Its a large number of writes not reads. Something like swap space gets allocated and even after it no longer reflects memory it stays empty for "a while". As long as "a while" is an hour or more on average you are fine.
No a mediocre SSD will crush a very expensive RAID-5 in read tests. If you go up to something like RAID-15 across 30 drives or so, yeah the raid will be better. SSD really is much faster otherwise no one would be all that excited about paying 20x the cost.
Thanks for that tip. I've been buying computer and printer memory from crucial but it was great to red the review on A&D and see how well their drives scored.
I'd agree with your 20x right now at say the 256 g - 500 g levels. But lets say when SSD is around 3t the spread might only be 10. At 20t the spread might only be 5. And maybe at 20t people start to switch in mass. This at first drives the spread back up close to 10 but within 2 years, the spread is like 2-3 with 35t SSD costing the same as 100t HDD. At that point only the most demanding users, those people who need huge amounts of cheap storage, exabytes worth still buy HDD. That's called "a retreat to quality". That small market kills their ability to get the innovations they need and so likely SSD becomes cheaper when it hits 200t or something.
Those are rough numbers but lets say around the time you are buying 1p SSD hard drive it will be cheaper than the HDD alternative.
If MSFT or Apple released updates every six months that broke at least a third of the hardware out there, wouldn't you deride them mercilessly? Of course you would. So why take that shit from your own OS?
Because it doesn't. Your facts are just wrong. Updates from OS version to OS version don't generally break anything. I don't see any evidence for Dell spending boatloads of money on anything Linux related, including the Poweredge. Suse, and RedHat do the suport work, Dell doesn't do much of anything. If they did then they could just take the data from Linux on Laptops and build their own port tree for a few thousand, not millions.
Linus changes to the kernels rarely break hardware. You simply recompile the kernel using the same static config, the way Fedora, RedHat, Mandriva, Debian.... do and the new stuff doesn't change much of anything. Ubuntu uses the Debian branch not Linus' branch (the vanilla branch) so his changes don't even affect them necessarily. You want a stable branch: Andi Kleen's, Greg Kroah-Hartman and Adrian Bunk maintain stable branches. Don't use vanilla.
And more importantly, if you like the more stable branches why use Ubuntu which doesn't support them? Frankly everything Ubruntu stands for is making Debian less stable. Another example of a very stable Linux is Cent.
Think about your argument in the last dozen posts, and ask yourself if this were true why Linux is so dominant on server, on supercomputing, on embedded. Your arguments aren't desktop specific anymore. Open Solaris had great binary compatibility I don't remember them being all that popular on the desktop.
And here we have the twin psychotic face of Linux thing. On the one hand time and time again (hell sites like/. are practically bursting with articles written by the community) we here you Linux advocates say "Linux is ready for the desktop!" yet when anyone points out the serious problems that give way to that lie you basically say "Oh its not an OS, its a community! Go back to Windows if you won't do things OUR way". So which is it? Can't eat your cake and have it too you know.
Well there is two things.
The first is that your only experience is with Windows desktop environments. There are millions of people all over the country using non Windows OSes right now on desktops. Companies like Burlington Coat Factory, Autozone, Pep Boys were able to switch easily, because they had a Unix not a Windows culture in their workforce. Education of HS or less so its not computer geeks. Millions of people still use mainframe OSes, and curses style interfaces. You simply lack experience as to what is else is possible. Your feeling is that if it is not done "the windows way" then its just wrong. The purpose of the GNU project was never to replace the NT kernel it was to change the computing culture. Having Linux do things the windows way, would subvert the broader goal. Having people use Windows while running ever increasing amounts of free software and become more concerned about open protocols (see the debate about net neutrality) advances the agenda.
Now as for the rest. It is ready for the desktop, and has been for a decade. Which is different then saying it is ready to overthrow Microsoft Windows. If Linux were the dominant desktop today, Windows would not be able to overthrow it either. The cost of switching between them is high in terms of the entire community of expectations. But for those people who have some reason to want to switch it is a fully viable desktop and has been for years. So I'm rejecting the equivalence you are making between an absolute worst case scenario for Linux and "the norm".
Assume for a moment Microsoft raised the OEM price of Windows to $1000, you don't think you would see computers shipping with high quality well maintained Linux distributions quickly? That's what ready means. That if Linux had to step in, it could step in. Now you made very specific claims about Dell which were easily refuted by looking at the history of the Power Edge for example. The point about the Power Edge is that with a motivated vendor long term driver updates were possible. Which I think should tell you that maybe you need to re-examine your assumptions about the way the Linux community works and the fact that it is able to support hardware when there is a genuine motivation. Which is different than your case where there isn't.
Richard Stallman, 1983 To begin with, GNU will be a kernel plus all the utilities needed to write and run C programs: editor, shell, C compiler, linker, assembler, and a few other things. After this we will add a text formatter, a YACC, an Empire game, a spreadsheet, and hundreds of other things. We hope to supply, eventually, everything useful that normally comes with a Unix system, and anything else useful, including on-line and hardcopy documentation.
