[Koha-devel] Technical issues with current workflow [ was: Release Manager 3.6]

[LAURENT Henri-Damien]

Le 12/05/2011 16:49, Scott Kushner a écrit :
> Maybe I'm not getting it here, BUT..have we forgotten how long it takes PROPRIETARY VENDORS to develop and debug their code? 
> 
> Maybe it's just me, but I'm amazed at how fast these releases are getting out there. (3.0, 3.2, 3.4...3.6...already?) It will be out soon, I'm sure. Also, and I know it's a hassle, but what's stopping you from writing your own code, integrating it into your instance of Koha, and using it while waiting for code to be tested and accepted into community? You're getting the benefits of your own written, or sponsored, code while having the peace of mind of getting the QA testing from the community. 
> 
> Again, maybe it's me, but am I missing something here? Have we become that impatient and spoiled that we want our pellet now, and let the process be damned? 
> 
> Scott Kushner
> Information Systems Librarian
> Middletown Public Library
> 
Scott,
The problem we have in BibLibre is that we developped things for 3.4 as
3.2 was not out. And that some other features have been pushed that
broke our code. Therefore we had to re engineer that code 4 times. And
that at some point, we are quite getting upset to see that still, that
code is not going to get seriously considered because it scares ppl away.
Having waited for over 2 years is quite a long time. And we would like
to get things validated. More and more libraries are working now with
that code... because they ask for Koha AND those features...
Kyle Hall may have the same issues, but his features are more external
modules, so he may be ok with that and handle that ok.
For us it has not been a problem not to see that in 3.2 .  It was
announced in RFCs for 3.4, but it was developped and delivered before
3.2 was out. And contains both new features and bug fixes... (Because,
believe it or not, as we developped, we discovered existing bugs...)
And we have our french bugtracker. But each and every bug we declared in
our bugtracker was not translated and updated in bugzilla because we
worked hard to get things done.
And the wonderful job done by the community (which I donot negate) is
making our lives a nightmare rebasing and redoing things.
So that we are wondering whether we could continue working and proposing
features, if we cannot make them in 2 years after they are delivered and
tested.
But it is not only a BibLibre problem. I think, many libraries are
working with a Koha fork (little or big, a fork is a fork, would it be
just an xslt change) that is fitting their needs. Our problem is that we
donot want this to happen for our libraries and our code, and we care.
And that we have vainly struggled to make them in in many ways. So that
at some point, financially, technically, and strategically some concerns
have to be discussed.
What will happen when we are more than 5 very active companies if some
big features are lost in the wild and re processed by another firm or
will have to be re engineered over and over again because there is no
planning, no developing group, no vision, no common goal.
It already happened with fine in days. Two companies did work on that
feature, we and xercode, and still, that feature is not in Koha. And it
is just an exemple.

Friendly
-- 
Henri-Damien LAURENT