[haiku] Re: Package Manager:
- From: Sean Collins <smc.collins@xxxxxxxxxxx>
- To: haiku@xxxxxxxxxxxxx
- Date: Sun, 27 May 2012 14:31:52 +0000
Charlie Clark wrote:
Hi,
I really don't like the way this conversation going.
Thats a choice your making
The conflation of "you" with the "user community" is, at best,
unfortunate. Similarly the recent parallel drawn between donor and
investor. It is important not to raise expectations artificially.
Donations to the project have been fantastic and have enabled some
developers to do small amounts of work but at well below market rates.
But the *vast* majority of code has been done by people in their spare
time.
I am not conflating anything, maybe get out in the user community, read
some of the commentary from users and third party devs.
I'm not a little proud to say that the BeGeistert coding sprints,
where developers pay for the privilege of devoting their free time to
program have been very important over the last few years. This has, at
times, led developers to devote two whole weeks a year to Haiku!
I appreciate your efforts as do many others. But lets be honest, BG has
been scaled back this year from years previous. the HSA has indicated
either a lack of funding or interest for the reduction in BG events
It is one of principles of open source. This, and the relationship of
money and open source, are frequent topics around open source. I
recommend Stormy Peters of Mozilla presention: "Would you do it again
for free?"
http://stormyscorner.com/would-you-do-it-again-for-free
I am asking the developers what they would need from the community to
get this OS to R1, and the conversation to that end will invariably
include monetary compensation, unless we all start living in communist
social systems. People need to eat.
What are the grounds for this assumption? Ingo's responses on this
mailing list?
Well, what other assumption is there to be made ? there is only one
developer working on PM
The theory of the need for software testing is well understood. As are
the limits of such testing and the time that can be devoted to it. It
is generally considered to be impossible to develop bug-free software
which is probably why software manufacturers don't get hit under
unlimited liability claims *all* the time.
It is impossible to develop bug free anything.
Haiku's unit test coverage is currently quite low; this is certainly
not helped by the amount of boilerplate that has to go into writing
and running them. Integration tests with real hardware are even lower.
It's actually close to a miracle that any of it works. Haiku has no
formal quality assurance, this is probably the one area where a
commercial approach, which would in theory provide resources for this,
would be an advantage. Though anyone with experience of the software
industry is replete with anecdotes of poor approaches to testing by
management and investors.
Charlie
The question is what kinds, types and amounts of resources will it take
to bring Haiku to R1 ?
Sean
Other related posts:
