Why doesn't Kubuntu run on my PC?
After the old KFN DB got corrupted some posts were lost, and were not moved to this new forum DB. I located a Google cache of this post made a year ago. I post it hear to explain why some machines which used to work well with Kubuntu (or any distro for that matter) no longer work as well, or at all.
(lost link to rant complaining about their ATI Radeon 9600 video chip no longer working with Linux)
Three computer generations ago (circa 2000 and before) the meaning of "release early, release often" was well known mantra among members of the Linux community. In a quid quo pro for free software community members gladly accepted early release software, knowing that for some of them the earlier the version number the more likely it would break their system, and they did their part by posting bug reports.
As time passed more and more people, unaware of the mantra or its meaning, came into the Linux community and rather than posting bug reports at the appropriate websites they often would post increasingly shrill rants at forums like this. The amazing thing is that many of those complaining came from Microsoft platforms, where they paid good money, and many times a large amount of it, for software that was no better than early version Linux software.
To counter the "release early, release often" component of the Bazaar Development Model (http://en.wikipedia.org/wiki/The_Cat...and_the_Bazaar), Microsoft introduced the idea of releasing software as "beta" releases, except that customers still had to pay for it. It reduced Microsoft's accountability while still generating revenue. That people would pay for "beta" software still amazes me.
Raymond's central proposition is that
"given enough eyeballs, all bugs are shallow"
(which he terms Linus' Law):
the more widely available the source code is for public testing, scrutiny, and experimentation, the more rapidly all forms of bugs will be discovered.
In contrast, Raymond claims that an inordinate amount of time and energy must be spent hunting for bugs in the Cathedral model, since the working version of the code is available only to a few developers.
Since that thesis was written developers in both camps have devised unit testing procedures and applications that significantly reduce the number of program defects, BUT, and it is a big BUT, the code has to be tested against a HUGE variety of hardware versions. A version of Kubuntu may work great on a particular HP model but the next release of Kubuntu may have problems because of a variety of reasons. The biggest reason is the there is a sliding window of compatibility in which most computers will run well with a particular distro release or range of releases. Computer models released after a distro is released may not work as well as previous models because Linux kernel and application developers don't yet have access to machines on which they can reverse engineer code. The Intel Mobil 4 Serious Express chipset in this 2008 Sony Notebook did not work well with Mandriva 2009 PP release. By February of 2009, the fix had filtered down to the distro developers and Kubuntu 9.04 Alpha gave this notebook excellent acceleration. Since Feb of 2009 Kubuntu has worked flawlessly on this box, as does Natty work on my relative's 2008 HP dv5. However, I harbor no illusions. And, as computers age, they may slide out of the compatibility window. Sometime in the future one of two things will happen, either this notebook will have a hardware failure which will be cheaper to fix by replacing it, or, its components compatibility will slide out of the compatibility window as drivers for its chips are moved to the legacy repository.
What causes the compatibility window? Frankly, it is a direct result of Microsoft's still functioning monopoly tilting the playing field. (As long as Microsoft can dictate to PC OEMs what OS they can install on their products an illegal monopoly exists, even if that dictation is powered by ad rebates.) There was a time when chip and peripheral makers products were OS neutral. The specs were public and any OS could write drivers to run them. That was pretty much the situation when I began using Linux in May of 1998. By the time Win95 came out Microsoft had convinced peripheral makers to remove their onboard CPUs (and pocket the savings as profit without reducing price) and rely on the computer's motherboard CPU, which they controlled access to because they had also convinced ($$$) PC OEMs to pre-install Win95 on many (and later ALL) of their computers. These devices became known as WinModems, WinPrinters, etc. because they only worked "out of the box" with Windows. To make matters WORSE, Microsoft also convinced ($$$) them to NOT publish the specifications to their devices, even though Microsoft was given the complete specifications. Linux, BEOS, Unix and other OS makers had to reverse engineer the software to run against the hardware, which took time. So, while Win98 or Win2K or Win7 can run a peripheral "out of the box", Linux users of computers and peripherals sometimes have to wait a bit longer before their favorite distro works well. Sometimes, some features never work as well on Linux as they do on Windows. This is especially true of niche products like my Watcom tablet, because demand is not great and the number of developers and users for such products is low. Go to Watcom's website and search for Linux and you will get no replies. SourceForge has a Linux driver for the Watcom Bamboo tablet. Installing it involves doing about 5 steps each time your kernel is upgraded, and requires you have some developer tools installed on your distro. (Note: As of Kubuntu Precise, the Wacom Bamboo tablet is recognized and configured, 3 1/2 years after I bought my Sony VIAO.)
To make matters even more difficult for Linux developers and users, PC OEMs often change the chip specifications within a given model as new machines are produced. My relative purchased a Gateway m675prr laptop. A year (2005) later I purchased the SAME model laptop. However, many of his chip drivers had "VER 1.0A" but mine showed "VER 1.1B" or some similar change. There was only a short period of about a year in which both of our laptops could run the same distro "out of the box". Mine was picky at first, and his became picky later. One example of why it was picky is the wireless chip, a Broadcom 4306. At first his wouldn't run, but by the time I got my laptop both would configure without problems, using ndiswrapper and the bcm5wl.sys and cat files. Later, bcm43-fwcutter was included in most distros and automatically picked the firmware out of the chip and created a driver, which installed automatically. Time moved on. Other wireless chips were released, reverse engineered, and added to the distros CDs to enable wireless chips on newer PCs. But, there is a limit to how much can go on a 700 MB CD. To make room for newer chip drivers older chip drivers were moved to the legacy archives. Some of the less common 43** Broadcom chips never had their drivers included on a CD and owners of those chips have always had to obtain their drivers from sources like SourceForge or the best source of all: http://linuxwireless.org/ (http://linuxwireless.org/)
Microsoft continued to throw roadblocks at Linux. When it became obvious that keeping specification sheets out of the hands of Linux developers didn't have the desired affect Microsoft and its proprietary allies lobbied ($$$) for DRM and other Constitution busting laws to prevent even reverse engineering. Those laws haven't had the effect they'd hoped for because Linux users in the US represent only a small fraction of the total number of Linux users in countries around the world, most of which have no obligation to enforce US laws. Then Microsoft tried a direct frontal attack, funding SCO to sue Linux by using IBM as a proxy. The goal was to hijack ownership of Linux by claiming that Linux was due almost in total to theft of SCO's Unix code. The judge didn't buy it and SCO lost. At about the same time Microsoft made vague threats about 235 patents which Linux supposedly "infringed". It proceeded to "negotiate" IP licenses with small fry corporations whose products incorporated Linux and used some old MS technology, like FAT. In 2006 Novell added smoke and mirrors to the issue by "admitting" (in exchange for $340M US upfront and installments of several tens of millions in subsequent years) that Linux uses MS proprietary IP and Novell agreed to pay Microsoft a ROYALTY for each SUSE Enterprise Linux Server CD it sold. (That's like me "admitting" that you are guilty of murder.) They then agreed to bifurcate the GPL by claiming that those who submitted GPL code to openSUSE and that code was included in SELS were "free" from threat of a Microsoft lawsuit but ALL other GPL developers contributing to openSUSE and ANY OTHER distro were under threat of Microsoft lawsuit for "IP infringement". That claim is a violation of the terms of the GPL and made distributing openSUSE or SUSE illegal, but no one who could afford to take them to court pushed it. The Novell agreement was to last 5 years. Both it and Novell's existence ended. Attachemate purchased Novell and laid off hundreds of developers, including de Icaza's MONO team. Attempts by Attachemate to transfer certain IP to Microsoft was blocked by the courts. (Note: Novell's "partnership" with Microsoft proved fatal to Novell, which no longer exists.) Now, Microsoft is pushing EUFI on OEM PCs. It's sole effect is to lock Linux out of dual booting with or replacing Windows 8.
I've given only a brief history of the difficulties of development of software under the GPL, the ONLY reason why free and open software exists at all. And, I have mentioned only a few of the things Microsoft has done to try and kill the Linux and the GPL.
So, as a retired software developer I want to thank everyone who has become part of the Linux community and contributed to its ongoing existence and welfare by contributing bug reports to the appropriate websites. (If you think you've found a bug and don't know where to report it ask here!), and encouragement to those who have experienced problems, for what ever reason. This is what builds Linux up and guarantees its continued existence. Childish ranting and overblown accusations or conclusions (if it doesn't work for you it must not work for anyone else) do not contribute to the GPL community. The pun "The beatings will continue until morale improves" makes a joke of the obvious, but the lesson should still be obvious.
Three computer generations ago (circa 2000 and before) the meaning of "release early, release often" was well known mantra among members of the Linux community. In a quid quo pro for free software community members gladly accepted early release software, knowing that for some of them the earlier the version number the more likely it would break their system, and they did their part by posting bug reports.
As time passed more and more people, unaware of the mantra or its meaning, came into the Linux community and rather than posting bug reports at the appropriate websites they often would post increasingly shrill rants at forums like this. The amazing thing is that many of those complaining came from Microsoft platforms, where they paid good money, and many times a large amount of it, for software that was no better than early version Linux software.
To counter the "release early, release often" component of the Bazaar Development Model (http://en.wikipedia.org/wiki/The_Cat...and_the_Bazaar), Microsoft introduced the idea of releasing software as "beta" releases, except that customers still had to pay for it. It reduced Microsoft's accountability while still generating revenue. That people would pay for "beta" software still amazes me.
Raymond's central proposition is that
"given enough eyeballs, all bugs are shallow"
(which he terms Linus' Law):
the more widely available the source code is for public testing, scrutiny, and experimentation, the more rapidly all forms of bugs will be discovered.
In contrast, Raymond claims that an inordinate amount of time and energy must be spent hunting for bugs in the Cathedral model, since the working version of the code is available only to a few developers.
Since that thesis was written developers in both camps have devised unit testing procedures and applications that significantly reduce the number of program defects, BUT, and it is a big BUT, the code has to be tested against a HUGE variety of hardware versions. A version of Kubuntu may work great on a particular HP model but the next release of Kubuntu may have problems because of a variety of reasons. The biggest reason is the there is a sliding window of compatibility in which most computers will run well with a particular distro release or range of releases. Computer models released after a distro is released may not work as well as previous models because Linux kernel and application developers don't yet have access to machines on which they can reverse engineer code. The Intel Mobil 4 Serious Express chipset in this 2008 Sony Notebook did not work well with Mandriva 2009 PP release. By February of 2009, the fix had filtered down to the distro developers and Kubuntu 9.04 Alpha gave this notebook excellent acceleration. Since Feb of 2009 Kubuntu has worked flawlessly on this box, as does Natty work on my relative's 2008 HP dv5. However, I harbor no illusions. And, as computers age, they may slide out of the compatibility window. Sometime in the future one of two things will happen, either this notebook will have a hardware failure which will be cheaper to fix by replacing it, or, its components compatibility will slide out of the compatibility window as drivers for its chips are moved to the legacy repository.
What causes the compatibility window? Frankly, it is a direct result of Microsoft's still functioning monopoly tilting the playing field. (As long as Microsoft can dictate to PC OEMs what OS they can install on their products an illegal monopoly exists, even if that dictation is powered by ad rebates.) There was a time when chip and peripheral makers products were OS neutral. The specs were public and any OS could write drivers to run them. That was pretty much the situation when I began using Linux in May of 1998. By the time Win95 came out Microsoft had convinced peripheral makers to remove their onboard CPUs (and pocket the savings as profit without reducing price) and rely on the computer's motherboard CPU, which they controlled access to because they had also convinced ($$$) PC OEMs to pre-install Win95 on many (and later ALL) of their computers. These devices became known as WinModems, WinPrinters, etc. because they only worked "out of the box" with Windows. To make matters WORSE, Microsoft also convinced ($$$) them to NOT publish the specifications to their devices, even though Microsoft was given the complete specifications. Linux, BEOS, Unix and other OS makers had to reverse engineer the software to run against the hardware, which took time. So, while Win98 or Win2K or Win7 can run a peripheral "out of the box", Linux users of computers and peripherals sometimes have to wait a bit longer before their favorite distro works well. Sometimes, some features never work as well on Linux as they do on Windows. This is especially true of niche products like my Watcom tablet, because demand is not great and the number of developers and users for such products is low. Go to Watcom's website and search for Linux and you will get no replies. SourceForge has a Linux driver for the Watcom Bamboo tablet. Installing it involves doing about 5 steps each time your kernel is upgraded, and requires you have some developer tools installed on your distro. (Note: As of Kubuntu Precise, the Wacom Bamboo tablet is recognized and configured, 3 1/2 years after I bought my Sony VIAO.)
To make matters even more difficult for Linux developers and users, PC OEMs often change the chip specifications within a given model as new machines are produced. My relative purchased a Gateway m675prr laptop. A year (2005) later I purchased the SAME model laptop. However, many of his chip drivers had "VER 1.0A" but mine showed "VER 1.1B" or some similar change. There was only a short period of about a year in which both of our laptops could run the same distro "out of the box". Mine was picky at first, and his became picky later. One example of why it was picky is the wireless chip, a Broadcom 4306. At first his wouldn't run, but by the time I got my laptop both would configure without problems, using ndiswrapper and the bcm5wl.sys and cat files. Later, bcm43-fwcutter was included in most distros and automatically picked the firmware out of the chip and created a driver, which installed automatically. Time moved on. Other wireless chips were released, reverse engineered, and added to the distros CDs to enable wireless chips on newer PCs. But, there is a limit to how much can go on a 700 MB CD. To make room for newer chip drivers older chip drivers were moved to the legacy archives. Some of the less common 43** Broadcom chips never had their drivers included on a CD and owners of those chips have always had to obtain their drivers from sources like SourceForge or the best source of all: http://linuxwireless.org/ (http://linuxwireless.org/)
Microsoft continued to throw roadblocks at Linux. When it became obvious that keeping specification sheets out of the hands of Linux developers didn't have the desired affect Microsoft and its proprietary allies lobbied ($$$) for DRM and other Constitution busting laws to prevent even reverse engineering. Those laws haven't had the effect they'd hoped for because Linux users in the US represent only a small fraction of the total number of Linux users in countries around the world, most of which have no obligation to enforce US laws. Then Microsoft tried a direct frontal attack, funding SCO to sue Linux by using IBM as a proxy. The goal was to hijack ownership of Linux by claiming that Linux was due almost in total to theft of SCO's Unix code. The judge didn't buy it and SCO lost. At about the same time Microsoft made vague threats about 235 patents which Linux supposedly "infringed". It proceeded to "negotiate" IP licenses with small fry corporations whose products incorporated Linux and used some old MS technology, like FAT. In 2006 Novell added smoke and mirrors to the issue by "admitting" (in exchange for $340M US upfront and installments of several tens of millions in subsequent years) that Linux uses MS proprietary IP and Novell agreed to pay Microsoft a ROYALTY for each SUSE Enterprise Linux Server CD it sold. (That's like me "admitting" that you are guilty of murder.) They then agreed to bifurcate the GPL by claiming that those who submitted GPL code to openSUSE and that code was included in SELS were "free" from threat of a Microsoft lawsuit but ALL other GPL developers contributing to openSUSE and ANY OTHER distro were under threat of Microsoft lawsuit for "IP infringement". That claim is a violation of the terms of the GPL and made distributing openSUSE or SUSE illegal, but no one who could afford to take them to court pushed it. The Novell agreement was to last 5 years. Both it and Novell's existence ended. Attachemate purchased Novell and laid off hundreds of developers, including de Icaza's MONO team. Attempts by Attachemate to transfer certain IP to Microsoft was blocked by the courts. (Note: Novell's "partnership" with Microsoft proved fatal to Novell, which no longer exists.) Now, Microsoft is pushing EUFI on OEM PCs. It's sole effect is to lock Linux out of dual booting with or replacing Windows 8.
I've given only a brief history of the difficulties of development of software under the GPL, the ONLY reason why free and open software exists at all. And, I have mentioned only a few of the things Microsoft has done to try and kill the Linux and the GPL.
So, as a retired software developer I want to thank everyone who has become part of the Linux community and contributed to its ongoing existence and welfare by contributing bug reports to the appropriate websites. (If you think you've found a bug and don't know where to report it ask here!), and encouragement to those who have experienced problems, for what ever reason. This is what builds Linux up and guarantees its continued existence. Childish ranting and overblown accusations or conclusions (if it doesn't work for you it must not work for anyone else) do not contribute to the GPL community. The pun "The beatings will continue until morale improves" makes a joke of the obvious, but the lesson should still be obvious.