Nominating things for the 3.2 roadmap
vladd at bugzilla.org
Sat Oct 7 18:18:43 UTC 2006
I wanted to share the Linux "roadmap" thing, maybe it's useful for
Bugzilla (or not).
According to Linus, Linux is "evolution", not "inteligent design". The
way Linux expandes is governed by its contributors. They have a 2-week
window in which they merge all the features that got coded during the
previous freeze, and after that it's stabilization, and release. And
after that it starts all over again.
There's no roadmap in the Linux world from this point of view.
Well, ok, there are lots of roadmaps in the Linux world. Each company
that depends on a specific device driver for the Linux OS has a roadmap
regarding the schedule of getting the driver into the kernel, and stuff
like that. But there is no "central", "approved" roadmap, because the
only way to put a thing in the Linux roadmap is to work hard enough and
be responsive enough in order to "carry" the patch in the Linus' tree.
I do welcome any kind of coordonation in a project, but I think
roadmaps, if mis-managed, can have a detrimental effect.
I've welcomed the Bugzilla 3.0 roadmap when it appeared, but I think
we're starting to push roadmap management too far, with possible
detriments to our developers and engineers. We're way too early in 3.2
lifecycle in order to "nominate" things. Developers that want to
nominate something should start working on that feature, or getting it
reviewed/feedbacked. But even that is too early (bitrotten is a fact).
If we want to have things on the radar, we can use a lot of mechanisms
already available (priority, target milestones, blocking flags)...
I think the Linux model has a lot of impact in motivating developers to
get stuff done. I, like Linus, prefer "evolution" to "inteligent
design", and too much roadmap management can be detrimental.
It's just my personal opinion, and I probably won't have the time to
reply to the replies, so consider it a useful insight :). Bugzilla has
become a 2-man game (LpSolit and mkanat), with Marc, RSZ, Colin and
others digging in occasionally. Compared to the state of things when I
was around, the top contributors are a lot more active, but the team is
somewhat smaller. I think it would be cool if we could manage to keep
the activity of top-contributors, while at the same time encourage new
devels to kick in. I do think roadmap over-management can hurt the new guys.
Max Kanat-Alexander wrote:
> Hey hey. So I was just mentioning to LpSolit, we can nominate things to
> show up on the 3.2 roadmap by putting "[roadmap: 3.2]" in their status
> That doesn't mean they *will* be on the roadmap. It just means that in
> two months or so when we write the roadmap, we'll have a list to go
> Don't discuss it here, just nominate things and we'll discuss them in a
> month or two.
More information about the developers