So what do you think I should do? Sell all my KDE stock? Use the same desktop that everyone else does, so that if the wolves attack they'll only eat the outlying sheep? Run around screaming in the streets that GNOME IZ DA SHITZ! Of course not. This is Linux, you are free to choose whatever you want. My only point is that people should be aware of the advantages and disadvantages of each platform, that's all. If you like/want to use KDE/Xfce/E17/etc., that's all fine with me.
Most Novell development is Gnome oriented? Is that why Novell developed so many of their tools using QT? For example, look at Mono, which is a big focus at Novell. GTK# is a very high priority (in fact, this is the preferred platform, instead of Windows.Forms). Note that Qt# exists, but is not developed by Novell (it appears to have a single developer and is hosted on the kde.org servers).
You also suggest there are only 3 major distros which is also pretty short sighted. Mandriva isn't a major distro? What about Gentoo? Or Debian? Or PCLinuxOS? Mandriva used to be more influential. But currently it appears to have fewer users than the top 3 I mentioned. The same is true for Gentoo and PCLinux OS - but I admit that accurate statistics are hard to find, so this is debatable. Regarding Debian, it isn't really a desktop-focused distro, so not relevant to the issue.
You say SLED defaults to Gnome, and I wouldn't know because I haven't tried it. Link: "In upcoming versions of Novell enterprise applications, the default desktop environment will be GNOME," but they do mention KDE will remain an option. Novell is focused on GNOME, as evidenced by their enterprise offering being GNOME-default and by Mono focused on GTK#, as mentioned above.
openSUSE - KDE predominately. Well, SLED has GNOME as default, and most Linux development in Novell is GNOME-oriented. OpenSUSE offers both GNOME and KDE (the first option is GNOME, though). But clearly Novell is betting on GNOME based on the first things I said.
In conclusion, the 3 major distros - Red Hat, SUSE, Ubuntu - default to GNOME. Yes, Ubuntu offers KDE as a secondary option and SUSE offers KDE in openSUSE, but still the top priority is GNOME. You gave a list of several other distros; yes, several use KDE, but these are smaller distros (nothing wrong with them, just not relevant to my point).
It is probably part of the reason why all major distros have moved to GNOME.
Only ONE major distro that standardizes on GNOME, and that's the distro that helps fund it. All of the Ubuntus are really just one distro, so they're not standardized on anything. All the other major distros (SuSE, Mandriva, Debian, etc.) either standardize on KDE or leave the choice up to the user during install. Red Hat is standardized on GNOME. Ubuntu has GNOME as the default, and if you test Ubuntu vs Kubuntu, you can see that the latter is far less polished; also, note how the next release will be LTS only for GNOME (not KDE, not Xfce) - GNOME is the top priority. Novell's enterprise offerings are all standardized on GNOME; openSUSE defaults to GNOME, and most development focuses on GNOME, but KDE is also an option.
That leaves Mandriva and Debian in your list. Debian isn't really a 'desktop distro', so it isn't that relevant here, but yes, it has no preference for any DE. Mandriva is standardized on KDE, but I wouldn't say it is among the 'major distros', by which it is commonly meant Red Hat, SUSE, Ubuntu.
Of course, this risk with KDE basing itself on Qt was obvious all the time due to the licensing model there. It is probably part of the reason why all major distros have moved to GNOME. What risk? Qt Free is licensed under the GNU GPL and has been for the last 7 years. There is absolutely NOTHING stopping the KDE developers, or anybody, from forking Qt, as long as the forks remain under the GPL. The risks were explained elsewhere in the thread. To summarize: the forks will always be GPL2 (and now GPL2/GPL3). We will never be able to write GPL4 apps for KDE in such a situation. And there are other problems with other licenses, etc.
Major distros have moved to GNOME mostly because GTK+ is LGPL, which means they (or third parties) can develop commercial apps for it without paying a licensing fee. If you want to write a commercial Qt app, you gotta pay the piper (Trolltech, and now Nokia)
That is part of it, yes. Another part aside from commercial apps is that the range of FOSS licenses is also limited. Red Hat doesn't want to arrive at a situation where it wants to write GPL4 apps but can't due to the restrictions of its chosen widget toolkit.
