This is a very serious problem of Ubuntu that is overlooked by the developers. Problems with specific hardware like those in the TFA, can be common especially these days with so many different combinations of monitors and video cards. I'd like to see some sort of "safe mode" that kicks in when there are problems, and a GUI to allow proper reconfiguration.
You are correct about this problem, however, the developers are not ignoring it. In fact they were considering implementing more or less what you suggested for Feisty. This has been deferred, however, and for good reason - X.Org, in a future release (7.3, IIRC) will offer related functionality. So Ubuntu developing it themselves would be a lot of effort, for just a few months.
Hopefully with the next X.Org and the next (after Feisty) Ubuntu we will see many of the typical X problems disappear.
Ubuntu is still its own OS (as are the other distros): See Ian Murdoch essay.
As such, no platform exists for PC software vendors to target.
The Ian Murdoch essay focuses on C/C++, which is where the main problem lies. If one develops in the 'bytecode' family of languages/environments - Java, Python, C# - then porting between Linux distros is actually quite simple, in my experience at least. Using a cross-platform GUI toolkit - WxWidgets, GTK+, Qt - can even allow porting to Windows and others. For example, look at the (old) 'official' Bittorrent client: written in Python and WxWidgets, it works on all OSes I am familiar with, no hassle at all.
As more development is done in bytecode languages, this problem is getting smaller and smaller. However, it does exist, I agree, and work needs to be done to improve it in the meanwhile (and work is being done, albeit slowly).
Other than Microsoft being able to claim that Windows runs on 99.99% of the PCs sold, you're absolutely correct. [...] So yeah, there's little point in whining, except for that statistics thing.
I completely disagree. This isn't just Microsoft getting the statistics of "Windows on 99% of desktops". They also get paid for each one! If the crapware offsets the price of Windows, the fact remains that Microsoft still gets paid for it - it is just the crapware manufacturers paying Microsoft, not you.
The entire system is built so that Microsoft gets paid no matter what. By buying a cheap PC with Windows, and installing Linux, you are (in effect) taking money out of the pockets of crapware manufacturers - which I am 100% (or more) ok with. But what I am not ok with is that that money goes to strengthen Microsoft's stranglehold on the PC market, which I see as causing a lot of harm to the industry.
You analyze things quite fairly. I do have one item on which I disagree, though: the most pro-consumer thing to do would be to notify the user if their copy is non-genuine. There is no need to, in addition, lock down the copy (into 'restricted mode').
The lockdown is certainly not in consumers' interest, only Microsoft's.
True, some of the people who get around activation are pirates. But WGA also causes a lot of issues for paying customers. In particular, the basic fact is that Microsoft will probably patch these issues, and any patch carries a risk of causing things to break.
True, windows simply is not trustworthy. I mean automatic updates are something great, but a company, which uses such a system to further their own interests and not that of their customers is simply unacceptable.
100% agreement with you. Notice, though, how (at the end of TFA) Microsoft's position is that product activation is for the benefit of their customers. Something along the lines of "products hacked to avoid activation may be faulty" and such. So, a forced patch through Windows Update would be 'for the good of the customers', to save them from the perils of running WGA-less Windows. War is peace, and all that.
One can only hope that in the long run such anti-consumer activity will come back to haunt them.
TCO is total cost of ownership, yes, but it is just the cost. To assess whether it is worth the cost, you need to know what you get from it. If product A costs 10 times more than product B, but gives 100 times the functionality, then it is worth it (from a price/performance standpoint).
If that is so, then the question I have can be rephrased as: why is Linux competing only with Sun, i.e. why isn't Windows a possible alternative to Linux and Sun (as Microsoft would have)?
I'm a Linux fan myself, but your reasoning here doesn't make sense. TCO means Total Cost of Ownership. It should take into account monetary benefits. How do you think it could be "worth it" if TCO is higher? A warm feeling in your tummy?
TCO is "total cost of ownership", so it doesn't consider the benefits of ownership. That is, it is just the 'price' out of 'price/performance'. Product A can cost 10 times more than product B, but do 100 times more useful things. So, even if Linux has a higher TCO (which personally I doubt), it might make up for that by being a better product (which, certainly in the server space, I believe it is).
Would not Cygwin be a viable solution, in that case, that is to port them to Cygwin on Windows rather than Linux? - not a rhetorical question, I am honestly asking (never used Cygwin myself).
