Chris Bainbridge posted <[EMAIL PROTECTED]>, excerpted below, on Thu, 23 Mar 2006 18:55:15 +0000:
>> No. It indicates nothing except that 58% of the 80 people who filled >> out the poll are "not really happy with the way the portage tree is >> handled" which by my counts, isn't even a drop in the bucket of our >> number of users, making the statistic completely worthless. > > True. Nevertheless, it is the only statistic I have seen regarding > users thoughts on this subject. Of course a larger sample size would > be preferable. There should be more feedback from users - bug voting > on bugzilla would be a start, why has it never been enabled? Two comments, one likely fairly obvious: Obvious: Both Gentoo users overall and the poll would tend to be self-selecting, Gentoo users toward those approving Gentoo's handling of most things -- or they'd be elsewhere, as there /are/ other choices, the poll toward that segment of Gentoo users not so happy -- since it's widely known that unhappy folks tend to be more vocal and more likely to be motivated to find and vote in such a poll than folks happy with things as they are. Even having the URL, for instance, I'm reasonably satisfied with the handling of the tree as it is, and the forums are out of the way, so it's unlikely I'll find it worth my time to go over there and vote. The other comment, answering your bug voting question: >From past threads where bug voting has been discussed, it appears a fairly vocal segment of developers oppose it. The reason given is that users don't necessarily have any sense of global priority (a data-loss situation on sparc, for instance, with maybe two users affected, vs an enhancement with a hundred votes that'll take weeks to implement and test), ease of solution, or knowledge of pending security vulns currently embargoed, for example, and how they affect developer's priorities. Neither are they likely to know how many other packages a developer may be responsible for, or what other factors an after-all volunteer developer may be balancing in their real-life. It'll be all too common, they are afraid, for users to ask why that bug with 100 votes seems to be being ignored, when all they see is 1 vote from the sparc guy on the bug getting all the action. Other developers in favor of bug voting have countered that a developer can ignore votes if they want, that its only an indicator of interest that need not be used if it doesn't fit the need. The counter to that is that the CC lists are already a decent indicator of interest, for those who want it, without implying the same obligation of the dev to prioritize bugs with the most votes. In any case, no consensus in favor large enough to turn on the feature has ever seemed to emerge, with a number of devs strongly and vocally opposed. I believe that's a fair summation of the arguments. My personal opinion, for whatever it's worth as a user on the dev list, is that the CC point is a valid one, the CC list should be a pretty decent measure of interest -- I know it has certainly proven so on some of the bugs I've CCed myself to, as I've seen others pile on too. Take a look at the CC list on this (resolved) one on AMD64 performance patches to glibc, for instance: http://bugs.gentoo.org/show_bug.cgi?id=100289 . How could voting make the interest any more evident than all those CCs, and the CCs actually continue to be useful after the "vote" has been registered, as well. Bug voting? I'd argue we already have bug voting! -- Duncan - List replies preferred. No HTML msgs. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman in http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html -- gentoo-dev@gentoo.org mailing list