Not sure I like "effectively LGPL", as it seems a funny way of saying userspace. The Linux Kernel is *not* LGPL, it is for sure GPL2. Seems like you could confuse people into thinking they have LGPL access to the kernel when they do not. Yes, you're absolutely right. I was just trying to say it in a short way that summarizes the essence of the situation. Basically, with LGPL code like GTK, you can base an app on it with any license, but change GTK itself and you need to follow its licensing instructions. The same is sort of true for the kernel - run apps on it, use any license, but if you change the kernel itself, then you have to abide by its license. In that sense the two situations are similar (and very different than that of Qt). If you have a better way to summarize this in a short sentence, I'd be grateful (this topic tends to come up from time to time).
Uhm, if it reverts to a BSD-style license, doesn't that mean we can, you know, change the license? That is the idea, yes. But the point was regarding the other case - where it doesn't revert to BSD. Then you have a problem.
Trolltech then nicely decided to agree to let people use the GPL3 - but they didn't have to. And when the GPL4 comes along, we'll have to hope that Nokia decides to allow it.
Do they have a track record of "doing the wrong thing" that they've only deviated from a time or two? Why assume it's going to get much worse now?
I don't assume anything. Not assuming anything, I am aware of possible future risks; this is one of them.
(Btw, on an 'emotional' level, I trust Trolltech more than Nokia. Trolltech has some likable guys there; Nokia is pushing for software patents in Europe. But I digress.)
When you're Red Hat, you don't want to build your OS in a way that lets another corporation control a critical aspect of it.
Kind of like how they're no longer using the GPLv2-only Linux kernel. Gotcha.
I never have understood why QT is always held to a different standard than other software. Even though it's GPLv2 + GPLv3 + closed if you wanna pay for it, that just doesn't please some people.
Red Hat knows that 99% of apps that run on Linux are in userspace. For those, Linux is effectively LGPL, not GPL - any license is possible. In that sense the Linux kernel is like GTK+, and unlike Qt: all apps that use Qt must follow its specific licensing guidelines.
Therefore Red Hat prefers Linux and GTK over Qt, as I read the map. I tend to agree with them.
Re:I've been waiting for *someone* to buy TrollTec
on
Nokia Buys Trolltech
·
· Score: 4, Interesting
I'm personally crossing my fingers for Nokia to change the license to LGPL. Yep, me too.
In retrospect, I consider Qt one of the two biggest 'misses' in open-source, the other being OpenSolaris. If Trolltech had 'gotten it' in time, GNOME wouldn't exist, and Qt/KDE would dominate the Linux desktop completely, a great vantage point from which to consider other markets. Likewise, if Sun had 'gotten it' way back then, OpenSolaris would be what Linux is today, Linux wouldn't exist, and Sun would be making a fortune. Yes, all of this is in retrospect, but the two stories are interesting, I think. And both revolve around fears of 'going all the way' with an open-source business model. Problem is, waiting too long is even worse. Maybe Nokia will get it right?
GPL for apps, LGPL for libraries. It's scary to a commercial enterprise, but it really works - at least better than anything else (except, maybe, having a monopoly on desktop operating systems...).
Of course, this risk with KDE basing itself on Qt was obvious all the time due to the licensing model there.
As others have pointed out, QT is available under the GPL. I seem to be feeling a little dense today. Could you please explain to me and others why building one GPLed project on another GPLed project is a risky venture?
It's explained elsewhere in this thread, but I'll do it again.
Until a month ago, Qt was GPL2. It couldn't include e.g. code from Samba for that reason. And you couldn't write apps for KDE that are GPL3 only, which some people want to do. (Note that this problem is not solvable by forking Qt.) Trolltech then nicely decided to agree to let people use the GPL3 - but they didn't have to. And when the GPL4 comes along, we'll have to hope that Nokia decides to allow it. The same goes for any new license, say, the Affero GPL (just as an example of a new type of license that may turn out to be very important).
The issue is that the foundations of a system are better off LGPLed than GPLed. Or some licensing model that is effectively LGPLed. That is what e.g. the Linux kernel does - you can write an app that runs on a Linux distro that uses any license. The same is true for GTK. With Qt, the control of what licenses you are permitted to use lies in the hands of a single commercial entity, whose actions you cannot predict (especially when you look a few years into the future).
