I have spent the entire day thus far going through bug reports on bugs.kde.org for plasma and krunner. This has resulted in a number of fixes and patches. Unfortunately the bulk of my time has been spent triaging bugs that really don't end up providing value; worst are the ones where I triage them and the reporter decides that, despite my response, I haven't quite understood things properly and so continues to engage on the bug report further wasting my time. As such the cost to the user (less software) is too well hidden compared to the cost of reporting a new bug or replying to an existing one (which is the least expensive thing to do).
There no real moral to this story, really. I do wish we had a better bug reporting and triaging system that supported real work flow, external bug repository tracking and other sane features, but I really see no solution to broken user habits. There are too many users, and too many stubborn people.
Update: It just occurred to me to mention that if you have a request or a thought about things (versus a straight forward bug report, e.g. "I did $FOO and Plasma crashed" or "I searched in KRunner for $FOO, but it returned $BAR instead"), try the mailing list (panel-devel at kde dot org) first. bugs.kde.org is insanely inneficient compared to the mailing list both in terms of time it takes to use it as well as discussion usefulness (threading, for one). If an action item comes out of the mailing list discussion, then create a report on bugs.kde.org that is succinct and to the point.
On a more positive note, I went out to Banff with my houseguest, W., on the weekend. We spent time wandering around the town (buying waaay too much fudge =), wandering through museums, wandering through forests (where we took pictures of various deer and ate wonderful little sandwiches we made amongst the trees) ... essentially a lot of peaceful wandering in the sunshine.
Much less stressful than dealing with bugs.kde.org all day, but also less productive for Plasma ;)
Subscribe to:
Post Comments (Atom)
10 comments:
One thing which could at least help a little would be to have a set of helpful standard phrases in KDE's Bugzilla.
For example, if I file a new bug report in RedHat's Bugzilla there are automatically phrases in the bug description field like "how can the bug be reproduced", "Steps to reproduce the bug", "What was expected" or "What actually happened".
That does indeed help me that much that I sometimes copy these phrases fom RedHat's Bugzilla to KDE's Bugzilla.
yeah, I also wish sometimes that bugzilla would be a bit more then some kind of bulletin board system. E.g. allowing to connect single reports dealing with a similar but not the same thing or a possibility to mark some wishes as "will be done but another way" (funny how some reports don't only ask for a special feature but also how it should be done technical) or a way to mark them as "don't show me this report till 4.1 is released" or... Well, it's just hard to keep track of so many issues and that some of them are not interesting till some other things are done while others need to be addressed asap doesn't make it more easy.
The workflow-part is also a rather interesting thing. To have here the mailinglist where changes to reports are mailed to helps already a bit since it allows to keep track of changes but still it does happen for every report rather then only those ones someone is interested in. What I also do is sometimes to add others to the CC_MAIL-list if I guess that they could help there somehow to future progress with the report. But imho that's also not really that effective.
And then there is still the problem that everything related to an issue is within one report and after 50+n replies it starts to get confusing and eats additional time to follow up. It even gets more worse if backtraces, screenshots and patches are attached and if someone is forced to review a patch within the same report (though I just discovered the possibility to reply to a patch. So, seems someone already saw the need for it).
Indeed a standard-template like suggested by liquidat would be probably really already a good start to at least prevent those "xyz does not work!!1!" reports and save the time needed to reask the first time ;)
Since some/most of the aggravation above was no doubt caused by me, allow me first to sincerely apologize, and then to elaborate...
The statement of "KDE has a problem" with the proprietary NVIDIA driver causing widget/panel corruption with KDE 4.x was not meant to be interpreted as meaning the issue is KDE's fault or KDE's responsibility, but rather it could be a major obstacle/hindrance for KDE4 adoption.
As I see it, linux (generic) is on the cusp of an OEM install explosion, fueled in part by the "invention" of the mini-notebook (and of course by the hard work of the community). When Asus and the like look forward to their next-gen Eee Pc's, will they choose a window manager that can't convince a vendor to get a major glitch fixed after nearly 7 months? Or, will a Third World school system decide to install a window manager on their existing workstations when it might not run properly on a significant portion of them?
NVIDIA driver bugs might not be KDE's responsibility but they are their problem. NVIDIA has been the default forum recommendation for graphics cards for as long as I can remember... those pre-8xxx cards are expected to "just work". When KDE doesn't work on them, it's KDE's problem until KDE makes it NVIDIA's problem.
KDE cannot first hand fix the bugs. That much is clear. But surely they can exert pseudo-political pressure on NVIDIA to resolve this issue, by being very vocal and visible with the frustration, rather than maintaining an attitude of "it's not my problem".
