I've been recommending people change flamebait to +5. Flamebait is code for "goes against herd thinking". The fact that I can change the moderation levels of tags basically means you have to accept that people do filter and want it. +5 Flamebait of course relies on herd thinking to work, so I in fact think the defaults are acceptable.
Just put in a job description for HR competent enough to hire engineers, with reasonable pay. Let them anguish over the advertised salary numbers. Or put in your own resume to make HR work for the company. Until your HR understands what it is your engineers do and what to look for, they're not serving the company correctly.
That's because it shouldn't be that way. The problems an engineer faces in business isn't mathematical, but social. An engineering team shouldn't have to temper what he tells his boss based on whether the boss will overstate his role in success or understate his role in failure. They shouldn't have to worry about whether they'll be passed over for promotion because they refused to sign off on the new design citing safety test failures. I don't think business school is all that hard, and judging by the parties the business majors next door hold, I don't think they've ever dedicated a weekend to working on a school project.
We need big firms that understand engineering, not engineers that understand business. The latter creates large numbers of contractors and small companies, while the former can take advantage of economies of scale.
Of course the second one was flagged, there was a fucking #ifdef PURIFY right next to it. An entropy pool basically has randomness added to it. The problem is that the destination is one of the operands in every case, so there's no initial starting value by design. Static analysis tools decide that adding a random number to an uninitialized field doesn't make the field initialized. So the pool is never initialized since that would remove some magic entropy that probably doesn't exist but certainly doesn't hurt.
Static analysis tools are common in the open source world. The lint name is well known enough that many projects make theirs a pun on it, ala lintian. A few years ago a local root exploit in X was discovered by running these sorts of checks. But generally, static analysis tools require human review -- with large code bases they generate large numbers of false positives, especially the dumber ones. This leads to trouble for perfectionists, a common trait among software developers interested in bug fix analysis. For example, the recent massive Debian vulnerability was caused by an overzealous developer trying to fix static analysis flags. One of these flags was valid, one was not, and removing both removed nearly all entropy from the RNG.
In the more general sense, static analysis cannot find all bugs. There's a trivial proof: a program stuck in an infinite loop is a bug, but finding all such loops would solve the halting problem. Handling interrupts and the like also causes reasoning problems, as it's very hard, if not computationally intractable, to prove multi-threaded software is safe. So static analysis won't rid the embedded world of watchdog timers and other software failure recovery crap.
Her laptop cost 400 dollars. You can't buy anything Mac for that, let alone a laptop.
But here's a surprising bigger problem that I've seen a couple times now: for the computer literate elderly, learning a new system might not have a positive ROI. One of my alma mater's retired faculty ran DOS until something like last year, and one reason I've heard is that he didn't think he could port the software he wrote to a new OS before he died. Apparently someone convinced him that Linux fit the bill, perhaps it was DOSbox that set him Free. The same reasoning sadly can apply in the grandmother case. It's a bit weird as a young man to see people apparently planning to die, but it's hard to deny there's a logic to it.
Windows and computers don't pass the grandmother test. I should know as she just visited today and very nearly the first topic was that her computer was broken and I should look at it later. The conversation goes along on the couch about the new house and the new internet service (ATT u-verse) my mother bought. Grandparents mention a horrible experience switching ISP providers. They switched from DSL to cable a few years back, and decided to switch back yet again. They didn't realize it at the time, but they were a bit screwed -- nobody knew the old password or how to reset it. 1-800 wasn't very helpful here and quite frustrating. Eventually a neighbor who worked for ATT came along and figured it out for em. "Never again will we change ISPs!" she says.
Later on we get to looking at that laptop. First thing I notice is it's quite new, still with stickers on it, yet there's a wifi pcmcia card. I check underneath and clearly see a Wireless MAC [random mac address]. Apparently she had bought a wifi card because she wanted wifi, without knowing that her laptop already had wifi. The problem? A common one to laptops: the kill switch was on. The fun doesn't stop there -- Toshiba value add (more like value subtract) was failing to find the device. A troubleshooting dialog suggests removing and re-adding the device in Device Manager. Fun times, but it worked (?!). Once I finally figure out that nightmare of XP wireless support, it's still not connecting to the shitty 2wire here. No suggestions, just an infinite waiting dialog. We haven't even gotten to her real problem (something online related), since neither wifi device is connecting to 2wire. Calling Microsoft here would be ridiculous, and ATT would be nightmarish. If this were Ubuntu, I'd have a much better shot at debugging this. I could see whether the driver loaded, whether the interface was up, check the logs for wifi status, etc. I also doubt this problem would have happened in the first place with Ubuntu, as MY Toshiba runs fine on this network (hurray Centrino). It could also be the 2wire, who the fuck knows, but we don't have time to debug this, let alone wait on hold.
Here's the crazy thing: thirty years ago, she was an RPG programmer for the State of Kansas, and has basically owned a computer since the Commodore. Lets stop pretending the grandmother test is perfect comprehension / acceptance or failure. What we need isn't a 1-800 line full of failing voice recognition software and foreign accents, we need software and hardware that works right the first time, and the way we expect.
On the lighter side, you could say that you'd only consider software that was "out of beta" or version 1.0 or greater, but that would leave out most open source projects and commercial "Web 2.0" products....
Then restrain yourself to "what Fedora ships" or "what Canonical supports in main". These are the presumably viable software products with a living upstream.
But you missed an interesting problem: failed commercial programs sometimes convert into open source projects. Its not clear to me whether this is a positive or negative effect. Are there more s out there or Blenders? Is the OpenOffice.org software good or bad?
No, the original intent of patents was to fill the Royal coffers by selling limited monopolies to merchants after the Magna Carta required democratic consent to levy taxes. The system was adopted in the US as a way to encourage people to bring ideas to an underdeveloped US with plenty of resources but not many highly educated inventors to employ those resources. When the founders wrote the constitution there were no "big companies". Eben Moglen had a pretty good talk on the origins of patents and the implications on software patents, but I've lost the links.
I'd really like to see a constitutional challenge against patents issued today based on the fact that they allow many patents that do not "advance the Science and useful Arts". The maximal utility of a patent from the filer's perspective is to disclose as little useful information as possible while broadly claiming as much as possible under the purview of their granted limited monopoly.
Shuttleworth has already stated he'd move Ubuntu scheduling to accommodate a release schedule. The puerile part is expecting other distros to walk into the trap of competing directly with a blockbuster release.
And there's already people out there testing ways around the system. Spelling out words in hints. The ESP game had a lot of wierd shit like people guessing the same sequence of words. I wonder if they've measured some sort of meme propagation for collusion like that.
The author has an excellent Google Tech Talk where he describes exactly how these games help computers get smarter. Basically, language is a shared set of common knowledge that a single researcher cannot accurately recreate without interviewing everyone. The games are tools to do exactly that. They generate datasets for analysis, and for further game playing. For example, you might find yourself describing the word "preserve" and start with marking it the opposite of destroy, while the partner guesses "strawberry" and "raspberry".
He's got lots of neat results from that system in the talk. Go watch it.
Well, judging by the comments this is how Fedora rolls. The bug you state is easily the most duplicated one in Ubuntu 8.04. I was curious if Fedora fixed that before release. From your original post I may presume they did not?
On the contrary when I try to fix bugs in Ubuntu I often find Debian Developers who refuse patches to their packages if they aren't in upstream. This is in spite of the upstream lead author saying things, when asked about a similar bug at a conference in front of a video, "Sure it's ugly, but until you the other guy does the work to change broken software we interact with, we'll do it this way. If you don't like it, fix it the right way yourself." Of course there's also people who don't follow that philosophy as hard; perhaps they feel upstream should focus on long term fixes instead of short term kludges, and so they patch locally but don't push for adoption upstream.
What I think needs to happen is perhaps realize that distributions do a different kind of work than upstream authors, and tools that allow, perhaps even encourage this is handy. Distributed source systems like git and bzr are great for tracking patches applied in Ubuntu, Debian, Redhat and so-on.
Developing software is a fixed cost. You only have to do it once per card. Fixed costs shouldn't affect the price of the device -- only if it happens at all. Price should be determined by how to maximize the term (Sale Price - Marginal Costs) * (# of Sales). Marginal costs are everything required to produce one more card (so unless quadro drivers are a larger binary, the driver is not marginally more expensive than a geforce). This works under the assumption, of course, that you can only charge one price for a device. Number of sales varies with price, so if you can charge people the people who'd pay $500 for it $500, and the people who'd pay $100 for it $100 , then your company "there to make money" could conceivably charge more to those high dollar customers by checking the PCI ID.
It's a shitty practice of course, since it relies on information asymmetry. nVidia knows more about their offerings than you or I do, and they can use this to their advantage.
And as usual, Apple is somewhere between just as bad or worse as MS. They do have undocumented APIs, which you're not supposed to use because of things like that. Of course they can use them, since they can fix it before the new library is released. Plus it gives them a competitive advantage over third parties.
From what I've heard from developers who've left Mac behind, though, is that last straw was a user base that's impossible to deal with. Being called an idiot because someone disagrees with your UI, or because your software doesn't support the latest OS immediately is not fun, apparently.
It's probably not that great on its own for able bodied people, but as one of many inputs it could be quite interesting. It could be used to subtly alter camera views towards what you find interesting, I suppose. This sort of stuff is already in use to generate "heat maps" of where people are looking when using a computer. Like where on a webpage you look, or in game, etc. It'd be useful to know what sort of things people don't care about, or how distracting something fading in / out might be to the task at hand.
It's also usable with dasher as a fairly fast text input for the disabled.
As a binary distro, Ubuntu refuses to fix bugs in other people's binaries. So compiling your own kernel is basically asking to no longer be supported, as it can affect just about everything. Compiling your own kernel packages is a huge pita on Ubuntu, a maze of twisty passages if you will. A basic kernel package is simple, but I've yet to figure out how to repackage the restricted driver packages. On an unrelated note, I don't think that many developers actually touch the forum and prefer IRC for rapid conversations and mailing lists for things that don't require immediate response. The forums simultaneously raise the noise level and require some amount of baby sitting.
Going back to the problem, I've heard that the kernel team is looking at releasing a new kernel to fix it. Unfortunately, random application problems like FF3b5 like to tank on disk I/O making it difficult to diagnose sometimes.
Clearly they should change the name and promote it as "it works 100 percent of the time, with everything, or double your money back."
Look, they're not trying to make a popular distro. They're trying to make a Linux distro that combines best of breed software from Ubuntu/Debian, while following a set of ethics. They don't want users who cry out for newer nvidia binaries, or fixing ndiswrapper breakage. In a sense, they're recognizing that many people feel FSF software ethics is a nuisance, that may call for boycotting software and hardware that works because it limits you. You don't need a distribution to convince people the ethics are important (the FSF seems to have demonstrated this), but you do need a distro for people convinced the ethics are important.
Making it popular can be good, but it would be stupid to make it popular at the expense of public understanding. Free Software is not necessarily less buggy or otherwise higher quality than open software, and in many cases you are making a sacrifice in the name of progress. Misleading people about the sacrifice or ignoring it totally is stupid.
I'm not a really good goto guy for improving/personalizing IM clients. I mostly just minimize the thing and wait for others to IM me. IRC is where it's at! Look at the number of irssi plugins and scripts versus adium;)
One of the new features in GNOME that I like is GIO. Basically gnome VFS gets replaced with FUSE, so you can launch mplayer while browsing smb shares etc. Apps no longer have to support GNOME VFS to integrate with nautilus, which is great.
GnomeDo mimics Quicksilver substantially, though still a work in progress. Apparently I missed the fact that it's in Ubuntu repos now. Remember that Ubuntu is supposed to be usable out of the box, it's up to you to make it awesome;)
I know this story's about Pidgin, but there's always Pidgin and Gnome-do. Adium's basically an OSX frontend to the pidgin libs anyways. There's a ton more apps besides Pidgin itself that uses libpurple. Most build for both OSX and Linux, so you can try before you leap;)
I don't know what world you live in, but Intel sells far more video chips than ATi and nvidia did combined. They were losing to Intel before, and now that intel is opening their drivers, they're still losing. If you view this loss as a nessecary evil resulting from pursuing high margin hardware, it's easy to attribute the open sourcing as opening the source to a low quality driver to a low feature chip. This mindset is already consistent with their support of Linux and production of fairly sophisticated 3d drivers for Linux.
Moreover, there's two sides to the appeal of source code. A small set of consumers may make decisions based on Linux support or open drivers, but an even smaller set of companies may be looking at purchasing nvidia. The ATi announcement came after CEO's declaring they'd never open the code. It took AMD buying ATi before that position could change, and the reason is simple: AMD engineers being able to evaluate the code could become a negotiating point where AMD says "look, we need to invest in making the drivers better". nVidia might be better off from that standpoint by pointing to their record of quality Linux binaries than a line by line review of the source. From a practical standpoint, it might even be a lose for potential buyers, who would probably need to review the GPL'd code before a purchase as a matter of due dilligence.
But honestly, I think the real reasons involve patent disputes and perhaps a bit of ugly runtime optimizations that they'd rather nobody knew about. I don't believe that nouveau has had any legal troubles, and if nouveau delivers the way you claim open source delivers, then at some point they can simply drop all that rhetoric and claim higher performance etc. They may have trouble obsoleting hardware like they've been doing the past few years with their "universal driver" system that keeps dropping older hardware, but that's just a reason to reinvest in GPU engineering.
Canonical sells support contracts to people and organizations. Places like system76 and Dell pay Canonical to provide engineering support for their systems. If Dell wants fingerprint reader device support in Ubuntu, Canonical will place an Ubuntu engineer on it. It still goes through the same software vetting and community process. Companies and individuals can also buy support contracts from Canonical, for things like requesting a bug fix or help configuring software (I guess). These are the people who might pay for Landscape, but I've never seen or used it.
In addition, I think Canonical recently landed a contract with Intel to port Ubuntu to their new Low Power Intel Architecture. This is where Ubuntu Mobile came from. They also appear to be working on designing a laptop or MID device themselves, to compete with Nokia and OQO and Eee. Their hiring page suggests this much at least.
Finally, I think they sell hosting services to companies like Opera that want to distribute software . Or perhaps they get a small fee for each install the way that google toolbar used to etc.
Remember that Ubuntu was started with two goals in mind: a proof of concept distribution for changes to Debian, and a proof of concept for their bug tracker software. Upstaging Debian itself wasn't planned on -- it took a lot of good timing and smart hires to make that happen. Fortunately those smart hires quickly set in place social protocols that set itself apart from Debian's perceived Internet cowboy theatrics. In many places I think Debian is slowly integrating the changes and concepts that Ubuntu tried out. Even if Canonical ultimately fails, it will easily have spawned the most successful Debian fork, and have been the most successful at providing, pushing and provoking change to Debian itself.
The problem is, the nVidias of the world are already losing to the Intels of the world. And they're firm in their belief that Intel is winning laptops because they're a low margin, low feature product. They're right, sadly. nVidia knows there's money to be made, but don't think sharing their driver code would make people pick them. It might lead to discovering a few interesting ideas in their drivers that propgate to other drivers, but more likely is a large set of bugs will be exposed and all those patent owners they don't pay will come looking to feed.
Slashdot Mirror
I've been recommending people change flamebait to +5. Flamebait is code for "goes against herd thinking". The fact that I can change the moderation levels of tags basically means you have to accept that people do filter and want it. +5 Flamebait of course relies on herd thinking to work, so I in fact think the defaults are acceptable.
Just put in a job description for HR competent enough to hire engineers, with reasonable pay. Let them anguish over the advertised salary numbers. Or put in your own resume to make HR work for the company. Until your HR understands what it is your engineers do and what to look for, they're not serving the company correctly.
That's because it shouldn't be that way. The problems an engineer faces in business isn't mathematical, but social. An engineering team shouldn't have to temper what he tells his boss based on whether the boss will overstate his role in success or understate his role in failure. They shouldn't have to worry about whether they'll be passed over for promotion because they refused to sign off on the new design citing safety test failures. I don't think business school is all that hard, and judging by the parties the business majors next door hold, I don't think they've ever dedicated a weekend to working on a school project.
We need big firms that understand engineering, not engineers that understand business. The latter creates large numbers of contractors and small companies, while the former can take advantage of economies of scale.
Of course the second one was flagged, there was a fucking #ifdef PURIFY right next to it. An entropy pool basically has randomness added to it. The problem is that the destination is one of the operands in every case, so there's no initial starting value by design. Static analysis tools decide that adding a random number to an uninitialized field doesn't make the field initialized. So the pool is never initialized since that would remove some magic entropy that probably doesn't exist but certainly doesn't hurt.
Static analysis tools are common in the open source world. The lint name is well known enough that many projects make theirs a pun on it, ala lintian. A few years ago a local root exploit in X was discovered by running these sorts of checks. But generally, static analysis tools require human review -- with large code bases they generate large numbers of false positives, especially the dumber ones. This leads to trouble for perfectionists, a common trait among software developers interested in bug fix analysis. For example, the recent massive Debian vulnerability was caused by an overzealous developer trying to fix static analysis flags. One of these flags was valid, one was not, and removing both removed nearly all entropy from the RNG.
In the more general sense, static analysis cannot find all bugs. There's a trivial proof: a program stuck in an infinite loop is a bug, but finding all such loops would solve the halting problem. Handling interrupts and the like also causes reasoning problems, as it's very hard, if not computationally intractable, to prove multi-threaded software is safe. So static analysis won't rid the embedded world of watchdog timers and other software failure recovery crap.
Her laptop cost 400 dollars. You can't buy anything Mac for that, let alone a laptop.
But here's a surprising bigger problem that I've seen a couple times now: for the computer literate elderly, learning a new system might not have a positive ROI. One of my alma mater's retired faculty ran DOS until something like last year, and one reason I've heard is that he didn't think he could port the software he wrote to a new OS before he died. Apparently someone convinced him that Linux fit the bill, perhaps it was DOSbox that set him Free. The same reasoning sadly can apply in the grandmother case. It's a bit weird as a young man to see people apparently planning to die, but it's hard to deny there's a logic to it.
Windows and computers don't pass the grandmother test. I should know as she just visited today and very nearly the first topic was that her computer was broken and I should look at it later. The conversation goes along on the couch about the new house and the new internet service (ATT u-verse) my mother bought. Grandparents mention a horrible experience switching ISP providers. They switched from DSL to cable a few years back, and decided to switch back yet again. They didn't realize it at the time, but they were a bit screwed -- nobody knew the old password or how to reset it. 1-800 wasn't very helpful here and quite frustrating. Eventually a neighbor who worked for ATT came along and figured it out for em. "Never again will we change ISPs!" she says.
Later on we get to looking at that laptop. First thing I notice is it's quite new, still with stickers on it, yet there's a wifi pcmcia card. I check underneath and clearly see a Wireless MAC [random mac address]. Apparently she had bought a wifi card because she wanted wifi, without knowing that her laptop already had wifi. The problem? A common one to laptops: the kill switch was on. The fun doesn't stop there -- Toshiba value add (more like value subtract) was failing to find the device. A troubleshooting dialog suggests removing and re-adding the device in Device Manager. Fun times, but it worked (?!). Once I finally figure out that nightmare of XP wireless support, it's still not connecting to the shitty 2wire here. No suggestions, just an infinite waiting dialog. We haven't even gotten to her real problem (something online related), since neither wifi device is connecting to 2wire. Calling Microsoft here would be ridiculous, and ATT would be nightmarish. If this were Ubuntu, I'd have a much better shot at debugging this. I could see whether the driver loaded, whether the interface was up, check the logs for wifi status, etc. I also doubt this problem would have happened in the first place with Ubuntu, as MY Toshiba runs fine on this network (hurray Centrino). It could also be the 2wire, who the fuck knows, but we don't have time to debug this, let alone wait on hold.
Here's the crazy thing: thirty years ago, she was an RPG programmer for the State of Kansas, and has basically owned a computer since the Commodore. Lets stop pretending the grandmother test is perfect comprehension / acceptance or failure. What we need isn't a 1-800 line full of failing voice recognition software and foreign accents, we need software and hardware that works right the first time, and the way we expect.
Then restrain yourself to "what Fedora ships" or "what Canonical supports in main". These are the presumably viable software products with a living upstream.
But you missed an interesting problem: failed commercial programs sometimes convert into open source projects. Its not clear to me whether this is a positive or negative effect. Are there more s out there or Blenders? Is the OpenOffice.org software good or bad?
I believe the novelty is that the button is designed to have a range of inputs, yet still have a "click" on the high end.
No, the original intent of patents was to fill the Royal coffers by selling limited monopolies to merchants after the Magna Carta required democratic consent to levy taxes. The system was adopted in the US as a way to encourage people to bring ideas to an underdeveloped US with plenty of resources but not many highly educated inventors to employ those resources. When the founders wrote the constitution there were no "big companies". Eben Moglen had a pretty good talk on the origins of patents and the implications on software patents, but I've lost the links.
I'd really like to see a constitutional challenge against patents issued today based on the fact that they allow many patents that do not "advance the Science and useful Arts". The maximal utility of a patent from the filer's perspective is to disclose as little useful information as possible while broadly claiming as much as possible under the purview of their granted limited monopoly.
Shuttleworth has already stated he'd move Ubuntu scheduling to accommodate a release schedule. The puerile part is expecting other distros to walk into the trap of competing directly with a blockbuster release.
And there's already people out there testing ways around the system. Spelling out words in hints. The ESP game had a lot of wierd shit like people guessing the same sequence of words. I wonder if they've measured some sort of meme propagation for collusion like that.
The author has an excellent Google Tech Talk where he describes exactly how these games help computers get smarter. Basically, language is a shared set of common knowledge that a single researcher cannot accurately recreate without interviewing everyone. The games are tools to do exactly that. They generate datasets for analysis, and for further game playing. For example, you might find yourself describing the word "preserve" and start with marking it the opposite of destroy, while the partner guesses "strawberry" and "raspberry".
He's got lots of neat results from that system in the talk. Go watch it.
Well, judging by the comments this is how Fedora rolls. The bug you state is easily the most duplicated one in Ubuntu 8.04. I was curious if Fedora fixed that before release. From your original post I may presume they did not?
On the contrary when I try to fix bugs in Ubuntu I often find Debian Developers who refuse patches to their packages if they aren't in upstream. This is in spite of the upstream lead author saying things, when asked about a similar bug at a conference in front of a video, "Sure it's ugly, but until you the other guy does the work to change broken software we interact with, we'll do it this way. If you don't like it, fix it the right way yourself." Of course there's also people who don't follow that philosophy as hard; perhaps they feel upstream should focus on long term fixes instead of short term kludges, and so they patch locally but don't push for adoption upstream.
What I think needs to happen is perhaps realize that distributions do a different kind of work than upstream authors, and tools that allow, perhaps even encourage this is handy. Distributed source systems like git and bzr are great for tracking patches applied in Ubuntu, Debian, Redhat and so-on.
Developing software is a fixed cost. You only have to do it once per card. Fixed costs shouldn't affect the price of the device -- only if it happens at all. Price should be determined by how to maximize the term (Sale Price - Marginal Costs) * (# of Sales). Marginal costs are everything required to produce one more card (so unless quadro drivers are a larger binary, the driver is not marginally more expensive than a geforce). This works under the assumption, of course, that you can only charge one price for a device. Number of sales varies with price, so if you can charge people the people who'd pay $500 for it $500, and the people who'd pay $100 for it $100 , then your company "there to make money" could conceivably charge more to those high dollar customers by checking the PCI ID.
It's a shitty practice of course, since it relies on information asymmetry. nVidia knows more about their offerings than you or I do, and they can use this to their advantage.
And as usual, Apple is somewhere between just as bad or worse as MS. They do have undocumented APIs, which you're not supposed to use because of things like that. Of course they can use them, since they can fix it before the new library is released. Plus it gives them a competitive advantage over third parties.
From what I've heard from developers who've left Mac behind, though, is that last straw was a user base that's impossible to deal with. Being called an idiot because someone disagrees with your UI, or because your software doesn't support the latest OS immediately is not fun, apparently.
It's probably not that great on its own for able bodied people, but as one of many inputs it could be quite interesting. It could be used to subtly alter camera views towards what you find interesting, I suppose. This sort of stuff is already in use to generate "heat maps" of where people are looking when using a computer. Like where on a webpage you look, or in game, etc. It'd be useful to know what sort of things people don't care about, or how distracting something fading in / out might be to the task at hand.
It's also usable with dasher as a fairly fast text input for the disabled.
As a binary distro, Ubuntu refuses to fix bugs in other people's binaries. So compiling your own kernel is basically asking to no longer be supported, as it can affect just about everything. Compiling your own kernel packages is a huge pita on Ubuntu, a maze of twisty passages if you will. A basic kernel package is simple, but I've yet to figure out how to repackage the restricted driver packages. On an unrelated note, I don't think that many developers actually touch the forum and prefer IRC for rapid conversations and mailing lists for things that don't require immediate response. The forums simultaneously raise the noise level and require some amount of baby sitting.
Going back to the problem, I've heard that the kernel team is looking at releasing a new kernel to fix it. Unfortunately, random application problems like FF3b5 like to tank on disk I/O making it difficult to diagnose sometimes.
Clearly they should change the name and promote it as "it works 100 percent of the time, with everything, or double your money back."
Look, they're not trying to make a popular distro. They're trying to make a Linux distro that combines best of breed software from Ubuntu/Debian, while following a set of ethics. They don't want users who cry out for newer nvidia binaries, or fixing ndiswrapper breakage. In a sense, they're recognizing that many people feel FSF software ethics is a nuisance, that may call for boycotting software and hardware that works because it limits you. You don't need a distribution to convince people the ethics are important (the FSF seems to have demonstrated this), but you do need a distro for people convinced the ethics are important.
Making it popular can be good, but it would be stupid to make it popular at the expense of public understanding. Free Software is not necessarily less buggy or otherwise higher quality than open software, and in many cases you are making a sacrifice in the name of progress. Misleading people about the sacrifice or ignoring it totally is stupid.
I'm not a really good goto guy for improving/personalizing IM clients. I mostly just minimize the thing and wait for others to IM me. IRC is where it's at! Look at the number of irssi plugins and scripts versus adium ;)
;)
One of the new features in GNOME that I like is GIO. Basically gnome VFS gets replaced with FUSE, so you can launch mplayer while browsing smb shares etc. Apps no longer have to support GNOME VFS to integrate with nautilus, which is great.
GnomeDo mimics Quicksilver substantially, though still a work in progress. Apparently I missed the fact that it's in Ubuntu repos now. Remember that Ubuntu is supposed to be usable out of the box, it's up to you to make it awesome
I know this story's about Pidgin, but there's always Pidgin and Gnome-do. Adium's basically an OSX frontend to the pidgin libs anyways. There's a ton more apps besides Pidgin itself that uses libpurple. Most build for both OSX and Linux, so you can try before you leap ;)
I don't know what world you live in, but Intel sells far more video chips than ATi and nvidia did combined. They were losing to Intel before, and now that intel is opening their drivers, they're still losing. If you view this loss as a nessecary evil resulting from pursuing high margin hardware, it's easy to attribute the open sourcing as opening the source to a low quality driver to a low feature chip. This mindset is already consistent with their support of Linux and production of fairly sophisticated 3d drivers for Linux.
Moreover, there's two sides to the appeal of source code. A small set of consumers may make decisions based on Linux support or open drivers, but an even smaller set of companies may be looking at purchasing nvidia. The ATi announcement came after CEO's declaring they'd never open the code. It took AMD buying ATi before that position could change, and the reason is simple: AMD engineers being able to evaluate the code could become a negotiating point where AMD says "look, we need to invest in making the drivers better". nVidia might be better off from that standpoint by pointing to their record of quality Linux binaries than a line by line review of the source. From a practical standpoint, it might even be a lose for potential buyers, who would probably need to review the GPL'd code before a purchase as a matter of due dilligence.
But honestly, I think the real reasons involve patent disputes and perhaps a bit of ugly runtime optimizations that they'd rather nobody knew about. I don't believe that nouveau has had any legal troubles, and if nouveau delivers the way you claim open source delivers, then at some point they can simply drop all that rhetoric and claim higher performance etc. They may have trouble obsoleting hardware like they've been doing the past few years with their "universal driver" system that keeps dropping older hardware, but that's just a reason to reinvest in GPU engineering.
Canonical makes their money any way they can.
Canonical sells support contracts to people and organizations. Places like system76 and Dell pay Canonical to provide engineering support for their systems. If Dell wants fingerprint reader device support in Ubuntu, Canonical will place an Ubuntu engineer on it. It still goes through the same software vetting and community process. Companies and individuals can also buy support contracts from Canonical, for things like requesting a bug fix or help configuring software (I guess). These are the people who might pay for Landscape, but I've never seen or used it.
In addition, I think Canonical recently landed a contract with Intel to port Ubuntu to their new Low Power Intel Architecture. This is where Ubuntu Mobile came from. They also appear to be working on designing a laptop or MID device themselves, to compete with Nokia and OQO and Eee. Their hiring page suggests this much at least.
Finally, I think they sell hosting services to companies like Opera that want to distribute software . Or perhaps they get a small fee for each install the way that google toolbar used to etc.
Remember that Ubuntu was started with two goals in mind: a proof of concept distribution for changes to Debian, and a proof of concept for their bug tracker software. Upstaging Debian itself wasn't planned on -- it took a lot of good timing and smart hires to make that happen. Fortunately those smart hires quickly set in place social protocols that set itself apart from Debian's perceived Internet cowboy theatrics. In many places I think Debian is slowly integrating the changes and concepts that Ubuntu tried out. Even if Canonical ultimately fails, it will easily have spawned the most successful Debian fork, and have been the most successful at providing, pushing and provoking change to Debian itself.
The problem is, the nVidias of the world are already losing to the Intels of the world. And they're firm in their belief that Intel is winning laptops because they're a low margin, low feature product. They're right, sadly. nVidia knows there's money to be made, but don't think sharing their driver code would make people pick them. It might lead to discovering a few interesting ideas in their drivers that propgate to other drivers, but more likely is a large set of bugs will be exposed and all those patent owners they don't pay will come looking to feed.