This inherent risk in Qt and KDE is, I believe, one reason for the major distros choosing GNOME. When you're Red Hat, you don't want to build your OS in a way that lets another corporation control a critical aspect of it. If Qt were LGPLed, like GTK, this would not be an issue.
> No, GTK is LGPL, not GPL. That is the critical difference.
OK, so it's not exactly the same, but it's not critical in any way. From KDE's point of view this changes nothing. Qt will still be GPL as it has been a number of years now and Nokia cannot take that away. GPL 'about the same' as the LGPL? My friend, wars have been fought over less:)
Here is a simple possible scenario. A critical tool is written in a license not compatible with Qt's licensing, as e.g. Samba was up until a few weeks ago. KDE was unable to utilize it, while GNOME had no such problem. Thankfully, meanwhile Trolltech decided to allow the community to use GPL3 code, so this particular matter is solved... but the same issue will arrive with GPL4, and other licenses. For a similar problem, see ZFS and the Linux kernel.
You might say, "ah, other licenses, I'll never need those"... and you might not. But you might, and it might turn out to be critical, as it almost did with Samba. The future is hard to predict. GTK is future-proofed since it is LGPL, Qt is not, it depends on the continuing goodwill of Nokia.
Of course, this risk with KDE basing itself on Qt was obvious all the time due to the licensing model there. It is probably part of the reason why all major distros have moved to GNOME. nonsense...
openSuse and Mandriva are more like KDE-distros, Debian isn't gnome based either.
KDE is more popular than gnome in Europe for instance... The default desktop on Red Hat(/Fedora), SUSE and Ubuntu is GNOME. I don't think Mandriva is doing so well these days as to be in the same list. As for Debian, it's an amazing distro, but not a desktop-focused one (however, yes, it has no default).
In summary, yes, GNOME is the default on all major distros, as I read the map. Perhaps in some geographical areas Mandriva is successful, that's true, but overall the top three are pretty much Red Hat, SUSE and Ubuntu these days.
I don't see Nokia as interested in the Linux desktop Are you kidding. Why not go and take a quick search on Nokia Internet Tablets such as the 700/N800/N810 and you'll see they are very active in linux development. Also check out Maemo.org, which is developed by Nokia and is debian based. You might say that is specialized and not the "desktop" but it is very end user and it would be in Nokia's best interest to keep the development rolling. Regarding the tablets and Maemo, note that these are GTK-based projects, so I'm not sure they are related to the purchase of Trolltech - there is no direct benefit, Trolltech and Maemo are orthogonal (will Nokia scrap Maemo? I doubt it). No, it seems far likelier that the purchase has all to do with mobile devices, phones in particular, an area Trolltech was working very much on getting into with their Qtopia platform, etc. Given who Nokia is, I think we can bet that mobile devices are the basis here.
Interesting. TFA states that Nokia plans to continue to develop Qt, though, and will continue to offer it under both open source and commercial licenses, just as things are now.
I assume that means as long as Nokia continues to develop Qt in the same manner (keeping Qt Free available for KDE), then the agreement doesn't apply.
Yes, all Nokia needs to do is keep Qt development on a low burner to avoid BSD-ization of their code. Not hard to do.
I don't see Nokia as interested in the Linux desktop, so I presume that part of Trolltech's work will not continue exactly as before; why pay the salaries of several KDE developers, for example - not sure Nokia will see the point in that. I don't predict immediate firings, though, but if I was one of them I wouldn't count on long-term job security. What I do see Nokia as wanting from Trolltech is everything related to mobile devices, Qtopia, all that stuff. So overall Qt may continue to be developed, but I'm not sure its focus won't move to one that is less useful for KDE.
Of course, this risk with KDE basing itself on Qt was obvious all the time due to the licensing model there. It is probably part of the reason why all major distros have moved to GNOME.
Licensing Issues for the Future
on
Nokia Buys Trolltech
·
· Score: 2, Interesting
In case anyone can feel the panic setting in while thoughts of closed source qt libraries swirl around their heads it may be as well to remind people that Troll Tech and KDE have this all worked out nicely already.
The KDE Foundation takes the code if qt is ever released closed. Not sure if it covers a buy out situation but I'm pretty sure it does. First, Qt is already released closed-source: it has several licenses, one of which is closed, others of which are GPL2, GPL3. But I presume you meant that they stop the FOSS releases. Then yes, you are essentially correct: if Qt stops being released in an (among others) FOSS license, it reverts to being BSD. However, this far from solves the issue. Qt may continue to be released as FOSS, but its development may stagnate, if e.g. Nokia's priorities are more towards mobile devices and less on desktop Linux (which makes sense, given what products Nokia specializes in). In that case, KDE will suffer, and little can be done.
You might say, "but then the community can fork Qt." Yes, a fork is possible. The fork will then be GPL2/GPL3, which is somewhat problematic, in that in the future we will never be able to write KDE apps in GPL4, should such be released (and I presume that Microsoft's attacks on FOSS will necessitate a GPL4 eventually, just as Microsoft's deal with Novell necessitated certain clauses in the GPL3). That is, yes, we can fork Qt, but we cannot add licenses to it (only the copyright holders can, and Nokia is now that entity). Thankfully Trolltech helped out KDE this time by letting Qt be GPL3, but next time, we have no assurances whatsoever.
Ubuntu is fast becoming a powerful player in this area; as the article says, the reason for supporting it was sizable customer demand. That is the logic here. People wanted to run Ubuntu on their enterprise desktops, they wanted IBM to have Notes on that platform, IBM agreed. No mystery. I'd like to know where this is happening. I haven't seen or read anything that would support this statement apart from the relentless ubuntu PR. It's all RH and Novel/Suse in the enterprise; ubuntu is virtually nonexistent in this space. Ignore the Ubuntu PR, all PR is suspect. TFA in fact is valid evidence: IBM has decided to support Ubuntu because of actual demand. TFA says, for example,
"We're doing pilots with customers now," Satyadas said. "Some of the requests came from big companies" with as many as 100,000 users that are interested in moving to Ubuntu Linux on the desktop. [empasis mine] Yes, Red Hat has most of the enterprise market, but for servers. That is 99% of the current Linux market, and is the reason you feel Ubuntu is 'nonexistent'. Desktop Linux is starting to slowly appear in enterprises, while this may not be the 'year of the Linux desktop', it is making progress. When it does, Ubuntu is often the name mentioned. It's a desktop-focused distribution, it has commercial backing, and it captured the top spot among desktop Linux enthusiasts.
I'm unable to understand the logic here. Is the word Ubuntu replacing Linux for marketing use, or is there some compelling reason to support just one distribution? In "the old days" (last year was it?) everything was SUSE. In "the REAL old days" (2 years ago was it?) it was RedHat. Linux is Linux is Linux. Ubuntu is fast becoming a powerful player in this area; as the article says, the reason for supporting it was sizable customer demand. That is the logic here. People wanted to run Ubuntu on their enterprise desktops, they wanted IBM to have Notes on that platform, IBM agreed. No mystery.
Of course 'support for Ubuntu' doesn't mean it won't run on random distro X. It might, but IBM won't recommend it/install it/support it for you. Which is fine if you want to do it all yourself. Most enterprises, however, are used to paying IBM (/Microsoft) a lot of money and not having to worry about support issues.
IBM, by the way, isn't supporting just one distro. They have various forms of support for various distros for their products. Their overall strategy seems quite simple; on the one hand, support the distros people ask for, on the other, keep that number a reasonable size. By which I mean, IBM doesn't want a single vendor (Microsoft sort of taught IBM a lesson there), but also IBM doesn't want too many vendors, which is hard to support and market. Simply put, that means we should expect IBM products to be supported on Ubuntu, Red Hat and SUSE. No surprises; these are the major distros these days (and for a few years now, too).
The updates most interesting to me are the anti-fragmentation patches,
Tests show that about 60-70% of physical memory can be allocated on a desktop after a few days uptime. In benchmarks and stress tests, it has been found that 80% of memory is available as contiguous blocks at the end of the test. To compare, a standard kernel was getting ~1% of memory as large pages on a desktop and about 8-12% of memory as large pages at the end of stress tests.
Perhaps someone can clarify exactly what this means? Reading the beginning, it talked about 4K pages, device drivers, and such, so I assumed it would just be relevant to the internal workings of the kernel. However, the quote I pasted above seems to indicate it might impact desktop performance as well.
I commonly see on my desktop, after several days uptime, that quite a lot of memory is being used (and I know how to ignore cache/buffers, as well as swapcache - that isn't the issue). Logging out and logging back in returns memory to reasonable levels (and the system becomes more responsive, but then I guess if I bought more memory I could accomplish that as well). Now, I've generally read that the problem was indeed memory fragmentation, e.g. here, but this would be internal fragmentation inside an app, and thus not relevant to the kernel, I believe? If someone can explain this issue I'd be grateful.
Radioshack settled a similar lawsuit with their store managers several years ago, and lowered their base salaries to offset the new overtime payouts. I've heard about lots of this sort of thing going on in smaller corporations, that you wouldn't hear about in the news. Fact is, the 'industry norm' is in many cases to not pay overtime for these sorts of jobs, even though people constantly work beyond the normal hours (these aren't 9 to 5 jobs!). As compensation, the base salaries are typically quite high. But it turns out that this norm is somewhat at odds with certain laws regarding overtime, and employees in many cases demand what they think they deserve.
The end result is exactly what IBM did. Suddenly starting to pay for overtime means IBM is raising effective salaries by 10-20% or more, so naturally IBM lowers base salaries. The end result is that we are exactly where we started - people work the same hours, and get the same pay.
Well, at least on average; for individuals who work more or who work less, there will be some change. There are also motivational issues - if you are paid for overtime, you have less incentive to work efficiently (one reason why hi-tech managers, and many workers, don't like paid overtime and prefer to raise the base pay). Overall, it is hard to say that the change is for the better. The old salaries and norms were already 'working' - they were comparable to industry norms, were arrived at after years of haggling, corrections, and so forth, and most importantly people knew what they were getting when they signed on.
If you would have bothered to read the article, the author did spend some time making sure that no server components like apache or mysql were installed. Although there probably were more programs on the linux installs anyway. First, it isn't in the linked article. I presume you refer not to TFA, but to the PDF linked to by an article that TFA links to.
In that PDF there are the details you mention. I still doubt the results, though. For example, on a typical Windows machine a user would install (say) Acrobat, iTunes, WinZip and some IM client other than Microsoft's (note that these are good targets for hackers). None of the vulnerabilities in these apps go into the Windows vulnerability figures. On Linux, comparable apps are included both in the installation media as well as in these vulnerability figures.
Perhaps the only really fair way to do it - which I have yet to see carried out, sadly - is a component-by-component analysis. That is, how many vulnerabilities in Windows' TCP/IP stack vs. Linux's, and so forth. Obviously it won't always be easy to determine categories, but it would be better than the methodology in this report.
Microsoft remains extremely wealthy... and the proceeds to charge money for Microsoft virtualization products. They've done it before.
Umm, right.
You mean like they charge money for Internet Explorer now (Netscape)? And Media Player (Real)? And the.NET Framework (Java)? And Silverlight (Flash)? Is Microsoft charging for IE? Well, what if tomorrow they said that Windows was $5 cheaper - a 'present to the world'. Then the day after, they said, we are charging $5 for IE per copy of Windows. How would that final stage be any different than where we are today?
Developing IE costs money. Microsoft passes that cost on to consumers. For further evidence, consider the rising price of Windows, and by that I mean, look at the price range for the various versions of Vista. How many people really pay for the lowest-end version and not one of the costlier ones (even if bundled into the cost of a new computer)?
Why is Apple able to deliver an OS that is faster instead of slower? It's got as much eye candy as vista. And Compiz on Linux offers eye candy as well, also with fewer resource requirements. But the reasons for this are fairly clear:
Microsoft isn't competing with OS X or Linux. It has a guaranteed monopoly market share. No reason to outdo the other players on technical merit.
Microsoft makes most of its Windows money from sales of new computers. The question is then, how do you convince them to buy a new one. It could be because the old computer is full of viruses, or because the new OS (which you must have for security reasons, or to run DirectX 10, or whatever other made up excuse) simply is too big and slow for the old hardware. This is basically the business model. Apple, on the other hand, convinces people to buy new computers for other reasons, one major one is that the new hardware is very sexy. The MacBook Air is a clear example of this - people want it just for the hardware (when did you ever hear someone say that about a Windows machine?).
