Linux imploding?
|
| Author | Content |
|---|---|
| Grishnakh Aug 22, 2011 2:21 PM EDT |
Is it me, or is Linux imploding? Between the Unity/Gnome3 debacle making it so it's hard to find a desktop Linux distro that actually works well on a desktop instead of a small touchscreen, continuing problems with video cards/drivers, and now stuff like this, it seems like Linux is going backwards instead of forwards. |
| dinotrac Aug 22, 2011 2:31 PM EDT |
No. Some bad things are happening in the Linux ecosphere, but that's been true for just about as long as Linux has been around. Lessee... kernel bugs and regressions? Yup. pretty much always. Desktop frolics? Lessee --- There was the QT licensing flap and the refusal of Debian to include KDE in their repositories, and -- who remembers xfree86? And -- ouch! -- The pre-Firefox days of Mozilla, when they decided that the world didn't really need any browser other than IE for the next few years. Stuff comes and stuff goes. It's all in the nature of the beast. |
| jdixon Aug 22, 2011 2:50 PM EDT |
> Between the Unity/Gnome3 debacle making it so it's hard to find a desktop Linux distro that actually works well on a desktop instead of a small touchscreen, XFCE has worked pretty well for a number of years now. And while LXDE may not be quite ready yet, it's working it's way into shape. And they're not the only options out there. But yes, it looks like we could use another good desktop environment that's actually oriented around the desktop. What a novel concept that would be. |
| montezuma Aug 22, 2011 3:17 PM EDT |
There seems to be a communication issue here. This problem has been reported on several distro bug lists for quite some time and nothing much has happened. Why is it not on the kernel bug list? Someone competent at git-bisect needs to go head to head with the kernel developers and sort this out once and for all. I am quite disappointed that such a person does not come from the ranks of distro kernel maintainers. It requires good knowledge of the kernel system and technical competencies. I think the guy from phoronix could also easily do this or is it more profitable for him to rack up page hits from frustrated laprop owners? |
| Steven_Rosenber Aug 22, 2011 3:34 PM EDT |
I have a Debian Squeeze LXDE system that I'm setting up. I still have about a dozen things I need to intall/configure before I can use the system in my work, but I'll write about it when I have something to report. |
| Fettoosh Aug 22, 2011 3:56 PM EDT |
Yeah? And the sky is falling too? Fear no more, Atlas is still supporting the heavens. |
| lcafiero Aug 22, 2011 8:29 PM EDT |
I don't know about this article, but then Phoronix articles that have to do with benchmarks (which is most of them) always cause me to read them several times, very slowly, before reaching for the nearest aspirin bottle. On one hand, it shows that there's increased power consumption on a notebook with an i5 Sandy Bridge Intel processor (a quad, if I recall correctly), but when they run the test on a Lenovo with an i7 processor (another quad?), the system, according to the article, is not affected by this regression. Obviously, I'm not a hardware guy and clearly there's a better than excellent chance I'm missing something here. But it looks like this is an issue around the Sandy Bridge processor and not the other one (the i7) or maybe even ones that haven't been tested. But to get back to the topic in question, Linux imploding? Um, no. This, the GNOME 3/Unity thing, video card travails -- we've weathered far, far worse. |
| fewt Aug 22, 2011 10:13 PM EDT |
omgphoronix - If you believe anything they have to say, I have a bridge to sell you! There is no active work because there is no bug... |
| dinotrac Aug 23, 2011 5:39 AM EDT |
>There is no active work because there is no bug... Do you mean no bug report or no bug? If you mean no bug, do you have a basis for saying that other than your disregard for Phoronix? |
| Scott_Ruecker Aug 23, 2011 6:06 AM EDT |
Linux Imploding? Really? Come on now..the answer is obviously no. What we have had here as of late is one thing and one thing only..drama. And I for one am not worried about there being some good ole' fashion drama surrounding some of the software entities in the Linux Universe. Drama is human..FOSS was created by humans, hence there is drama in FOSS. ;-) |
| fewt Aug 23, 2011 7:19 AM EDT |
@dinotrac - The supposed bug only impacts a handful of machines, and was inserted into the kernel intentionally to work around buggy BIOS implementations. It isn't a bug, nor is it a regression. Relevant: https://bugzilla.redhat.com/show_bug.cgi?id=681017 http://lwn.net/Articles/449648/ http://lwn.net/Articles/449448/ |
| montezuma Aug 23, 2011 8:41 AM EDT |
fewt, That may well be the case and I recall reading an article on LWN sometime back by Matthew Garrett blaming lousy BIOS implementation for this issue. Nevertheless I have a highend Lenovo thinkpad laptop which have a reputation for good BIOSes and it runs about 10C hotter with the 2.6.38 kernel compared with the 2.6.32 kernel. There is a bug open on launchpad (Ubuntu) where literally hundreds of people have reported problems. This needs to be sorted. |
| dinotrac Aug 23, 2011 10:14 AM EDT |
@fewt : I can appreciate that the regression -- and a regression it definitely seems to be -- doesn't affect everyone. The fact that it may be the result of an intentional workaround makes it no less a regression and -- for those who were working fine before -- should call the workaround itself into question. That said, I completely agree that it is not a matter of linux itself going to hell in a handbasket. Lots of users -- probably most, maybe even a vast majority -- will be completely unaffected by it. |
| fewt Aug 23, 2011 10:54 AM EDT |
@montezuma - Install Jupiter. [url=https://launchpad.net/~webupd8team/ archive/jupiter]https://launchpad.net/~webupd8team/ archive/jupiter[/url] - The bug report at Ubuntu is just a bunch of "me toos" and not enough actual testing. Last I looked, the majority of me too bug confirmations had tons of apps open which of course all drive up power usage. @dinotrac - Fair enough, I'll concede that it would be considered a regression for some. |
| montezuma Aug 23, 2011 11:39 AM EDT |
fewt, I see that jupiter is your app. I have used the intel app powertop to carefully check this issue over several kernels (and with no other apps open) and the number of wakeups and the power consumption have definitely increased. I am frankly pissed that no kernel developers have asked any users on that bug report to run any more detailed tests. Unless that occurs all you get are a large number of people venting and a series of uncontrolled reports. |
| fewt Aug 23, 2011 2:56 PM EDT |
@montezuma - Correct, it is my app. The number of wakeups and power consumption will likely continue to rise over time because the kernel doesn't dynamically tune itself. That job is in userspace where it belongs. ;) Don't be pissed at kernel developers for fixing bugs, if you research these issues, they all boil down to hardware / BIOS bugs that the code was altered to work around. Talk to distribution manufacturers to include tools like Jupiter or tuned which are designed to interface with the kernel and tune for the desired state. |
| montezuma Aug 23, 2011 3:32 PM EDT |
fewt, This issue affects hundreds of users of Ubuntu. It should not be ignored. It should be properly investigated by a developer not left to unqualified users to research. Also please provide a link documenting the "dynamical tuning" of kernels and why this must be done in userspace. |
| Grishnakh Aug 23, 2011 3:42 PM EDT |
dinotrac wrote:There was the QT licensing flap and the refusal of Debian to include KDE in their repositories, and -- who remembers xfree86? While I'm glad to read that this particular issue is minor and only affects a few systems (and will probably be worked around with a "quirk" or something), it still seems like FOSS is regressing in many ways. Remember, regression is different from simply starting in a bad state. Yes, the days of many websites being IE-only were bad, but that's not a regression, that was just an initially bad state. FOSS was very young back then, and before Firefox, there weren't exactly a lot of FOSS browsers. Netscape, the main competitor to IE back then, was not FOSS, it was closed and proprietary (though a free download). Things have obviously gotten a lot better since then as far as browsers go, but there is a little regression; just look at all the complaints about the latest Firefox releases. Luckily, these days there's other browser choices, but FF is still probably the most popular FOSS one. The QT licensing flap, again, was an initially bad state. It's not like Qt started out with a GPL license and then switched to proprietary; it started out prop., and after the flap, they made it GPL, and then the problem was gone. There's been no regression. As far as XFree86, I don't remember any problems there, at all. That, in fact, was a prime example of why FOSS can be much better. XFree86 was the only FOSS X system around for a long time, used by all the distros. But there were a lot of problems in its organization, with many developers feeling constrained by the guy in charge. Then this guy made a stupid licensing change, caused a big controversy that was essentially the hair that broke the camel's back, and a big group of other devs got together and forked the project. This project then went into a development frenzy implementing lots of new features that had been held back under the old leadership. Within a few months, all the distros had switched to the fork, now called x.org, and no one used XFree86 any more. For users, there was absolutely zero disruption; they just upgraded their distros at some point and found Xfree86 packages replaced with x.org packages, and all was well. There was no regression there, only improvement by way of forking. It was much the same with the more recent OpenOffice->LibreOffice fork. No disruption to users, just a new version with a new name, and no regression. It's not like this with Ubuntu/Unity and Gnome3, however, because of the sheer number of people who have been depending on these projects; they've woken up one day to find a giant disruption to their workflow and their way of doing things, and it wasn't a positive improvement in the opinion of most users. That's a regression. It was much the same 3 years ago with KDE4.0, except the big difference there was that KDE wasn't nearly as popular as Gnome2 back then (and still isn't), so not nearly as many people were directly affected by these sudden changes, but otherwise it was the same: users upgraded to the next distro version only to find that their desktop environment was basically unusable. You'd think the Gnome devs, the distro maintainers, etc. would have learned from this bit of history, but obviously they didn't. jdixon wrote:XFCE has worked pretty well for a number of years now. And while LXDE may not be quite ready yet, it's working it's way into shape. And they're not the only options out there. But yes, it looks like we could use another good desktop environment that's actually oriented around the desktop. What a novel concept that would be. Luckily, with FOSS, we do have these options, as well as Trinity and perhaps the new Gnome2 fork. But it's suboptimal, in that none of these are well-supported by mainstream distros, which mostly just tend to use the latest Gnome (and a few use KDE). In the Linux desktop world, these other choices are much smaller, though that may be changing, esp. now that Linus himself has publicly declared his support for XFCE. In the meantime, however, you can't deny that this is all a big disruption to Linux users. |
| Steven_Rosenber Aug 23, 2011 4:42 PM EDT |
Would it kill GNOME and KDE to keep the older DEs in "maintenance mode" for a year or so until the newer version could catch up in terms of features and bug fixes? |
| dinotrac Aug 23, 2011 4:52 PM EDT |
>but that's not a regression, that was just an initially bad state. It was most definitely a regression. Once upon a time, Netscape was the leading internet browser, and it ran on Unix -- including Linux. It had lost significant share when the Mozilla team abandoned it, but retained a significant player. Mozilla abandoned Netscape in order to do its own thing and ceded the web to IE. |
| fewt Aug 23, 2011 7:18 PM EDT |
@montezuma - Welcome to Linux where every user has to pull his weight if they want something fixed. Don't like it, pay for Windows or OSX. As for documentation, http://kernel.org/doc/, enjoy. :D |
| JaseP Aug 24, 2011 10:44 AM EDT |
fewt, You are assuming that M$ and Apple actually fix their problems. Some of them are now, "features." User work arounds don't just exist in the Linux world. There are plenty of them in every OS. Historically, in fact, developers used broken Windoze APIs to extend particular functionalities, taking advantage of software quirks to accomplish certain goals. Specific, examples escape me, but graphics drivers come to mind. |
| fewt Aug 24, 2011 11:42 AM EDT |
@JaseP - I didn't make any assumptions. Proprietary software is designed for people that don't want to fix their own problems. That doesn't mean I implied that they were guaranteed a fix, however those platforms provide paid support options to get issues resolved faster. If Microsoft and Apple didn't fix problems they wouldn't have any users. They may not fix every single problem, but no software development shop does. |
| JaseP Aug 24, 2011 12:35 PM EDT |
In my experience, most problems get fixed faster in FOSS... The one's that don't are the ones that we are talking about,... The disconnect between the developer and the user (this seems to reflecting these power management code changes and the Gnome/Unity issues). ... But that same kind of problem exists in Windoze too... case in point; switching from Office 2003 to Office 2007 means a huge (and unnecessary) interface change. It's actually easier to switch to OpenOffice/LibreOffice than make the M$ Office 2003 to 2007 change. It seems there's a lot of leaping before looking in the developer community. |
| Fettoosh Aug 24, 2011 1:18 PM EDT |
Quoting:Welcome to Linux where every user has to pull his weight if they want something fixed. Don't like it, pay for Windows or OSX. Linux user don't have to switch and pay for Windows or OSX to get support, Red Hat or Suse do have full support for their products. Canonical have the same. Look around the Internet and I am sure one would find other outfits that are willing to offer Linux support to users who are willing to pay for it. Quoting:If Microsoft and Apple didn't fix problems they wouldn't have any users. They may not fix every single problem, but no software development shop does. Not always true. Users of proprietary software are pretty much locked-in and pretty hard to just pack up and switch. Case in point, the Windows security issue has never been fixed for a long time, but kept its users for a very long time. It is only when competition began to threaten their business they started to take the issue seriously. Problems in proprietary software are fixed when its vendor is willing and only on its own schedule. Problems in FOSS software are fixed much quicker and most of the time for no charge. If one needs immediate action taken to resolve an issue/bug in FOSS, most of the time developers would be more than happy to oblige if compensated. |
| fewt Aug 24, 2011 1:46 PM EDT |
@Fettoosh - You keep telling yourself that. The rest of us who know better will continue to know better. |
| skelband Aug 24, 2011 3:22 PM EDT |
@fewt: The landscape is far from straightforward. Good luck with getting most large proprietary software houses to look at any kind of enhancement or bug fix (unless it is very serious) unless you are a very important user. Examples abound: Oracle and Microsoft are two that I can think of. Games houses like EA tend to be quite receptive to bug reports however. Apart from a few notable exceptions, free software that is well supported can turn around bugs and enhancement pretty quickly. Your mileage will definitely depend on how busy they are, how active the development is and how receptive they are to requests. |
| dinotrac Aug 24, 2011 3:22 PM EDT |
@fewt -- Fettoosh is right -- sort of. Though it's a gross over-simplification, FOSS problems seem to come in two varieties: Those that are fixed quickly, and Those that are never fixed at all. Interesting problems, or, more specifically, those that are important to the developers, tend to get handled quickly. Not-so-interesting problems may languish forever. There is one GIANT asterisk to that characterization, however: FOSS is not the exclusive domain of volunteers. IBM, Oracle, Red Hat, Canonical, Suse, assorted equipmnet vendors, and many others pay people to work on FOSS. Problems in areas that interest those companies may get resolved even if they don't excite a single developer anywhere. |
| montezuma Aug 24, 2011 3:31 PM EDT |
fewt, I have contributed much to the Ubuntu community through bugreports which in the past has been appreciated by many Ubuntu developers. If Canonical wants to take my contribution for granted and exploit me for commercial gain I will take imy efforts elsewhere. The whole point of Ubuntu in the beginning was as a community. This seems to be rapidly fading. |
| tracyanne Aug 24, 2011 10:18 PM EDT |
Quoting:The whole point of Ubuntu in the beginning was as a community. Only in so far as it provided an excellent marketing opportunity for a commercial company. This is not to say there is anything wrong with doing that, but that was the real point. |
| dinotrac Aug 25, 2011 7:30 AM EDT |
@ta - Yup. I'm amazed that anybody could have thought differently. Canonical has never hidden their role. |
| fewt Aug 25, 2011 8:04 AM EDT |
Quoting:Good luck with getting most large proprietary software houses to look at any kind of enhancement or bug fix (unless it is very serious) unless you are a very important user. @skelband - Well Microsoft has fixed bugs for me, and written knowledge base articles about problems I have found so I know this is not really true. All you have to do is pay for the support incident and follow the process. Your implication that they won't tells me that you haven't really tried. It wasn't a fun process, but I was shoved into it and asked to fix it and that was the net result. It was related to MS Exchange, and SSL if you are curious, 6 or 7 years ago. Quoting:Those that are fixed quickly, and Those that are never fixed at all. @Dinotrac - This is absolutely true. Launchpad for example has many of examples of this, including bugs that are 3 or more years old. My point wasn't to imply that bugs just don't get fixed. My point was if you want someone to be accountable for bugs, then proprietary software may be a better route to take. That doesn't mean it's ideal, but that's just how it is. Canonical is the only "company" that is focused on Linux "desktops", and they don't really have anyone working in the kernel. (note: See rants by Gregg K-H) Quoting:The whole point of Ubuntu in the beginning was as a community. This seems to be rapidly fading. @montezuma - I disagree. The whole point of Ubuntu was to build a business by exploiting the community. I'm glad to hear that you do issue bug reports, but I might recommend another distribution if you actually want those bugs fixed. |
| tuxchick Aug 25, 2011 8:51 AM EDT |
Why should Ubuntu be criticized for not being a significant kernel contributor? Greg K-H doesn't deliver a desktop distro. I don't agree with the characterizations that Canonical is exploiting 'the community' for profit, either. For one thing after all these years there is yet no profit. We can't have it both ways, and be all happy that someone like Canonical spends squillions- including developer salaries- supporting a distro, and then go all peevish that they might want to earn some revenues. |
| montezuma Aug 25, 2011 9:02 AM EDT |
fewt, You missed my point. I have lodged over 50 bug reports for Ubuntu and in the first few years they were nearly all taken seriously by the developers which is the reason I stuck with the distro since that indicated they were taking the community thing seriously. This has changed in the past 2-3 years. Being cynical about Canonical's role is easy. I can sneer as well as the next person. The problem for Canonical is that a business model based on exploiting a community and not giving back is a business model doomed to failure. The only way Ubuntu will work is with an engaged community. Without this it will rapidly go downhill. I am beginning to think that Ubuntu/Canonical isn't a business model at all. More like a big ego trip by Mark Shuttleworth. |
| fewt Aug 25, 2011 9:10 AM EDT |
Quoting:Why should Ubuntu be criticized for not being a significant kernel contributor? @tuxchick - Why shouldn't they be? Quoting: We can't have it both ways, and be all happy that someone like Canonical spends squillions- including developer salaries- supporting a distro, and then go all peevish that they might want to earn some revenues. They will need a real business model if they want to be successful. Performing actions such as changing the affiliate code in Banshee to take money that was being earned for GNOME is valid evidence that Canonical is willing to exploit the community to earn a buck. Quoting:I am beginning to think that Ubuntu/Canonical isn't a business model at all. More like a big ego trip by Mark Shuttleworth. @montezuma - +1. I realized that as I was writing my Ubuntu rant a few years ago. |
| tuxchick Aug 25, 2011 9:26 AM EDT |
Ah fewt, dodgey as usual. |
| Fettoosh Aug 25, 2011 9:54 AM EDT |
Quoting:You keep telling yourself that. The rest of us who know better will continue to know better looks like you can't handle facts and this is your easy way out. Hmm, must be the training in Redmond. |
| fewt Aug 25, 2011 9:59 AM EDT |
Quoting:Ah fewt, dodgey as usual. @tuxchick - Hardly, I gave a specific example. Quoting:looks like you can't handle facts and this is your easy way out. Hmm, must be the training in Redmond. @Fettoosh - ok, like I said, believe what you want. |
| dinotrac Aug 25, 2011 11:43 AM EDT |
@fewt - Why shouldn't you be criticised for not bothering to broker a Middle East peace deal? Or not making a battery that can power an electric car for 500 miles and be recharged in 5 minutes? You certainly have a vested interest in a world that doesn't explode into global war, melt down in global warming, or run out of energy. But you have other things to do and other contributions to make. We don't criticise police officers for not putting out fires or firemen for not catching criminals. Why would we criticise Ubuntu for not contributing more to the kernel? There are other things to do in Linux, too. I remember listening to Linus address attendees at Linuxworld in New York more than ten years ago. At that time, he said that it was more important to make contributions to user space than the kernel. I believe that still to be true. The kernel works pretty damned well and lots of people contribute to it. User space? Not so much. As to exploiting the community, so what? Where is it written that there is anything whatsoever wrong with that? Has Canonical lied to anybody about who they are or what they are? People who willingly take part in something where nobody is hiding the truth have no right to whine because somebody makes money. Ditto for those who write GPL'd code only to see somebody take their code and run with it -- while turning a few dollars. The license specifically allows that. |
| montezuma Aug 25, 2011 11:51 AM EDT |
Dino, I agree that Canonical are open about their motivation and good luck to them in making money. What I object to is that some of the value that has been created in Ubuntu is down to a large number of volunteers. Canonical's role has been in the past to provide a crack team of developers who worked with the community to provide a great distro. If they take that army of volunteers for granted by not responding to serious bugs which are a major annoyance to users then that army (which has included me) will take our contributions elsewhere. That strikes me as very stupid of Canonical. |
| gus3 Aug 25, 2011 11:53 AM EDT |
@fewt: A product fails to work as advertised, but you have to *pay*to get the fix? I ain't buyin'. |
| fewt Aug 25, 2011 12:57 PM EDT |
Quoting:A product fails to work as advertised, but you have to *pay*to get the fix? @gus3 - I challenge you to find a product that doesn't have bugs. Having bugs != not working as advertised. |
| Fettoosh Aug 25, 2011 1:33 PM EDT |
Quoting:I challenge you to find a product that doesn't have bugs. Having bugs != not working as advertised. Dodgy , you are missing the point. It isn't software without bugs, it is having to pay for fixing it after you already paid for it up front. |
| fewt Aug 25, 2011 2:02 PM EDT |
Quoting:Dodgy , you are missing the point. It isn't software without bugs, it is having to pay for fixing it after you already paid for it up front. @fetoosh - You are just making excuses about something that you don't like. |
| skelband Aug 25, 2011 2:21 PM EDT |
@fewt: "All you have to do is pay for the support incident and follow the process." Well I didn't say they wouldn't fix your bug, just that you have to be an important customer. Paying for support makes you important. You have to wonder though that if you have bought one of their products and you are taking the trouble to let MS know that there is an issue in their software, that they feel that they need to charge you for that privilege, there is something wrong with their culture. You are helping them, not the other way around. I guess that is my point. |
| fewt Aug 25, 2011 2:42 PM EDT |
Quoting:You have to wonder though that if you have bought one of their products and you are taking the trouble to let MS know that there is an issue in their software, that they feel that they need to charge you for that privilege, there is something wrong with their culture. Their business model requires that they have an income to feed the people that fix bugs. It isn't a privilege. Besides, they in many cases will waive the fee if it is in fact a bug. |
| dinotrac Aug 25, 2011 2:57 PM EDT |
@montezuma : Well, can't argue with you there. Stupidity has been know to show it's ugly little pointed head in more than a few corporate cultures. |
| skelband Aug 25, 2011 3:30 PM EDT |
@fewt: "Their business model requires that they have an income to feed the people that fix bugs. It isn't a privilege.." That there is the problem. After sales. The money is got, anything after that is a liability. From a pure business perspective, I can see why they would think that. Everything before the sale, marketing, r&d, management costs, the guys that were supposed to fix the bugs before release, pretty much everything that it costs to run a company is paid for by the purchase price of the software. Then there is support, that unfortunate bit afterwards, where you look after your customers. It speaks volumes about the ethos of the company. It's unusual in the business world though, isn't it? If I buy a vacuum cleaner and it has a fault, the manufacturer will repair it, or replace it free of charge, within a reasonable period of time. After all, they want me to buy another one of their products in the future don't they? If I buy a car and it has a safety issue, then they go to huge lengths to get them all fixed, cost be damned (unless you're Toyota of course, and you see where that ended up). If I buy some corn flakes and they're not quite up to what I expected, they will give me my money back, no questions asked. We as consumers expect perfection and most manufacturers go out of their way to give it to us as far as is humanly possible. Strange then that if a software house fixes a critical bug in a piece of software, they potentially make a lot of their customers happier, for relatively small cost by making the fix available on their website, and for my assistance in this, I have to pay them. It's been that way for as long as I can remember and for that reason, we accept it as normal. |
Posting in this forum is limited to members of the group: [ForumMods, SITEADMINS, MEMBERS.]
Becoming a member of LXer is easy and free. Join Us!