If HSBC thinks that Linux has a higher TCO than Windows, then why do they even have Linux machines?
The only reasons I can think of are that
They have Linux-only apps that they can't run on Windows. (Is that likely? Perhaps someone here can shed some light on that matter.)
Linux has a higher TCO, but is worth it.
Linux had a higher TCO when using multiple distros, but after consolidating to Novell SUSE, they expect Linux's TCO to be below Windows'. TFA does focus on their moving to a single Linux distro to cut costs, but doesn't mention whether after that cut Linux will have a competitive TCO vs. Windows or not.
Well, besides 'good vs. evil', there is also 'rooting for the underdog'. Nintendo is certainly the underdog when competing against corporations the size of Microsoft and Sony. Personally I like rooting for the underdog sometimes.
What's going to happen when Dell releases a flavor that can't play MP3s, or some media files, out of the box?
Well, why would Dell do that? It doesn't make much sense. They should just install a Linux setup that does play MP3s and so forth (for home PC usage, at least). Legal codecs are available, they can get them from Fluendo, or via CNR.
If they charged for it, there'd be a political price to pay. People would ask why they're charging for something OO.o offers for free.
Well, the answer is simple: they are selling support.
There would be calls to ask Dell to donate to OO.o. People would complain about internet updates costing OO.o. They'd have to pay careful attention to this or risk bad PR, primarily.
This might have a PR influence, but only among computer geek types. And among those, the majority would applaud Dell installing OO.o, I would think (but I might be wrong). So my personal guess about the PR issue is that it would be an overall win.
it would be tricky for Dell to attempt to charge customers for [OpenOffice]
Tricky how? Legally there is no issue here. The only thing they need to do is set up the support for OpenOffice, as far as I can tell. Or am I missing something else that you intended?
As far as I know, Microsoft provides the support for Microsoft Office, and I can't see why Dell would offer support for OpenOffice since they have nothing to do with its development.
Dell might want to offer support for OpenOffice for a very good reason - because it makes them money. On the one hand selling Microsoft Office gives Dell a small profit (I presume that nearly all the money goes to Microsoft), and low support costs (they can't be zero, since people will call even if they are immediately redirected to Microsoft; still, they are low). On the other hand, the grandparent post was right - they can charge whatever they want for OpenOffice - even much less than MS Office - and still make a nice profit, because they pocket all of that profit themselves, only needing to spend on support. My guess is that this could be very profitable. It is probably not happening because (1) Dell are conservative, and this is a big change, and (2) possibly Microsoft would retaliate and raise prices for Windows and Office, phrased as 'reducing their discounts'.
Exactly, this is well-known. Piracy is a crucial part of the Microsoft strategy, and it works very well.
What I think is important to realize is that this is something somewhat unique to software. You don't see BMW being happy that their cars are being stolen - although there might be some 'prestige' factor in being the car thief's favorite, more theft can quickly (1) cause people to fear owning BMWs, and (2) cause the insurance costs for owning a BMW to skyrocket.
Microsoft's piracy strategy is only possible because piracy isn't theft; it's copyright violation. No actual product is stolen. And this is because generating copies of software has no cost, i.e., zero marginal value. This interesting property of software is the basis both for Microsoft's piracy strategy, and for FOSS in general - but with completely different results.
I also think that the hard sciences are dying, but for completely different reasons. Note that by 'hard sciences' I really mean the harder ones, Physics and CS, and Math.
Basically, these are dying out because they have (1) run into a wall, and (2) encountered tough 'competition'.
Regarding the wall: look at those fields in the first half of the 20th century, and the second half. There is really no comparison, the first half was revolutionary, the 2nd far, far less. Einstein's amazing insights led to the atomic bomb, a fitting end for the first half of the century: a theoretical masterpiece that led to incredible practical applications. Likewise mathematics had some amazing stuff (Godel, for example). But more recently, far less. There is a basic inability of the hard sciences to deal with problems of interest today, it seems. This leads to hybrid approaches, such as Neural Computation (neural networks and such). Analytic investigation may just have a limit, and we may have already reached it; all the low-hanging fruit has been plucked.
On the other hand, the competition for bright young minds is harder now. In particular, medicine and biology are making amazing discoveries, that impact people's lives - finding the cause of ulcers and their cure, a vaccine for HPV, brain implants to give some simple visual capabilities to the blind, just to give a few examples. And the amount of money flowing into these fields is incredible. Donors just seem to care less about some equation in physics that interests only 100 people worldwide, versus healing the ill (which, in the case of many old rich donors, includes themselves).