If they [Microsoft] do give their hypervisor away for free, then VMware will release ESX too and nothing will really change! I think this can only be a good thing for us:-) Unless after a few years of giving away their products for free, VMware goes bankrupt. Microsoft, on the other hand, remains extremely wealthy due to Windows and Office revenue, and then proceeds to charge money for Microsoft virtualization products. This is a good thing only for Microsoft, and they've done it before.
Slashdot Mirror
Mandriva used to be more influential. But currently it appears to have fewer users than the top 3 I mentioned. The same is true for Gentoo and PCLinux OS - but I admit that accurate statistics are hard to find, so this is debatable. Regarding Debian, it isn't really a desktop-focused distro, so not relevant to the issue. You say SLED defaults to Gnome, and I wouldn't know because I haven't tried it. Link: "In upcoming versions of Novell enterprise applications, the default desktop environment will be GNOME," but they do mention KDE will remain an option. Novell is focused on GNOME, as evidenced by their enterprise offering being GNOME-default and by Mono focused on GTK#, as mentioned above.
Well, SLED has GNOME as default, and most Linux development in Novell is GNOME-oriented. OpenSUSE offers both GNOME and KDE (the first option is GNOME, though). But clearly Novell is betting on GNOME based on the first things I said.
In conclusion, the 3 major distros - Red Hat, SUSE, Ubuntu - default to GNOME. Yes, Ubuntu offers KDE as a secondary option and SUSE offers KDE in openSUSE, but still the top priority is GNOME. You gave a list of several other distros; yes, several use KDE, but these are smaller distros (nothing wrong with them, just not relevant to my point).
It is probably part of the reason why all major distros have moved to GNOME.
Only ONE major distro that standardizes on GNOME, and that's the distro that helps fund it. All of the Ubuntus are really just one distro, so they're not standardized on anything. All the other major distros (SuSE, Mandriva, Debian, etc.) either standardize on KDE or leave the choice up to the user during install. Red Hat is standardized on GNOME. Ubuntu has GNOME as the default, and if you test Ubuntu vs Kubuntu, you can see that the latter is far less polished; also, note how the next release will be LTS only for GNOME (not KDE, not Xfce) - GNOME is the top priority. Novell's enterprise offerings are all standardized on GNOME; openSUSE defaults to GNOME, and most development focuses on GNOME, but KDE is also an option.
That leaves Mandriva and Debian in your list. Debian isn't really a 'desktop distro', so it isn't that relevant here, but yes, it has no preference for any DE. Mandriva is standardized on KDE, but I wouldn't say it is among the 'major distros', by which it is commonly meant Red Hat, SUSE, Ubuntu.
The risks were explained elsewhere in the thread. To summarize: the forks will always be GPL2 (and now GPL2/GPL3). We will never be able to write GPL4 apps for KDE in such a situation. And there are other problems with other licenses, etc. Major distros have moved to GNOME mostly because GTK+ is LGPL, which means they (or third parties) can develop commercial apps for it without paying a licensing fee. If you want to write a commercial Qt app, you gotta pay the piper (Trolltech, and now Nokia)
That is part of it, yes. Another part aside from commercial apps is that the range of FOSS licenses is also limited. Red Hat doesn't want to arrive at a situation where it wants to write GPL4 apps but can't due to the restrictions of its chosen widget toolkit.
Do they have a track record of "doing the wrong thing" that they've only deviated from a time or two? Why assume it's going to get much worse now?
I don't assume anything. Not assuming anything, I am aware of possible future risks; this is one of them.(Btw, on an 'emotional' level, I trust Trolltech more than Nokia. Trolltech has some likable guys there; Nokia is pushing for software patents in Europe. But I digress.) When you're Red Hat, you don't want to build your OS in a way that lets another corporation control a critical aspect of it.
Kind of like how they're no longer using the GPLv2-only Linux kernel. Gotcha.