The goal was to provide a Unix system. First and foremost Linux has to be a Unix and do things the Unix way. The goal was not to build a free Amiga. There is another project that aims to do that, Syllable and maybe what you should do is work with them.
And the reason your arguments much credit is I remember very much the same arguments being made in the late 1990s and early 2000s by big box Unixes. "Linux is fine for hobbyists but its never going to replace Solaris"...
I have yet to hear any reason from you would want to use Linux or for that matter any Unix.
a) You hate the GUIs b) You hate the apps c) You don't want to support non typical usage.
As for drivers, features like kernel patch protection. And yes, the NT mode that Cutler proposed is also closer to the Linux philosophy.
Your post is basically a rant of: "I hate Unix, I hate everything that's different between Unix and Windows. I have no need for any Unix features or applications". Well in that case use Windows. I don't like golf, I don't playing golf, I don't want a golf membership; ergo you aren't going to be able to sell me golf clubs regardless of what you do with them. Why would anyone expect your small hardware shop to support or offer Linux? Were you supporting SCO or OS/2 15 years ago?
Dell's PowerEdge Line has had solid Linux support for most of a decade now. Including excellent driver support and compatibility with most distributions. No serious problems remotely like what you are describing. Today, they ship with Suse Enterprise or RHEL with no need for specialized repository. The M Series seems to be doing well and the N series has existed off and on with no problems. Because those people want to run Unix applications they are willing to do things the Unix way. And there Dell gets excellent support.
But in terms of Dell not getting support they've gotten great support. I've been dealing with Dell's and Linux for 17 years, I can always find excellent information from the community on how to configure Dell's and deal with their quirks. That's good support. Now if you mean working out the box. That's Dell's job, they create an OEM version config file just like they do for Windows today.
As far as the Ubuntu line. Horrible garbage low end hardware for cheap machines.... I can see Linux going downmarket in the 3rd world but what 1st world customer wouldn't rather just pay $30-40 for an OEM version of windows 7? Netbooks were a missed opportunity. Cell phones are working out.
In terms of GUI features: * -- XML building so that the GUI can be complete reconfigured by end users (or OEMs). Apple sorta supports something like this, Microsoft doesn't allow it at all. * -- Entourage like window management * -- desktop grid: allows zooming in and out of virtual windows. * -- transparency and opacity
Those things didn't exist on GUI's in the 1990s at all. Apple introduced them last decade and Microsoft with Vista. So lets cut the non sense about Windows 9X. Try doing those things on 9X.
I think you need some realism. Linux is not a product its a community. You want people to work to service your needs, call Microsoft. You want to help build a culture of information sharing, then we have something to talk about.
a) The kernel abi is not binary stable. b) The operating systems rarely require a CLI interaction. c) Commercial applications are better, in particular easier to use.
In all 3 areas I think you are failing to take into account the difference in spend.
(a) has been a major area of debate. The advantages of stable binary drivers have been discussed ad nauseam, and that's because they really do exist. The problems of bad binary drivers corrupting systems have been discussed ad nauseam. I'm not going to get into the details but there really are tradeoffs. The general direction over the last 15 years has been for * Windows to move closer to the Linux way of doing things * Linux to move closer to the Windows way of doing things * Most vendors to just write their own and limit hardware
Microsoft's driver solution literally costs multiple billions of dollars per year. It is simply not an option of the Linux community in terms of manpower needed. If you need excellent hardware support across the full range of PC hardware, you have one option in terms of OSes. No one else is going to be comparable for a long time. Linux does an excellent job at less than 1/4000th of the cost. But yes 99.98% reduction is spending does have end user impact. If you have $2b a year for a driver lab I'm sure Linus would love the help and would be happy to do it Microsoft's way. The Linux solution has worked so well that the Linux kernel is the dominant embedded kernel because of the huge range of hardware it support. So I think you either being completely unrealistic about what is achievable and failing to appreciate how good a job the Linux kernel team has done in terms of hardware support given the tremendous differences in spend.
On (b) CLI bleed is a result of cost savings. Just to pick an example, Excel never implemented the GUI features of ideas like Jazz from Lotus because of the additional complexity; so Excel functions are CLI "sum(a7..b23)" with some level of GUI interface and not "sum(sales table)" like you had on Lotus products because at the time Excel wasn't making that kind of money. CLI is much cheaper than GUI, by about two orders of magnitude.
The first attempt to even build a real GUI for Linux didn't start until 1996. By 2001 or so there was a GUI, which still required some CLI, on par with say Windows 3.1 (so say about 7 years behind). Today's Linux GUIs are better but the commercial GUIs have massively improved over the last 10 years so Linux has lost ground. I definitely think you can argue Linux GUIs for the last 5 years are as good as Windows 2000 was and are approaching XP. So maybe they've lost a little ground and are more like 9 years behind but its been a very tough decade to keep pace. Further Windows 2000 was a premium product. I had to mess with config files on Windows 2000 quite a bit. Similarly I think you can compare Linux GUIs favorably to OSX 10.1 where people frequently did have to drop to terminal to get things to work. Those systems were usable by end users at the time, I think you are exaggerating.
If you point is that Linux GUIs are worse than those that represent tens of billions of dollars in GUI development, well yeah. But again I think you can say that Linux has far and away the best GUIs of any non mainstream operating system. I don't disagree that there is no particular reason an end user should want to downgrade their experience. I will comment those same end users often run software that is 5, 10, 15 years old, as you mentioned so perhaps they dislike change more than they like features and so it really doesn't matter very much.
In terms of apps., your point (c) I think its worth noting that widgets sets like QT and GTK are now the standard for most non platform specific software development so overtime you will see Linux applications improving in the commercial sector..NET and Cocoa are really really good so for applications makers willing to inc
I think you ignored the core of my comment. The goal of the GNU project was to have a free Unix that was reasonable close in quality to the commercial offerings so that a free software could exists which didn't tie the non commercial software community to expensive software vendors. In other words a system user friendly enough that a moderately skilled Unix system admin could handle it. My point is that goal has been exceeded, the free BSDs and Linux met and exceeded that goal so much that today the commercial Unixes don't exist.
The second level of the GNU project, in some sense the Minix project, was a desktop Unix that a programmer, not a Unix system administrator, could successfully install, configure in unusual ways if they needed to, and use. Again I think Linux has clearly met that goal. Using myself as an example in 1994 I couldn't get Linux to work right, but in 1996 I could. Where "work right" for me meant running the Unix software I cared about, not playing.mp3s, for that stuff I still used (and use) a "business OS" which was Windows until about 2001 and then OSX. But no question by 2000 installing Linux plus a software suite was way easier than installing Solaris, Irix, HPUX or AIX. And in terms of cost, Linux was been incredibly successful. Dell's Unix which was a good OEMed SCO in 1993 was $1000, and a Dell with SCO was much cheaper than what a full featured Unix workstation from Sun, SGI, IBM, DEC would cost you where the OS was $2-4k in addition to the hardware being a few thousand extra. Today Dell will give me a Unix for nothing and even if they won't I can get Unixes that run on Dells for free. The PC hardware is far better and more reliable than it was. The dream of a $2k PC Unix workstation beating a $7k workstation was realized by the early 2000s.
Then there was the goal of replacing Windows. That project has failed, because the target market changed. The Windows consumer line died, and was replaced by their enterprise NT line with the emergence of Windows XP. How do you replace something that doesn't exist anymore? The Windows enterprise line was naturally a better choice for consumers than Linux since it was so much closer to how they did things. And arguably the NT kernel is a better desktop kernel to boot. The binary driver problem that you are focusing on being a good example. And the cost of the OEM Windows enterprise is now (and was since the XP days) very low, so its hard to make a financial case. The cost today of commercial operating systems is tiny and the quality is excellent today. How can Linux compete?
But even here I think FOSS is doing a good job though in doing what the GNU project was doing in its very earliest day on the Unix platform. Getting people to run some open source software alongside their mostly commercial offerings on their commercial OSes. For example in the early 1990s the FOSS people were getting programs on Suns to use GCC (FOSS) instead of CC (Sun's $2000 C compiler). Firefox / Chrome, Open Office, VLC... are all thriving on Windows and OSX. Apache, GCC, MySQL/Postgres, Perl/PHP, have broken the barriers in the commercial market and now thrive. And we now have close 13% of the population running an open source kernel even. Yes 90% of them are running a BSD kernel (Darwin) but that's not a bad number.
On the cell phone market where the commercial offering are weaker we see something half way between embedded and desktop and open source is doing very well. You point out a lot of things that ultimately come down to poor user interface and driver problems. And yes those are problems. But there is a more substantial problem. What significant advantage has Linux ever had to offer the common man in a first world country?
So in short: I reject your analogy to Commodore. The Unix has grown not shrunk in the last 20 years and Linux is a huge player. Even on the desktop being 2nd in that market isn't bad. Did IRIX, which was at its height in 1991 when Linux came out have 1/2% ma
Turns? Take a look at/. from a decade ago. Flamebait articles, especially on Linux politics, have been key to/. attracting the crowd they want since it started.
Except now people are starting to wake up and realize that as well. It is a buggy bloated distro that got lucky because they mailed cd's to people. However now people are starting to realize Fedore and Mandriva exist. It's that simple.
Mandriva doesn't exist anymore. The company went broke again. http://www.mageia.org/en/ is the replacement if it works out.
I think you are being unfair in saying he's changing things for the sake of it. He's doing what NeXT did when they started breaking from the X / paradigm.
X works pretty well on the Mac. Its not vaporware it ships with every OSX and has worked pretty well since OSX 10.1 (though often requiring separate installation). The thing is most of the apps aren't X apps.
I don't think Canonical will pull this off. I think they've underestimated the changes they want by 1.5-2.5 orders of magnitude in terms of cost. That being said....
This isn't unnecessarily obscure. They are doing it to be able to get integrations that just aren't offered by X/Unix apps. You could layer all the features of Aqua on top of X but by forcing every app to use Cocoa you force every app to have a standard way of dealing with specifics. Like for example video content and thus cut and paste for video works.
Yes but Wayland is more than just an X server. Its primary mode is running a different type of windowing system. The situation is far more analogous to running an X server under GDI (windows) or Aqua (Mac) and primarily using the native system with a few X apps thrown in.
Was it ever any better? I like to think that journalists like Woodward and Bernstein were honest, find-the-truth, no-agenda journalists. If they were looking into Watergate now they would've been lambasted as the 'liberal media' with a 'Democrat bias' and the corruption of the Nixon Whitehouse all would've gotten lost in the noise of the day.
They were blasted that way at the time. Do a web query on Spirow Agnew and his relationship with the media. What's changed now is that we have the kind of variety we had in the 1960s in newspapers and magazine on the TV, while we don't have variety in newspapers in most cities.
The numbers are terrible. Of the 10 largest hosting companies 9 use Linux/Apache and 1 uses a custom OS/Apache. While they all offer windows its a fraction of their sites.
Linux is the 2nd most popular desktop Unix variant.
20 years ago: IRIX, SunOS, AIX, SCO (Unixware), Digital Unix, NeXT, Xenix, HPUX would have all beaten it. So in 2 years Linux is at about 7% of desktop Unix share, and NeXT (essentially) came in 1st and everyone else with the exception of AIX is essentially dead.
That ain't so bad. I wouldn't call that a failure of the GNU project.
But Linux has had fifteen years to get people to switch using the CLI heavy nerd way and what has it gotten you? 22% of the server and falling, and so low on the desktop it is literally below the margin for error.
Huh? Well actually its more like 20 years but. Lets look at where things are:
a) Server. Linux is now the dominant Unix variant, using your data having a higher percentage than all other Unixes combined 4.5x over. That's crushing defeat of companies like Sun. Look back at the discussions even a decade ago. Its even around 50% in terms of revenue which is amazing for a free system.
b) Applications. FOSS software is standard in many companies. Open Office, Firefox, gcc, apache, MySQL are huge players in their fields. The way Linux beat Sun was: i) SunOS or Solaris running proprietary applications became ii) Solaris running mostly open source applications with a few commercial thrown in iii) Linux running mostly open source applications with a few commercial thrown in
The (i) -> (ii) transition is happening (but very slowly) on the Windows world.
c) In areas of Linux's traditional strength it continues to be dominant. For example of the top 10 hosting companies 9 are Linux and the other 1 is using its own custom OS. None are primarily Windows based.
d) Embedded has been a huge win for Linux over the last 13 years.
Desktop is a disaster but the Unix desktop crowd has a pretty excellent solution in OSX/Darwin. Linux got beat there.
Heey, don't say that, Canonical's biggest contribution to the open source world has been to Gnome as they created a a set of icons. Accidently that is pretty much the only contribution they have made upstreams.
No Canonical's biggest contribution was marketing. With the rise of the Mac as a Unix, Linux as a desktop OS would have likely died with Canonical.
You target niches and go after those. And that includes creating apps for the those niches which are stronger than the commercial applications or where the price differences are substantial enough to make a difference. A good example where FOSS has done that is the programmer niche. In 1995 the vast majority of programmers used proprietary, platform specific languages and paid for development tools. FOSS has changed that.
High end hardware manufacturers (server) have Linux drivers because Linux has real market share on the high end server market. They don't have share in the Music market. To get share they would need FOSS apps that steal market share from commercial apps.
The whole idea of the rapture was invented by the fundamentalists. Other Christian sects had different end times philosophies.
The whole idea of Christ being divine was decided by the First Council of Nicaea.,
When you start convening councils to decide on which details will make up your religion, it really reminds me of the fights between which version fo emacs or vi is the 'one true way'. And just as relevant.:)
But that doesn't change the fact that the fundamentalists (Dabney, et. al) invented the rapture idea. They also don't like councils.
Moore's law hasn't held up this whole decade. Look at what's happening with CPU speeds. But...
1) We are doing way better in optimizing libraries. 2) We are making improvements in parallelism. 3) We just had a major improvement in disk speeds (SSD).
So I think we get the speed increase. 1000 ~ 2^10. We can get 10 more doubles I think.
Slashdot Mirror
Its a large number of writes not reads. Something like swap space gets allocated and even after it no longer reflects memory it stays empty for "a while". As long as "a while" is an hour or more on average you are fine.
No a mediocre SSD will crush a very expensive RAID-5 in read tests. If you go up to something like RAID-15 across 30 drives or so, yeah the raid will be better. SSD really is much faster otherwise no one would be all that excited about paying 20x the cost.
Thanks for that tip. I've been buying computer and printer memory from crucial but it was great to red the review on A&D and see how well their drives scored.
-- How much longer will these be 20x the per GB cost of a HDD?
They spread will come down but there will be a spread until long after no one is using HDD. http://en.wikipedia.org/wiki/File:Disruptivetechnology.gif
I'd agree with your 20x right now at say the 256 g - 500 g levels. But lets say when SSD is around 3t the spread might only be 10. At 20t the spread might only be 5. And maybe at 20t people start to switch in mass. This at first drives the spread back up close to 10 but within 2 years, the spread is like 2-3 with 35t SSD costing the same as 100t HDD. At that point only the most demanding users, those people who need huge amounts of cheap storage, exabytes worth still buy HDD. That's called "a retreat to quality". That small market kills their ability to get the innovations they need and so likely SSD becomes cheaper when it hits 200t or something.
Those are rough numbers but lets say around the time you are buying 1p SSD hard drive it will be cheaper than the HDD alternative.
If MSFT or Apple released updates every six months that broke at least a third of the hardware out there, wouldn't you deride them mercilessly? Of course you would. So why take that shit from your own OS?
Because it doesn't. Your facts are just wrong. Updates from OS version to OS version don't generally break anything. I don't see any evidence for Dell spending boatloads of money on anything Linux related, including the Poweredge. Suse, and RedHat do the suport work, Dell doesn't do much of anything. If they did then they could just take the data from Linux on Laptops and build their own port tree for a few thousand, not millions.
Linus changes to the kernels rarely break hardware. You simply recompile the kernel using the same static config, the way Fedora, RedHat, Mandriva, Debian.... do and the new stuff doesn't change much of anything. Ubuntu uses the Debian branch not Linus' branch (the vanilla branch) so his changes don't even affect them necessarily. You want a stable branch: Andi Kleen's, Greg Kroah-Hartman and Adrian Bunk maintain stable branches. Don't use vanilla.
And more importantly, if you like the more stable branches why use Ubuntu which doesn't support them? Frankly everything Ubruntu stands for is making Debian less stable. Another example of a very stable Linux is Cent.
Think about your argument in the last dozen posts, and ask yourself if this were true why Linux is so dominant on server, on supercomputing, on embedded. Your arguments aren't desktop specific anymore. Open Solaris had great binary compatibility I don't remember them being all that popular on the desktop.
And here we have the twin psychotic face of Linux thing. On the one hand time and time again (hell sites like /. are practically bursting with articles written by the community) we here you Linux advocates say "Linux is ready for the desktop!" yet when anyone points out the serious problems that give way to that lie you basically say "Oh its not an OS, its a community! Go back to Windows if you won't do things OUR way". So which is it? Can't eat your cake and have it too you know.
Well there is two things.
The first is that your only experience is with Windows desktop environments. There are millions of people all over the country using non Windows OSes right now on desktops. Companies like Burlington Coat Factory, Autozone, Pep Boys were able to switch easily, because they had a Unix not a Windows culture in their workforce. Education of HS or less so its not computer geeks. Millions of people still use mainframe OSes, and curses style interfaces. You simply lack experience as to what is else is possible. Your feeling is that if it is not done "the windows way" then its just wrong. The purpose of the GNU project was never to replace the NT kernel it was to change the computing culture. Having Linux do things the windows way, would subvert the broader goal. Having people use Windows while running ever increasing amounts of free software and become more concerned about open protocols (see the debate about net neutrality) advances the agenda.
Now as for the rest. It is ready for the desktop, and has been for a decade. Which is different then saying it is ready to overthrow Microsoft Windows. If Linux were the dominant desktop today, Windows would not be able to overthrow it either. The cost of switching between them is high in terms of the entire community of expectations. But for those people who have some reason to want to switch it is a fully viable desktop and has been for years. So I'm rejecting the equivalence you are making between an absolute worst case scenario for Linux and "the norm".
Assume for a moment Microsoft raised the OEM price of Windows to $1000, you don't think you would see computers shipping with high quality well maintained Linux distributions quickly? That's what ready means. That if Linux had to step in, it could step in. Now you made very specific claims about Dell which were easily refuted by looking at the history of the Power Edge for example. The point about the Power Edge is that with a motivated vendor long term driver updates were possible. Which I think should tell you that maybe you need to re-examine your assumptions about the way the Linux community works and the fact that it is able to support hardware when there is a genuine motivation. Which is different than your case where there isn't.
Richard Stallman, 1983
To begin with, GNU will be a kernel plus all the utilities needed to write and run C programs: editor, shell, C compiler, linker, assembler, and a few other things. After this we will add a text formatter, a YACC, an Empire game, a spreadsheet, and hundreds of other things. We hope to supply, eventually, everything useful that normally comes with a Unix system, and anything else useful, including on-line and hardcopy documentation.
The goal was to provide a Unix system. First and foremost Linux has to be a Unix and do things the Unix way. The goal was not to build a free Amiga. There is another project that aims to do that, Syllable and maybe what you should do is work with them.
And the reason your arguments much credit is I remember very much the same arguments being made in the late 1990s and early 2000s by big box Unixes. "Linux is fine for hobbyists but its never going to replace Solaris"...
I have yet to hear any reason from you would want to use Linux or for that matter any Unix.
a) You hate the GUIs
b) You hate the apps
c) You don't want to support non typical usage.
As for drivers, features like kernel patch protection. And yes, the NT mode that Cutler proposed is also closer to the Linux philosophy.
Your post is basically a rant of: "I hate Unix, I hate everything that's different between Unix and Windows. I have no need for any Unix features or applications". Well in that case use Windows. I don't like golf, I don't playing golf, I don't want a golf membership; ergo you aren't going to be able to sell me golf clubs regardless of what you do with them. Why would anyone expect your small hardware shop to support or offer Linux? Were you supporting SCO or OS/2 15 years ago?
Dell's PowerEdge Line has had solid Linux support for most of a decade now. Including excellent driver support and compatibility with most distributions. No serious problems remotely like what you are describing. Today, they ship with Suse Enterprise or RHEL with no need for specialized repository. The M Series seems to be doing well and the N series has existed off and on with no problems. Because those people want to run Unix applications they are willing to do things the Unix way. And there Dell gets excellent support.
But in terms of Dell not getting support they've gotten great support. I've been dealing with Dell's and Linux for 17 years, I can always find excellent information from the community on how to configure Dell's and deal with their quirks. That's good support. Now if you mean working out the box. That's Dell's job, they create an OEM version config file just like they do for Windows today.
As far as the Ubuntu line. Horrible garbage low end hardware for cheap machines.... I can see Linux going downmarket in the 3rd world but what 1st world customer wouldn't rather just pay $30-40 for an OEM version of windows 7? Netbooks were a missed opportunity. Cell phones are working out.
In terms of GUI features:
* -- XML building so that the GUI can be complete reconfigured by end users (or OEMs). Apple sorta supports something like this, Microsoft doesn't allow it at all.
* -- Entourage like window management
* -- desktop grid: allows zooming in and out of virtual windows.
* -- transparency and opacity
Those things didn't exist on GUI's in the 1990s at all. Apple introduced them last decade and Microsoft with Vista. So lets cut the non sense about Windows 9X. Try doing those things on 9X.
I think you need some realism. Linux is not a product its a community. You want people to work to service your needs, call Microsoft. You want to help build a culture of information sharing, then we have something to talk about.
Your specific complaints are:
a) The kernel abi is not binary stable.
b) The operating systems rarely require a CLI interaction.
c) Commercial applications are better, in particular easier to use.
In all 3 areas I think you are failing to take into account the difference in spend.
(a) has been a major area of debate. The advantages of stable binary drivers have been discussed ad nauseam, and that's because they really do exist. The problems of bad binary drivers corrupting systems have been discussed ad nauseam. I'm not going to get into the details but there really are tradeoffs. The general direction over the last 15 years has been for
* Windows to move closer to the Linux way of doing things
* Linux to move closer to the Windows way of doing things
* Most vendors to just write their own and limit hardware
Microsoft's driver solution literally costs multiple billions of dollars per year. It is simply not an option of the Linux community in terms of manpower needed. If you need excellent hardware support across the full range of PC hardware, you have one option in terms of OSes. No one else is going to be comparable for a long time. Linux does an excellent job at less than 1/4000th of the cost. But yes 99.98% reduction is spending does have end user impact. If you have $2b a year for a driver lab I'm sure Linus would love the help and would be happy to do it Microsoft's way. The Linux solution has worked so well that the Linux kernel is the dominant embedded kernel because of the huge range of hardware it support. So I think you either being completely unrealistic about what is achievable and failing to appreciate how good a job the Linux kernel team has done in terms of hardware support given the tremendous differences in spend.
On (b) CLI bleed is a result of cost savings. Just to pick an example, Excel never implemented the GUI features of ideas like Jazz from Lotus because of the additional complexity; so Excel functions are CLI "sum(a7..b23)" with some level of GUI interface and not "sum(sales table)" like you had on Lotus products because at the time Excel wasn't making that kind of money. CLI is much cheaper than GUI, by about two orders of magnitude.
The first attempt to even build a real GUI for Linux didn't start until 1996. By 2001 or so there was a GUI, which still required some CLI, on par with say Windows 3.1 (so say about 7 years behind). Today's Linux GUIs are better but the commercial GUIs have massively improved over the last 10 years so Linux has lost ground. I definitely think you can argue Linux GUIs for the last 5 years are as good as Windows 2000 was and are approaching XP. So maybe they've lost a little ground and are more like 9 years behind but its been a very tough decade to keep pace. Further Windows 2000 was a premium product. I had to mess with config files on Windows 2000 quite a bit. Similarly I think you can compare Linux GUIs favorably to OSX 10.1 where people frequently did have to drop to terminal to get things to work. Those systems were usable by end users at the time, I think you are exaggerating.
If you point is that Linux GUIs are worse than those that represent tens of billions of dollars in GUI development, well yeah. But again I think you can say that Linux has far and away the best GUIs of any non mainstream operating system. I don't disagree that there is no particular reason an end user should want to downgrade their experience. I will comment those same end users often run software that is 5, 10, 15 years old, as you mentioned so perhaps they dislike change more than they like features and so it really doesn't matter very much.
In terms of apps., your point (c) I think its worth noting that widgets sets like QT and GTK are now the standard for most non platform specific software development so overtime you will see Linux applications improving in the commercial sector. .NET and Cocoa are really really good so for applications makers willing to inc
I think you ignored the core of my comment. The goal of the GNU project was to have a free Unix that was reasonable close in quality to the commercial offerings so that a free software could exists which didn't tie the non commercial software community to expensive software vendors. In other words a system user friendly enough that a moderately skilled Unix system admin could handle it. My point is that goal has been exceeded, the free BSDs and Linux met and exceeded that goal so much that today the commercial Unixes don't exist.
The second level of the GNU project, in some sense the Minix project, was a desktop Unix that a programmer, not a Unix system administrator, could successfully install, configure in unusual ways if they needed to, and use. Again I think Linux has clearly met that goal. Using myself as an example in 1994 I couldn't get Linux to work right, but in 1996 I could. Where "work right" for me meant running the Unix software I cared about, not playing .mp3s, for that stuff I still used (and use) a "business OS" which was Windows until about 2001 and then OSX. But no question by 2000 installing Linux plus a software suite was way easier than installing Solaris, Irix, HPUX or AIX. And in terms of cost, Linux was been incredibly successful. Dell's Unix which was a good OEMed SCO in 1993 was $1000, and a Dell with SCO was much cheaper than what a full featured Unix workstation from Sun, SGI, IBM, DEC would cost you where the OS was $2-4k in addition to the hardware being a few thousand extra. Today Dell will give me a Unix for nothing and even if they won't I can get Unixes that run on Dells for free. The PC hardware is far better and more reliable than it was. The dream of a $2k PC Unix workstation beating a $7k workstation was realized by the early 2000s.
Then there was the goal of replacing Windows. That project has failed, because the target market changed. The Windows consumer line died, and was replaced by their enterprise NT line with the emergence of Windows XP. How do you replace something that doesn't exist anymore? The Windows enterprise line was naturally a better choice for consumers than Linux since it was so much closer to how they did things. And arguably the NT kernel is a better desktop kernel to boot. The binary driver problem that you are focusing on being a good example. And the cost of the OEM Windows enterprise is now (and was since the XP days) very low, so its hard to make a financial case. The cost today of commercial operating systems is tiny and the quality is excellent today. How can Linux compete?
But even here I think FOSS is doing a good job though in doing what the GNU project was doing in its very earliest day on the Unix platform. Getting people to run some open source software alongside their mostly commercial offerings on their commercial OSes. For example in the early 1990s the FOSS people were getting programs on Suns to use GCC (FOSS) instead of CC (Sun's $2000 C compiler). Firefox / Chrome, Open Office, VLC... are all thriving on Windows and OSX. Apache, GCC, MySQL/Postgres, Perl/PHP, have broken the barriers in the commercial market and now thrive. And we now have close 13% of the population running an open source kernel even. Yes 90% of them are running a BSD kernel (Darwin) but that's not a bad number.
On the cell phone market where the commercial offering are weaker we see something half way between embedded and desktop and open source is doing very well. You point out a lot of things that ultimately come down to poor user interface and driver problems. And yes those are problems. But there is a more substantial problem. What significant advantage has Linux ever had to offer the common man in a first world country?
So in short: I reject your analogy to Commodore. The Unix has grown not shrunk in the last 20 years and Linux is a huge player. Even on the desktop being 2nd in that market isn't bad. Did IRIX, which was at its height in 1991 when Linux came out have 1/2% ma
Turns? Take a look at /. from a decade ago. Flamebait articles, especially on Linux politics, have been key to /. attracting the crowd they want since it started.
Except now people are starting to wake up and realize that as well. It is a buggy bloated distro that got lucky because they mailed cd's to people. However now people are starting to realize Fedore and Mandriva exist. It's that simple.
Mandriva doesn't exist anymore. The company went broke again. http://www.mageia.org/en/ is the replacement if it works out.
I think you are being unfair in saying he's changing things for the sake of it. He's doing what NeXT did when they started breaking from the X / paradigm.
X works pretty well on the Mac. Its not vaporware it ships with every OSX and has worked pretty well since OSX 10.1 (though often requiring separate installation). The thing is most of the apps aren't X apps.
I don't think Canonical will pull this off. I think they've underestimated the changes they want by 1.5-2.5 orders of magnitude in terms of cost. That being said....
This isn't unnecessarily obscure. They are doing it to be able to get integrations that just aren't offered by X/Unix apps. You could layer all the features of Aqua on top of X but by forcing every app to use Cocoa you force every app to have a standard way of dealing with specifics. Like for example video content and thus cut and paste for video works.
Yes but Wayland is more than just an X server. Its primary mode is running a different type of windowing system. The situation is far more analogous to running an X server under GDI (windows) or Aqua (Mac) and primarily using the native system with a few X apps thrown in.
Was it ever any better? I like to think that journalists like Woodward and Bernstein were honest, find-the-truth, no-agenda journalists. If they were looking into Watergate now they would've been lambasted as the 'liberal media' with a 'Democrat bias' and the corruption of the Nixon Whitehouse all would've gotten lost in the noise of the day.
They were blasted that way at the time. Do a web query on Spirow Agnew and his relationship with the media. What's changed now is that we have the kind of variety we had in the 1960s in newspapers and magazine on the TV, while we don't have variety in newspapers in most cities.
The numbers are terrible. Of the 10 largest hosting companies 9 use Linux/Apache and 1 uses a custom OS/Apache. While they all offer windows its a fraction of their sites.
Linux is the 2nd most popular desktop Unix variant.
20 years ago: IRIX, SunOS, AIX, SCO (Unixware), Digital Unix, NeXT, Xenix, HPUX would have all beaten it. So in 2 years Linux is at about 7% of desktop Unix share, and NeXT (essentially) came in 1st and everyone else with the exception of AIX is essentially dead.
That ain't so bad. I wouldn't call that a failure of the GNU project.
But Linux has had fifteen years to get people to switch using the CLI heavy nerd way and what has it gotten you? 22% of the server and falling, and so low on the desktop it is literally below the margin for error.
Huh? Well actually its more like 20 years but. Lets look at where things are:
a) Server. Linux is now the dominant Unix variant, using your data having a higher percentage than all other Unixes combined 4.5x over. That's crushing defeat of companies like Sun. Look back at the discussions even a decade ago. Its even around 50% in terms of revenue which is amazing for a free system.
b) Applications. FOSS software is standard in many companies. Open Office, Firefox, gcc, apache, MySQL are huge players in their fields. The way Linux beat Sun was:
i) SunOS or Solaris running proprietary applications became
ii) Solaris running mostly open source applications with a few commercial thrown in
iii) Linux running mostly open source applications with a few commercial thrown in
The (i) -> (ii) transition is happening (but very slowly) on the Windows world.
c) In areas of Linux's traditional strength it continues to be dominant. For example of the top 10 hosting companies 9 are Linux and the other 1 is using its own custom OS. None are primarily Windows based.
d) Embedded has been a huge win for Linux over the last 13 years.
Desktop is a disaster but the Unix desktop crowd has a pretty excellent solution in OSX/Darwin. Linux got beat there.
Heey, don't say that, Canonical's biggest contribution to the open source world has been to Gnome as they created a a set of icons. Accidently that is pretty much the only contribution they have made upstreams.
No Canonical's biggest contribution was marketing. With the rise of the Mac as a Unix, Linux as a desktop OS would have likely died with Canonical.
1) Mac uses /dev as well. They just shield end users from knowing about it. There is no reason a noob needs to know about the dev file system.
2) I agree. Creative names are a problem. Though directories organized by function can fix that somewhat.
3) I'm not sure what you mean here. Most distributions have pretty good GUI package managers for over a decade.
4) Methinks you haven't counted how many executables there are on a typical Unix system. You don't want 10,000 icons.
You target niches and go after those. And that includes creating apps for the those niches which are stronger than the commercial applications or where the price differences are substantial enough to make a difference. A good example where FOSS has done that is the programmer niche. In 1995 the vast majority of programmers used proprietary, platform specific languages and paid for development tools. FOSS has changed that.
High end hardware manufacturers (server) have Linux drivers because Linux has real market share on the high end server market. They don't have share in the Music market. To get share they would need FOSS apps that steal market share from commercial apps.
OK so what specifically is the biology thing I fail to know?
The whole idea of the rapture was invented by the fundamentalists. Other Christian sects had different end times philosophies.
The whole idea of Christ being divine was decided by the First Council of Nicaea.,
When you start convening councils to decide on which details will make up your religion, it really reminds me of the fights between which version fo emacs or vi is the 'one true way'. And just as relevant. :)
But that doesn't change the fact that the fundamentalists (Dabney, et. al) invented the rapture idea. They also don't like councils.
Moore's law hasn't held up this whole decade. Look at what's happening with CPU speeds. But...
1) We are doing way better in optimizing libraries.
2) We are making improvements in parallelism.
3) We just had a major improvement in disk speeds (SSD).
So I think we get the speed increase. 1000 ~ 2^10. We can get 10 more doubles I think.