Between these two things, the hard/theoretical sciences are dwindling somewhat. As someone who works in them, I certainly hope they won't die out, and I don't think they will. But their standing is certainly not on the rise.
I think he does a good job in pointing out the 'brand loyalty' issue with OSS - people prefer distros etc. for reasons that are, in part, not completely rational. But he also points out other issues relevant here, on slide 38, issues that are relevant to both proprietary software and OSS: that software has risks, and has a high degree of inability to estimate those risks. This leads to brand loyalty for fairly rational reasons.
For example, it makes sense to stick to Red Hat as opposed to switching to Oracle's new offering, given that software is tricky stuff. Oracle may apply Red Hat patches late, they may apply them wrongly, and they make enter new bugs of their own. The same is true for CentOS - no matter how similar it is supposed to be to Red Hat, it still depends on someone applying the Red Hat patches and so forth.
Not that I wouldn't use CentOS, I have in the past, and I might in the future. But for large corporations, it is perfectly rational to be loyal to a successful brand like Red Hat, even though they sell OSS. Thus, OSS is not a commodity market, even though it might seem like one at first glance. At least that is how I understand the slides.
If google is helping crazy regimes stay in power that is a very bad thing.
Hang on - India a crazy regime? You can't be serious. India is a democracy - just as democratic as the West, actually. There is no 'crazy regime staying in power' in India, by any stretch of the imagination.
Sure, this specific censorship issue sounds a bit odd to some (including me), but no more odd than things happening in, say France and the UK, just to mention very recent Slashdot stories.
I know nobody wants to admit it... but this is why Microsoft is actually a good thing on the desktop market.
I disagree completely. Ignoring the question of whether the 'one OS is simpler to develop for' argument was ever true, it certainly isn't anymore. Develop once in a cross-platform tool - say, Java or Python, using GTK+ or Qt - and basically you can run anywhere. Yes, yes, you do need to test on lots of platforms, and yes, there will be problems. But then there are also problems running Windows 2000-age apps on Vista (possibly more, even).
There is no need for a single OS to make life simple for developers. What developers need are universal platforms - and not necessarily just one! Choice is a good thing to have. One can write in any of the 'bytecode languages/platforms' (Java, Python,.Net), using any of the GUI toolkits that is available for all platforms (GTK+, Qt, WxWidgets), and life is pretty good.
In Ubuntu or Kubuntu you need to replace both GNOME and KDE to get something stable. They apply a bunch of experimental patches to "improve" the experience, but the patches often creates more bugs.
What? I am using Ubuntu right here with GNOME, it works fine. Sure there are a few bugs, but all software has bugs (I've seen more bugs on other 'major' distros, actually). So I really don't know what you are complaining about. Perhaps you should be more specific about what bugs you have encountered.
As for Kubuntu and KDE, I don't use them, so I can't say.
Second Life servers run Debian and use MySQL. They are transitioning to use Mono as a scripting language (from their own scripting language, which apparently isn't working out so well).
Which is nice. However, not open-sourcing their server code is somewhat disappointing. Oh well, at least the client is open, someone else can create a FOSS server if the interest ever arises.
Slashdot Mirror
You are correct about this problem, however, the developers are not ignoring it. In fact they were considering implementing more or less what you suggested for Feisty. This has been deferred, however, and for good reason - X.Org, in a future release (7.3, IIRC) will offer related functionality. So Ubuntu developing it themselves would be a lot of effort, for just a few months.
Hopefully with the next X.Org and the next (after Feisty) Ubuntu we will see many of the typical X problems disappear.
The Ian Murdoch essay focuses on C/C++, which is where the main problem lies. If one develops in the 'bytecode' family of languages/environments - Java, Python, C# - then porting between Linux distros is actually quite simple, in my experience at least. Using a cross-platform GUI toolkit - WxWidgets, GTK+, Qt - can even allow porting to Windows and others. For example, look at the (old) 'official' Bittorrent client: written in Python and WxWidgets, it works on all OSes I am familiar with, no hassle at all.
As more development is done in bytecode languages, this problem is getting smaller and smaller. However, it does exist, I agree, and work needs to be done to improve it in the meanwhile (and work is being done, albeit slowly).
I completely disagree. This isn't just Microsoft getting the statistics of "Windows on 99% of desktops". They also get paid for each one! If the crapware offsets the price of Windows, the fact remains that Microsoft still gets paid for it - it is just the crapware manufacturers paying Microsoft, not you.
The entire system is built so that Microsoft gets paid no matter what. By buying a cheap PC with Windows, and installing Linux, you are (in effect) taking money out of the pockets of crapware manufacturers - which I am 100% (or more) ok with. But what I am not ok with is that that money goes to strengthen Microsoft's stranglehold on the PC market, which I see as causing a lot of harm to the industry.
You analyze things quite fairly. I do have one item on which I disagree, though: the most pro-consumer thing to do would be to notify the user if their copy is non-genuine. There is no need to, in addition, lock down the copy (into 'restricted mode').
The lockdown is certainly not in consumers' interest, only Microsoft's.
True, some of the people who get around activation are pirates. But WGA also causes a lot of issues for paying customers. In particular, the basic fact is that Microsoft will probably patch these issues, and any patch carries a risk of causing things to break.
True, but - if (for example) further security updates require this one, then it is essentially forced.
100% agreement with you. Notice, though, how (at the end of TFA) Microsoft's position is that product activation is for the benefit of their customers. Something along the lines of "products hacked to avoid activation may be faulty" and such. So, a forced patch through Windows Update would be 'for the good of the customers', to save them from the perils of running WGA-less Windows. War is peace, and all that.
One can only hope that in the long run such anti-consumer activity will come back to haunt them.
TCO is total cost of ownership, yes, but it is just the cost. To assess whether it is worth the cost, you need to know what you get from it. If product A costs 10 times more than product B, but gives 100 times the functionality, then it is worth it (from a price/performance standpoint).
If that is so, then the question I have can be rephrased as: why is Linux competing only with Sun, i.e. why isn't Windows a possible alternative to Linux and Sun (as Microsoft would have)?
I'm a Linux fan myself, but your reasoning here doesn't make sense. TCO means Total Cost of Ownership. It should take into account monetary benefits. How do you think it could be "worth it" if TCO is higher? A warm feeling in your tummy?
TCO is "total cost of ownership", so it doesn't consider the benefits of ownership. That is, it is just the 'price' out of 'price/performance'. Product A can cost 10 times more than product B, but do 100 times more useful things. So, even if Linux has a higher TCO (which personally I doubt), it might make up for that by being a better product (which, certainly in the server space, I believe it is).
Would not Cygwin be a viable solution, in that case, that is to port them to Cygwin on Windows rather than Linux? - not a rhetorical question, I am honestly asking (never used Cygwin myself).
The only reasons I can think of are that
Well, besides 'good vs. evil', there is also 'rooting for the underdog'. Nintendo is certainly the underdog when competing against corporations the size of Microsoft and Sony. Personally I like rooting for the underdog sometimes.
Well, the answer is simple: they are selling support.
This might have a PR influence, but only among computer geek types. And among those, the majority would applaud Dell installing OO.o, I would think (but I might be wrong). So my personal guess about the PR issue is that it would be an overall win.
As mentioned in other threads, this isn't obvious at all. Dell don't need to give OpenOffice away, they can charge for it.
Tricky how? Legally there is no issue here. The only thing they need to do is set up the support for OpenOffice, as far as I can tell. Or am I missing something else that you intended?
Dell might want to offer support for OpenOffice for a very good reason - because it makes them money. On the one hand selling Microsoft Office gives Dell a small profit (I presume that nearly all the money goes to Microsoft), and low support costs (they can't be zero, since people will call even if they are immediately redirected to Microsoft; still, they are low). On the other hand, the grandparent post was right - they can charge whatever they want for OpenOffice - even much less than MS Office - and still make a nice profit, because they pocket all of that profit themselves, only needing to spend on support. My guess is that this could be very profitable. It is probably not happening because (1) Dell are conservative, and this is a big change, and (2) possibly Microsoft would retaliate and raise prices for Windows and Office, phrased as 'reducing their discounts'.
Exactly, this is well-known. Piracy is a crucial part of the Microsoft strategy, and it works very well.
What I think is important to realize is that this is something somewhat unique to software. You don't see BMW being happy that their cars are being stolen - although there might be some 'prestige' factor in being the car thief's favorite, more theft can quickly (1) cause people to fear owning BMWs, and (2) cause the insurance costs for owning a BMW to skyrocket.
