Buyer Beware: A Major Change in NFS is about to happen

Kevin Kofler kevin.kofler at chello.at
Thu Oct 1 01:37:42 UTC 2009


Jesse Keating wrote:
> This isn't a post-facto justification.  The only "one-off" for F12 was
> the removal of the milestone previously known as alpha.

Making the renaming a one-time-only change as I'm proposing would be "post 
facto".

> The rest of the milestone adjustment proposal came out of the Fedora
> Activity day, had lots of time to be communicated, discussed, and voted on
> by the community at large, FESCo specifically.

I don't remember FESCo ever voting on that issue and I can't find it in the 
meeting summaries either (and I also checked the ones from before I joined, 
back to April). AFAIK, this was just posted to the fedora-devel-list for 
feedback, got almost none, which was taken by you as "everyone is fine with 
it" (whereas I think most people probably just ignored it as "yet another 
wacky proposal which is never going to get implemented sent to fedora-devel-
list") and a few days later it was a rel-eng decision. (I remember having 
been really surprised by this having been implemented ("Huh, there's a F12 
Alpha now?"), given that there was no consensus at all on the mailing list.)

I don't doubt there was some in-person discussion at the FAD, but that will 
never be as inclusive as a mailing list discussion (which never happened 
because a huge list of "brainstorming results" (which turned out to actually 
mean "almost decided items", which also wasn't clear from the wording or you 
might have gotten more objections) was dumped onto the mailing list in one 
e-mail). I think a small circle of people flying to some location to make 
lots of decisions in person in one day is really bad for transparency in an 
international project like ours. Ideas really need to be sent one at a time 
to the public mailing lists.

And there's an additional important point: we have additional evidence now 
that the renames were harmful! So, independently from how the decision was 
originally achieved, we should revisit it now based on the new evidence. 
Back when the renames were proposed, I didn't quite see the point, but I 
didn't think it'd be a big issue either. But as time has progressed, I have 
seen many developers come to FESCo with things like "oh, I was supposed to 
have this feature ready for Alpha? I thought it was Beta as always!". These 
renames turned out to have produced lots of lateness in the feature process, 
and Fedora development as a whole. So I'm now proposing to solve this 
problem by restoring the names people are used to. I think having our 
developers deliver on time is much more important than using pedantically 
"correct" names which lost most of their meaning anyway ("beta" can be 
everything from unusable pre-alpha code to Google's "betas" these days) in 
end-user communication. Those users are going to be the ones hurt the most 
by buggy, incomplete or entirely missing features, so having the milestone 
names make sense for developers is also what's most important for THEM.

> I'm really not in favor of changing it again.  A group of developers,
> release engineers, and QA folks brainstormed on fixing the milestone
> issues and what we have now is the product of that effort.  If you
> really think it wasn't for the eventual better, float your own proposal
> and get approval from developers, release engineers, QA folks, and
> eventually FESCo.

My proposal is plain and simple: go back to the 3 milestones and their 
naming from F11. It worked and developers all knew what the milestones 
corresponded to. The change was completely unnecessary and just caused 
confusion.

        Kevin Kofler




More information about the fedora-devel-list mailing list