I never have understood why QT is always held to a different standard than other software. Even though it's GPLv2 + GPLv3 + closed if you wanna pay for it, that just doesn't please some people.
Red Hat knows that 99% of apps that run on Linux are in userspace. For those, Linux is effectively LGPL, not GPL - any license is possible. In that sense the Linux kernel is like GTK+, and unlike Qt: all apps that use Qt must follow its specific licensing guidelines.Therefore Red Hat prefers Linux and GTK over Qt, as I read the map. I tend to agree with them.
Yep, me too.
In retrospect, I consider Qt one of the two biggest 'misses' in open-source, the other being OpenSolaris. If Trolltech had 'gotten it' in time, GNOME wouldn't exist, and Qt/KDE would dominate the Linux desktop completely, a great vantage point from which to consider other markets. Likewise, if Sun had 'gotten it' way back then, OpenSolaris would be what Linux is today, Linux wouldn't exist, and Sun would be making a fortune. Yes, all of this is in retrospect, but the two stories are interesting, I think. And both revolve around fears of 'going all the way' with an open-source business model. Problem is, waiting too long is even worse. Maybe Nokia will get it right? GPL for apps, LGPL for libraries. It's scary to a commercial enterprise, but it really works - at least better than anything else (except, maybe, having a monopoly on desktop operating systems...).
Good summary, I agree completely.
As others have pointed out, QT is available under the GPL. I seem to be feeling a little dense today. Could you please explain to me and others why building one GPLed project on another GPLed project is a risky venture?
It's explained elsewhere in this thread, but I'll do it again.Until a month ago, Qt was GPL2. It couldn't include e.g. code from Samba for that reason. And you couldn't write apps for KDE that are GPL3 only, which some people want to do. (Note that this problem is not solvable by forking Qt.) Trolltech then nicely decided to agree to let people use the GPL3 - but they didn't have to. And when the GPL4 comes along, we'll have to hope that Nokia decides to allow it. The same goes for any new license, say, the Affero GPL (just as an example of a new type of license that may turn out to be very important).
The issue is that the foundations of a system are better off LGPLed than GPLed. Or some licensing model that is effectively LGPLed. That is what e.g. the Linux kernel does - you can write an app that runs on a Linux distro that uses any license. The same is true for GTK. With Qt, the control of what licenses you are permitted to use lies in the hands of a single commercial entity, whose actions you cannot predict (especially when you look a few years into the future).
This inherent risk in Qt and KDE is, I believe, one reason for the major distros choosing GNOME. When you're Red Hat, you don't want to build your OS in a way that lets another corporation control a critical aspect of it. If Qt were LGPLed, like GTK, this would not be an issue.
OK, so it's not exactly the same, but it's not critical in any way. From KDE's point of view this changes nothing. Qt will still be GPL as it has been a number of years now and Nokia cannot take that away. GPL 'about the same' as the LGPL? My friend, wars have been fought over less
Here is a simple possible scenario. A critical tool is written in a license not compatible with Qt's licensing, as e.g. Samba was up until a few weeks ago. KDE was unable to utilize it, while GNOME had no such problem. Thankfully, meanwhile Trolltech decided to allow the community to use GPL3 code, so this particular matter is solved... but the same issue will arrive with GPL4, and other licenses. For a similar problem, see ZFS and the Linux kernel.
You might say, "ah, other licenses, I'll never need those"... and you might not. But you might, and it might turn out to be critical, as it almost did with Samba. The future is hard to predict. GTK is future-proofed since it is LGPL, Qt is not, it depends on the continuing goodwill of Nokia.
In summary, yes, GNOME is the default on all major distros, as I read the map. Perhaps in some geographical areas Mandriva is successful, that's true, but overall the top three are pretty much Red Hat, SUSE and Ubuntu these days.
That is certainly how Trolltech saw it. But Nokia controls things now, and time will tell whether it sees things the same way.
I assume that means as long as Nokia continues to develop Qt in the same manner (keeping Qt Free available for KDE), then the agreement doesn't apply.
Yes, all Nokia needs to do is keep Qt development on a low burner to avoid BSD-ization of their code. Not hard to do.
