- Site menu:
- Latest Discussions
- Latest Newswire
- Archives
- Submit Story
- Story HOWTO
- Syndicate
- E-mail Alerts
This week at LWN: KDE 4, distributors, and bleeding-edge software
|
|
|
Comments (25)
5 earlier comments in the full thread.
Well well well, interesting, at last I feel somewhat vindicated. This is the bloke that basically told me, more or less, in an email, that I should be grateful for all the work he and the KDE4 team have, even though 4 (4.1 at the time I contacted him by email) was pretty much unuasable.
I suppose this is the closest we'll ever get to an apology from the bloke and the KDE4 team.
I suppose this is the closest we'll ever get to an apology from the bloke and the KDE4 team.
> I should be grateful for all the work he and the KDE4 team have, even though 4 (4.1 at the time I contacted him by email) was pretty much unuasable.
And Ballmer cheered. That "open source is stuff we do in mom's basement" is just the mentality Microsoft wants to see. If you can't take the feedback, get out of the kitchen, so to speak.
And Ballmer cheered. That "open source is stuff we do in mom's basement" is just the mentality Microsoft wants to see. If you can't take the feedback, get out of the kitchen, so to speak.
bigg -
**Applause**
The thing that really irked me on all this is the notion that we should just take what we're given when, in fact, on reason developers go to a major project is the notoriety of working on something everybody knows and thinks is important.
Well, kiddies, with that notoriety comes a certain amount of responsibility. As a wise individual once wrote, if you can't take the feedback, get out of the kitchen, so to speak.
**Applause**
The thing that really irked me on all this is the notion that we should just take what we're given when, in fact, on reason developers go to a major project is the notoriety of working on something everybody knows and thinks is important.
Well, kiddies, with that notoriety comes a certain amount of responsibility. As a wise individual once wrote, if you can't take the feedback, get out of the kitchen, so to speak.
Quoting:Whatever they do, the immutable law is simple:
If you want a lot of people to use something, it had better be something a lot of people can use.
That, my friend, is exactly the problem. They don't wanted people to _use_ it, because they know it had / has problems, but they want people to _test_ it. They assumed calling it 4.0 would make people test it, and it turned out calling it 4.0 made people use it while it was not that usable.
It's quite all right a user doesn't know KDE 4.0 is for testing purposes and not for day to day use, and sure mr. Seigo is to blame for that.
However, I have to agree the distributions were stupid to put something that's for 'testing' in 'using' releases; those making a distribution are insiders and they should know better. Now matter how something is called, if it breaks the stability of a distro you should not include it in a distro meant to be stable. I mean, taking a major platform like KDE4 that even a distro with a lot of testers / bleeding edge users such as Gentoo deems unstable and masks and then putting that platform in a distro meant for end-users is indeed really, really stupid. Release engineers should not let themselves be tricked by release numbers, they should know better. And if Fedora or Mandriva ends up to be far more unstable than Gentoo, then it's really cheap to blame the KDE people for that; because if some hobbyists project such as Gentoo without very much users can get it right, then there is simply no excuses why big distro's with even a company behind it can't and have to screw up like this.
I see it like this: If you as an end-user install Vista because Microsoft says it's ready and it's not, then Microsoft is to blame. If some ISP you'd normally trust installs Windows Vista for you and it makes your system unusable because Vista is full of bugs, you shouldn't blame Microsoft because your ISP should have known that. They are the one providing service to you and they are the one who should know if something is stable or not. So they are the one who screwed up by installing Vista, assumed they could have installed something else that did work (clearly in the case of KDE this was the fact, they could have provided 3.5). ISP's / release managers of distro's should immediately see through the PR-stuff and know what's working and what not. If an ISP pulled such a trick / screwed up like that by just blindly believing claims from some company or group, I'd be looking for another ISP. If my distro pulled such a trick, I'd be looking for another one which in fact _does_ have a clue about which software is stable and which is not.
Obviously however, I ended up as the only one defending Aaron and KDE again. Hmm, I don't know why, maybe because I saw the bloke in real life once, or because I just like to point at how some popular distro's really screwed up and now try blaming KDE while Gentoo didn't screw up for a change. Needlessly to say I don't use KDE; maybe not having to deal with the bugs in it might make one more objective and less angry at the KDE-team because it's a "far-from-my-bed show" as they tend to say over here.
However, all this finger-pointing doesn't get anyone any further, so the thing I'd like to ask to my fellow LXer-readers is: Certainly this model is flawed. According to Mr. Seigo, but of course he can be wrong, calling it 3.99 or 4.0 Alpha instead was not a solution either, because then there would have be too little people testing it. So these two solutions fail. Given this boundaries, what is the solution (except for the logical answer distro relengs should do a better job)?
hans -
The solution at the end starts in the beginning. It's deciding how much you're going to bite off and how much you can chew. It's remembering that you have a user base and are not just working on some hobby.
KDE folks are not the first to fall into the trap -- Mozilla, anybody? -- and I'm sure some valuable learning will be had by all.
The solution at the end starts in the beginning. It's deciding how much you're going to bite off and how much you can chew. It's remembering that you have a user base and are not just working on some hobby.