Now certainly "it's not my problem" is the correct response from the plasma-dev, and I again apologize for using a bug as a forum for discussion. I regret that. I honestly do. But when I read through the multiple "duplicate" bugs, and the cited nvforum threads that question whether or not the different rendering bugs are related, I raise my hand have have to ask, after 7 months, does anyone really know what the problem is? Is there one bug or three? Is anything being done to fix it?
And finally, if the "stubborn people" comment was directed at me, I would like to point out that I commented on the closed bug first (half expecting no action due to its status) and then posted relevant info. in the open bug... your post would suggest it was the opposite. And again, sorry. I am an advocate and an enthusiast (enthusiasm and frustration together rarely a good thing)... I certainly don't want to be counterproductive to the project. Please feel free to ignore this post. I'll move my focus to NVIDIA and my distro where it best belongs, as directed. Thank you for your time.
@Heath: that was hardly the most annoying bug of the day. trust me. there were far worse.
the biggest problem is that there were not just 1 or 2 but several.
to be honest, i have to say that as a group our user base is out of control. i have some ideas on how to address that, since the user base itself seems unable or unwilling to do anything effective about it internally.
anyways .. nvidia.
have you ever wondered why those little systems such as the ASUS one use Intel chipsets for these things? in part because they work and have open drivers.
NVidia did fix the worst of their bugs, but then they regressed on some of them, and others have emerged ...... it's just not a pretty sight.
i'd rather see such companies fail in the marketplace than us feel compelled to solve problems only they can actually solve.
and right now, when we look at small devices, that's exactly what's happening.
Just move kde's development to LaunchPad.net
First of all, I am an ardent fan of kde4 and i am loving its progress.
I love your as well as the whole kde team's work :)
My one humble opinion/request to you is, please don't rant about userbase/bug reporters in a write up like blog because it has such a huge audience.
This will surely have -ve impact.
I can understand the reason for such, but once you are out of it you will be blaming yourself for blogging about it.
Happy hacking :)
Don't be so stressed ;)
Well, one thingthat can be improved is to be able to submit the crashes back to KDE automatically.
I'm not suggesting to make automatically bugs out of them, but submitting crash reports to some kind of crash-report-repository can be very useful I think I think.
What is the reason kde is still using bugzilla anyway? I don't want to put you down but posting kde-bugreports on bugzilla is one of the worst experiences you could give new users and betatesters. I've resorted to posting my stuff on Ubuntu/Launchpad and hoping that the supplied fixes then hit mainline through that way at some point(which worked quite fine with a few Kopete-issues I had with 4.0.0). And since you say we should provide concrete starting points when criticizing stuff, here are some for Bugzilla:
- It's not visually appaling
- It's not providing the same amount of hints/failsafes as launchpad
- It needs to be purged of all the "change behavior x, and y instead", or at least should move them to some kind of "suggestions/wishlist"-category like launchpad does
>> to be honest, i have to say that as a group our user base is out of control.
I don't think it's such a good idea to insult the people that are making the effort to report bugs.
>> i have some ideas on how to address that, since the user base itself seems unable or unwilling to do anything effective about it internally.
You seem to think this is an organized group that can make changes to its own behaviour.
Anyway, in your next post you accuse Celeste of waving her hands in the air without concrete examples. Well that's precisely what you're doing here. Give some links to example bug reports where the reporters were being obstinate and you would have a much stronger case.
@Leo_S: "I don't think it's such a good idea to insult the people that are making the effort to report bugs. "
it's not an insult, it's an observation. in a participatory model where we allow anyone to touch the bug database, the consumer-customer relationship is changed dramatically.
"You seem to think this is an organized group that can make changes to its own behaviour."
it is organized, only not consciously. it's a chaotic system with real patterns of movement and behaviour. this isn't a necessity, however, and something that i think can, should and perhaps even must change if we are to deal with the coming years successfully.
"you accuse Celeste of waving her hands in the air without concrete examples. Well that's precisely what you're doing here. Give some links to example bug reports where the reporters were being obstinate and you would have a much stronger case."
firstly, the difference is i'm not asking someone els to figure out this problem. i'm thinking out loud rather than trying to engage others in a discourse with questions. there's a difference.
secondly, i've actually dealt with these individuals one on one versus publicly humiliating them in my blog entry. you talked earlier about respect.
if you want to look at some examples of the same sort of behaviour, though, go read my blog entry on "no more desktop icons" and just watch it all unroll frome there.
it's insane.
Post a Comment