounds like a wet cough when
pronounced is a happy accident.