KDE folks are not the first to fall into the trap -- Mozilla, anybody? -- and I'm sure some valuable learning will be had by all.
hkwint, where is the line drawn? boy howdy thats an easy one to answer, you draw the line prior to the point where you intentionally mislead people and sucker them into something that they normally would not have chosen to do.
So I think when you have to deviate from / bend honesty and integrity to accomplish your goal ... the line needs to be drawn short of that point. Deceitful practices are never acceptable.
So I think when you have to deviate from / bend honesty and integrity to accomplish your goal ... the line needs to be drawn short of that point. Deceitful practices are never acceptable.
az.... Thank you!
> ...calling it 3.99 or 4.0 Alpha instead was not a solution either, because then there would have be too little people testing it.
And exactly how would they know this? It's not like they tried it.
And exactly how would they know this? It's not like they tried it.
jdixon -
And, of course, that very argument means that they were aware that their actions represented a lie: Something to get people using the software who would not use it if they knew the truth.
And, of course, that very argument means that they were aware that their actions represented a lie: Something to get people using the software who would not use it if they knew the truth.
jdixon:
Even Microsoft has observed that. Bugs that were never observed during alpha or beta, suddenly become glaringly obvious in general release.
Even Microsoft has observed that. Bugs that were never observed during alpha or beta, suddenly become glaringly obvious in general release.
Distributors and the KDE team must each take part of the blame. Why don't we just get over it and start discussing KDE 4.2?
@mort KDE 4.2 looks like it might be what should have been the release to general users version, as far as I can tell from testing Mandriva 2009.1 Alpha 1 - I can't test Alpha 2 as it seems to have broken VirtualBox compatibility
I like to stick to the general rule that one should never expect a .0 release of a product to be entirely right. And indeed KDE4.1 is quite usable. There were some shortcomings, but I fixed em with a few tweaks.
Will install 4.2 soonish.
Will install 4.2 soonish.
Now don't get me wrong, I am currently using 4.2 and have been for a week or so. Except for the odd bit or piece that is incomplete, and I am unsure exactly where that those bits come from, although using Sabayon/Gentoo I can be fairly certain it KDE.
It is the manner in which they arrived at this point that bothers me.
It is the manner in which they arrived at this point that bothers me.
> Bugs that were never observed during alpha or beta, suddenly become glaringly obvious in general release.
And that's true no matter when your release occurs. That's no excuse for shipping alpha code as production ready, which is what the KDE folks did. They act like no one who uses Linux ever trys beta code, while from what I've seen it's a fairly common event. Some folks even use alpha code if they need the functionality it provides.
Regardless, if they don't have enough alpha or beta testers, that's their problem, not the distros and not the general users of KDE, who by definition probably don't know enough to be alpha/beta testers.
I'm sorry, but the "we had to release our alpha code as .0 to get enough testers" excuse doesn't cut it. As has already been noted, that's the way Microsoft works. They're hardly and example we want to emulate.
And that's true no matter when your release occurs. That's no excuse for shipping alpha code as production ready, which is what the KDE folks did. They act like no one who uses Linux ever trys beta code, while from what I've seen it's a fairly common event. Some folks even use alpha code if they need the functionality it provides.
Regardless, if they don't have enough alpha or beta testers, that's their problem, not the distros and not the general users of KDE, who by definition probably don't know enough to be alpha/beta testers.
I'm sorry, but the "we had to release our alpha code as .0 to get enough testers" excuse doesn't cut it. As has already been noted, that's the way Microsoft works. They're hardly and example we want to emulate.
Given the widespread use of virtualization, I really doubt that they would have difficulty finding people to test. Make packages and Virtualbox images available, and you will have plenty of people willing to test.
bigg, jdixon, et al...
The sad thing is that KDE never lacked for testers in the past. They had a history of putting out betas that were solid enough to use. I can't remember the last time -- before now -- that I waited for a x.0 release to upgrade.
Sure, I suffered some glitches, but, overall, things were livable.
Had they only followed that approach -- figured out how to make a livable product -- they'd have gotten all the testers they wanted.
The sad thing is that KDE never lacked for testers in the past. They had a history of putting out betas that were solid enough to use. I can't remember the last time -- before now -- that I waited for a x.0 release to upgrade.
Sure, I suffered some glitches, but, overall, things were livable.
Had they only followed that approach -- figured out how to make a livable product -- they'd have gotten all the testers they wanted.
Quoting:the general users of KDE, who by definition probably don't know enough to be alpha/beta testers.That's a pretty broad brush in your hand...
> That's a pretty broad brush in your hand.
Possibly, though I'm just assuming historical statistical probabilities and distribution. Are you offering to perform the survey of KDE users to find out what percentage actually are comfortable performing alpha/beta testing? :)
Possibly, though I'm just assuming historical statistical probabilities and distribution. Are you offering to perform the survey of KDE users to find out what percentage actually are comfortable performing alpha/beta testing? :)
Post a comment
A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.
Your message:Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help
Sign up and post (it is free) or Log in and post
If you want a lot of people to use something, it had better be something a lot of people can use.
Whether you call it Vista or KDE 4.0, people tend not to take well to half-baked crap.
At least Microsoft has its huge network effects and market position to buffer it.
The KDE user base has more alternatives.