I can't tell whether this is a failing of open source software users, people who file Mozilla bugs, or American society in general, but I'm running into more and more people who think that everyone owes them someting:
Scenario 1: Someone files a feature request on an area of code that has so many issues that it needs a total rewrite or close to it instead of new features (helper apps come to mind, as do some parts of layout and anything dealing with view-source). The bug gets five votes from people who think this is the best feature idea since the <img> tag. This is interpreted as a mandate from "The People" that someone should drop everything and fix that bug right now.
Scenario 2: Someone files a bug. 36 hours later they complain about the bug not having gotten looked at. Nevermind that during those 36 hours there were 250 bugs reported (most of them duplicates, just like the bug in question). Clearly the bug should be looked at yesterday.
Scenario 3: Someone files yet another duplicate of a "bug" that has been marked INVALID. This is seen as indication that the bug should obviously be reopened and fixed right this instant.
These examples demonstrate what is, in my mind, the biggest danger to the Mozilla project—the explosion of abusive and bratty bug reporters. This is the major contributor to my current attempts to reduce the amount of time I spend on Mozilla. It's just not worth my time to be threatened and cursed at every time I read my mail.
I haven't had to deal with such continuous displays of childishness since my brothers passed the age of 2 and it became possible to actually explain something to them and for them to understand that they are not the only people in the world who matter.
Posted by bzbarsky at February 10, 2003 5:01 PMThere is no solution for this because
a) you can create x Accounts in bugzilla
(no problem if you loose one)
b) you can't close bugs for new comments.
And people will always try to change a decision with stupid comments if they see that this sometimes work. (see the grippy bug)
You must learn to ignore such ppl and we need you as great developer in this project.
Please don't let they destroy this project with their stupid comments.
Yeah, well... The fact that you can't close bugs to comments is a bug in bugzilla as far as I'm concerned. And the grippy bug is an example of spineless UI "owners" combined with pressure from Netscape UI design "experts" overriding decisions made by the people in charge of the Mozilla UI.
I'm not even talking about UI here. The UI situation is fucked beyond belief. I'd settle for people staying the hell out of the rendering engine and other back end code if they don't know what they're talking about....
Posted by: Boris on February 11, 2003 2:34 AMI've given up on the UI situation in mozilla. I have no opinion on the presence of the grippies themselves (although a grippy for the menu bar is 'interesting' - even ns4unix doesn't have that). I just wish that someone could make up their mind. Of course, then you have to define who that someone is... Maybe we could ask botbot for a UUID - odd ones mean the feature gets in, even ones mean it doesn't. For contraversial cases, we can do a best of three showdown with mozbot, bobot, and ssdbot.
And don't get me started on the GRE, mozilla's answer to so versioning. Just because windows can't have multiple versions of a shared lib installed... Like we have so many embeddors, all of whom will be using a build built on the exact same day (see bug 180383 - how do I linkify stuff in a comment?) hopefully pulled on that day, with no additional patches, and from the same branch. Oh, and built with the same compiler + flags.
Posted by: Bradley on February 11, 2003 6:36 AMBoris, absolutely, the amount of people who actually post useful technical information in bugzilla is minimal (5% maybe?). Maybe a solution would be to have a "cancomment" permissons thingy so random people can't comment on bugs (and/or slashdot style comment rating?). Its a shame people can't just stop themselves from posting crap...
Bradley, absolutely also, someone must be responsible and have the conviction to say "yes this is going to go in" or "no it won't" instead pissing about for years deciding what should happen. But then i thought that's what drivers were for....
Haven't thought about these much, just what came to mind...
Posted by: spark on February 11, 2003 1:23 PMBoris, absolutely, the amount of people who actually post useful technical information in bugzilla is minimal (5% maybe?). Maybe a solution would be to have a "cancomment" permissons thingy so random people can't comment on bugs (and/or slashdot style comment rating?). Its a shame people can't just stop themselves from posting crap...
Bradley, absolutely also, someone must be responsible and have the conviction to say "yes this is going to go in" or "no it won't" instead pissing about for years deciding what should happen. But then i thought that's what drivers were for....
Haven't thought about these much, just what came to mind...
Posted by: spark on February 11, 2003 1:24 PMBoris, absolutely, the amount of people who actually post useful technical information in bugzilla is minimal (5% maybe?). Maybe a solution would be to have a "cancomment" permissons thingy so random people can't comment on bugs (and/or slashdot style comment rating?). Its a shame people can't just stop themselves from posting crap...
Bradley, absolutely also, someone must be responsible and have the conviction to say "yes this is going to go in" or "no it won't" instead pissing about for years deciding what should happen. But then i thought that's what drivers were for....
Haven't thought about these much, just what came to mind...
Posted by: spark on February 11, 2003 1:24 PMThe solution to this situation is obvious: have two Bugzillas.
The first one, located at bugzilla.mozilla.org, would be the pretend one. Real developers would pay no attention to it and the great unwashed could post comments saying, "We should have a pref for this" and, "Please *fix* this, it's IMPORTANT" to their hearts' content. To keep the ruse going, every now and again a script would post comments that appear to be from real developers. This script would be smart enough to make the comments seem authentic, so you'd have David Baron saying, "According to section 5.12.3 of the CSS 2 spec..." or Samir Gehani saying, "Nav triage team: nsbeta1+/adt2". Occasionally timeless would appear to pop in to set everybody straight.
