burtonator <[EMAIL PROTECTED]> wrote:
> I don't think that the Open Source way of releasing software is any
better or
> worse than they way large (sometimes proprietary) companies release
software.
Hi Kevin. I didn't mean to say one was better or worse than the other.
What I meant was, procedure that has been developed for traditional
development within companies doesn't necessarily work well for open source
communities. I think the primary issue is control. If I'm developing
software within a group in a software company, someone has pretty absolute
control over the developers, and usually has strict delivarable contract to
customers. In open source, you simply can't control when things are going
to be delivered, because for many developers the open source coding is only
a sideline... they have day jobs. The same goes for customers who are
testing the software. Also, the software for open source often needs to be
far more general than for proprietary solutions... and thus can have a much
broader testing landscape. I don't think this is only an issue caused by
open source, but also by general standards that allow the software to be
used in a great variety of situations.
> IE.. Do allow feedback from your users as to a potential release
I take it this means you think we should go through a proper final beta
period? I tend to agree, though it may slow down the process yet more.
> and also run
> regression tests and QA. This last step is often neglected by smaller
Open
> Source projects due to lack of resources.
I think the Xalan project has been exemplary in building a testing
structure and running these tests before release. However, we've not been
as good about building proper regression tests for each bug fixed to date.
This is what the "Bugzilla test automation available!" thread is all
about... so hopefully this should get much better real soon.
-scott