Microsoft's piracy strategy is only possible because piracy isn't theft; it's copyright violation. No actual product is stolen. And this is because generating copies of software has no cost, i.e., zero marginal value. This interesting property of software is the basis both for Microsoft's piracy strategy, and for FOSS in general - but with completely different results.
I also think that the hard sciences are dying, but for completely different reasons. Note that by 'hard sciences' I really mean the harder ones, Physics and CS, and Math.
Basically, these are dying out because they have (1) run into a wall, and (2) encountered tough 'competition'.
Regarding the wall: look at those fields in the first half of the 20th century, and the second half. There is really no comparison, the first half was revolutionary, the 2nd far, far less. Einstein's amazing insights led to the atomic bomb, a fitting end for the first half of the century: a theoretical masterpiece that led to incredible practical applications. Likewise mathematics had some amazing stuff (Godel, for example). But more recently, far less. There is a basic inability of the hard sciences to deal with problems of interest today, it seems. This leads to hybrid approaches, such as Neural Computation (neural networks and such). Analytic investigation may just have a limit, and we may have already reached it; all the low-hanging fruit has been plucked.
On the other hand, the competition for bright young minds is harder now. In particular, medicine and biology are making amazing discoveries, that impact people's lives - finding the cause of ulcers and their cure, a vaccine for HPV, brain implants to give some simple visual capabilities to the blind, just to give a few examples. And the amount of money flowing into these fields is incredible. Donors just seem to care less about some equation in physics that interests only 100 people worldwide, versus healing the ill (which, in the case of many old rich donors, includes themselves).
Between these two things, the hard/theoretical sciences are dwindling somewhat. As someone who works in them, I certainly hope they won't die out, and I don't think they will. But their standing is certainly not on the rise.
I think he does a good job in pointing out the 'brand loyalty' issue with OSS - people prefer distros etc. for reasons that are, in part, not completely rational. But he also points out other issues relevant here, on slide 38, issues that are relevant to both proprietary software and OSS: that software has risks, and has a high degree of inability to estimate those risks. This leads to brand loyalty for fairly rational reasons.
For example, it makes sense to stick to Red Hat as opposed to switching to Oracle's new offering, given that software is tricky stuff. Oracle may apply Red Hat patches late, they may apply them wrongly, and they make enter new bugs of their own. The same is true for CentOS - no matter how similar it is supposed to be to Red Hat, it still depends on someone applying the Red Hat patches and so forth.
Not that I wouldn't use CentOS, I have in the past, and I might in the future. But for large corporations, it is perfectly rational to be loyal to a successful brand like Red Hat, even though they sell OSS. Thus, OSS is not a commodity market, even though it might seem like one at first glance. At least that is how I understand the slides.
Hang on - India a crazy regime? You can't be serious. India is a democracy - just as democratic as the West, actually. There is no 'crazy regime staying in power' in India, by any stretch of the imagination.
Sure, this specific censorship issue sounds a bit odd to some (including me), but no more odd than things happening in, say France and the UK, just to mention very recent Slashdot stories.
I disagree completely. Ignoring the question of whether the 'one OS is simpler to develop for' argument was ever true, it certainly isn't anymore. Develop once in a cross-platform tool - say, Java or Python, using GTK+ or Qt - and basically you can run anywhere. Yes, yes, you do need to test on lots of platforms, and yes, there will be problems. But then there are also problems running Windows 2000-age apps on Vista (possibly more, even).
There is no need for a single OS to make life simple for developers. What developers need are universal platforms - and not necessarily just one! Choice is a good thing to have. One can write in any of the 'bytecode languages/platforms' (Java, Python,
What? I am using Ubuntu right here with GNOME, it works fine. Sure there are a few bugs, but all software has bugs (I've seen more bugs on other 'major' distros, actually). So I really don't know what you are complaining about. Perhaps you should be more specific about what bugs you have encountered.
As for Kubuntu and KDE, I don't use them, so I can't say.
Second Life servers run Debian and use MySQL. They are transitioning to use Mono as a scripting language (from their own scripting language, which apparently isn't working out so well).
Which is nice. However, not open-sourcing their server code is somewhat disappointing. Oh well, at least the client is open, someone else can create a FOSS server if the interest ever arises.