Meanwhile, the real Bugzilla, located deep within Cheyenne Mountain, would be where all the real work is done. To keep new blood flowing into the Mozilla project, people who post consistently good comments to the fake Bugzilla would be given access to the real one. Who would be crazy enough to want to sift through all those mindless comments to find the occasional gems? Why, the same crazy people who triage the UNCONFIRMED bug lists every day. As marking duplicates and the like wouldn't really matter in the fake Bugzilla, the R.K.Aas of the world would be freed up to participate in this important task.
This would give us the best of both worlds and keep everybody happy. Whoops, I'm starting to sound like a bad Bugzilla commenter now...
Posted by: Alex Bishop on February 11, 2003 1:50 PMBradley, do we use the GRE on Linux too in preference to .so versioning?
Alex, I'm not sure we want to go _that_ far. We get a lot of good bugs reported by people who have never reported anything before.
What we _do_ need are ways to mark comments irrelevant to a bug (and have the irrelevant comments not be shown by default). That's the best solution I've been able to come up with so far... :(
Posted by: Boris on February 11, 2003 7:19 PMas long as the person who makes the comment never sees that it's irrelevant, it would probably work :)
Posted by: db48x on February 11, 2003 10:25 PMbz: Yes. We do. And we don't look in the ld lib path (either standard or from the env variable)
I asked several people about this, and haven't got an answer I'm happy with. See some comments in bug 190907.
I'm perfectly happy with having a set of standard libraries. Its trivial to do that though - we already split up nspr and nss in the rpms, and defining a 'core set' of stuff which is in another rpm is really not that much work.
Posted by: Bradley on February 12, 2003 3:19 AMBoris, it was a joke. ;-)
"What we _do_ need are ways to mark comments irrelevant to a bug (and have the irrelevant comments not be shown by default). That's the best solution I've been able to come up with so far... :("
I like that idea. Could be kind of like Slashdot's moderation system, except without the goatse.cx links and "IN SOVIET RUSSIA" jokes.
Posted by: Alex Bishop on February 12, 2003 4:16 AMBoris, those "requirements" arent that seldom in tables land. (1) The first time I have seen them I got angry, (2) the second time I asked whether I can reassign the bug to the requestee. Oh man that gives usually silence. (3)Now I am at a weighted approach: the potential use of a comment is weighted by the number of patches, testcases or correct judgements in bugs by the author. For you that weighted approach would mean that you need to *read* only dbarons and rocs comments. They don't post silly requirements because they know already the danger that the bug can be reassigned to them (see (2)). I remember the times when at mozilla items like http://www.cafeshops.com/cp/prod.aspx?p=mozilla1.641337
have been popular to silence the lamers.
I don't like Slashdot's point system. Here's why. Back in the early days, people could say something unique and interesting. Nowadays, they still can, but it will never get points. Only posts that correspond to the Slashdot groupthink mentality, or are technically authoritative, get high scores. The Mozilla Project needs not only technically authoritative comments, but also a steady supply of fresh and unique ideas.
What I think you're addressing here, Boris, is a Bugzilla problem. The problem originates in that a lot of important information is contained in comments. Unfortunately, in many bugs, comments also contain a lot of useless information. It can be a waste of time when you're separating the good from the bad. Bugzilla does not yet offer a technical solution to do this. Instead, it just dumps all comments sequentially, like a big text file.
Maybe some kind of XUL or DHTML interface could be built in to bug pages. For example, if you click one of the buttons, all of the "Bug xxxxxx has been marked as a duplicate..." comments would be hidden. Click another button, and all the posts of a killfiled commenter would disappear. Click a third button and all comments would be displayed, like a raw view.
Posted by: Andrew Hagen on February 12, 2003 2:42 PMBradley, can we shoot someone for that crap?
Alex, I know it was a joke. But I was serious.
Andrew, I am not suggesting that _everyone_ have moderator access. That is the problem with slashdot. I'm also not advocating wide-ranging use of rating comments. But the owner of a bug should be able to mark all the comments irrelevant to the actual fix so. It's very rare that a bug really needs a "creative idea"; more often what's needed is just time to fix it (and getting rid of all the "I want this fixed too" comments makes the bug easier to deal with when such time is found).
Bernd, I try to do that. I really do....
Posted by: Boris on February 12, 2003 3:57 PMAnother possible improvement might be to automatically mark all serverity:enhancement bugs as helpwanted and assigned to nobody@mozilla.org. Then anyone actually implementing these features could assign the bug to themselves. It would also create the more accurate impression that suggesting a feature does not mean that a team of devs will automatically jump on it and start implementing, and maybe encourage more people to try coding something without fear that their work will be useless. A more extreme example of the same idea would be to move all RFEs to a seperate database.
The heart of the problem seems to be that people confuse 'Open Source' with 'developers working directly for my whims', which is something of a spectacular misinterpretation of 'providing the code so users are able to impement changes they require'.
Speaking from the point of view of someone who has posted his fair share of inane/inaccurate bugzilla comments, (but hopefully no 'me too' type comments), the scoring system idea sounds. It would have the additional advantage of improving the quality of the comments that people made, since there would be no incentive to post a whiny comment that would just be ignored. It would also provide a way of determining the qualities of a good comment.
Posted by: James Graham on February 12, 2003 5:24 PMSomewhat tangentially: I use the "votes" feature in Bugzilla as a quiet CC: -- I don't want to send everyone on the CC: list a notice that I want to be updated on the bug.
bz: Talk to dougt :)
Posted by: Bradley on February 13, 2003 1:17 AMI'll stick with Phoenix, where most of the whining occurs on the forum. Of course then I have to search Bugzilla and tell them to blame Mozilla, but at least it helps keep down some of the tripe.
I guess we could all use more triagers.
Come over to the wonderful world of Phoenix!
Posted by: alanjstr on March 11, 2003 11:15 PM