I don't see Nokia as interested in the Linux desktop, so I presume that part of Trolltech's work will not continue exactly as before; why pay the salaries of several KDE developers, for example - not sure Nokia will see the point in that. I don't predict immediate firings, though, but if I was one of them I wouldn't count on long-term job security. What I do see Nokia as wanting from Trolltech is everything related to mobile devices, Qtopia, all that stuff. So overall Qt may continue to be developed, but I'm not sure its focus won't move to one that is less useful for KDE.
Of course, this risk with KDE basing itself on Qt was obvious all the time due to the licensing model there. It is probably part of the reason why all major distros have moved to GNOME.
The KDE Foundation takes the code if qt is ever released closed. Not sure if it covers a buy out situation but I'm pretty sure it does. First, Qt is already released closed-source: it has several licenses, one of which is closed, others of which are GPL2, GPL3. But I presume you meant that they stop the FOSS releases. Then yes, you are essentially correct: if Qt stops being released in an (among others) FOSS license, it reverts to being BSD. However, this far from solves the issue. Qt may continue to be released as FOSS, but its development may stagnate, if e.g. Nokia's priorities are more towards mobile devices and less on desktop Linux (which makes sense, given what products Nokia specializes in). In that case, KDE will suffer, and little can be done.
You might say, "but then the community can fork Qt." Yes, a fork is possible. The fork will then be GPL2/GPL3, which is somewhat problematic, in that in the future we will never be able to write KDE apps in GPL4, should such be released (and I presume that Microsoft's attacks on FOSS will necessitate a GPL4 eventually, just as Microsoft's deal with Novell necessitated certain clauses in the GPL3). That is, yes, we can fork Qt, but we cannot add licenses to it (only the copyright holders can, and Nokia is now that entity). Thankfully Trolltech helped out KDE this time by letting Qt be GPL3, but next time, we have no assurances whatsoever.
Of course 'support for Ubuntu' doesn't mean it won't run on random distro X. It might, but IBM won't recommend it/install it/support it for you. Which is fine if you want to do it all yourself. Most enterprises, however, are used to paying IBM (/Microsoft) a lot of money and not having to worry about support issues.
IBM, by the way, isn't supporting just one distro. They have various forms of support for various distros for their products. Their overall strategy seems quite simple; on the one hand, support the distros people ask for, on the other, keep that number a reasonable size. By which I mean, IBM doesn't want a single vendor (Microsoft sort of taught IBM a lesson there), but also IBM doesn't want too many vendors, which is hard to support and market. Simply put, that means we should expect IBM products to be supported on Ubuntu, Red Hat and SUSE. No surprises; these are the major distros these days (and for a few years now, too).
I commonly see on my desktop, after several days uptime, that quite a lot of memory is being used (and I know how to ignore cache/buffers, as well as swapcache - that isn't the issue). Logging out and logging back in returns memory to reasonable levels (and the system becomes more responsive, but then I guess if I bought more memory I could accomplish that as well). Now, I've generally read that the problem was indeed memory fragmentation, e.g. here, but this would be internal fragmentation inside an app, and thus not relevant to the kernel, I believe? If someone can explain this issue I'd be grateful.
The end result is exactly what IBM did. Suddenly starting to pay for overtime means IBM is raising effective salaries by 10-20% or more, so naturally IBM lowers base salaries. The end result is that we are exactly where we started - people work the same hours, and get the same pay.
Well, at least on average; for individuals who work more or who work less, there will be some change. There are also motivational issues - if you are paid for overtime, you have less incentive to work efficiently (one reason why hi-tech managers, and many workers, don't like paid overtime and prefer to raise the base pay). Overall, it is hard to say that the change is for the better. The old salaries and norms were already 'working' - they were comparable to industry norms, were arrived at after years of haggling, corrections, and so forth, and most importantly people knew what they were getting when they signed on.
In that PDF there are the details you mention. I still doubt the results, though. For example, on a typical Windows machine a user would install (say) Acrobat, iTunes, WinZip and some IM client other than Microsoft's (note that these are good targets for hackers). None of the vulnerabilities in these apps go into the Windows vulnerability figures. On Linux, comparable apps are included both in the installation media as well as in these vulnerability figures.
Perhaps the only really fair way to do it - which I have yet to see carried out, sadly - is a component-by-component analysis. That is, how many vulnerabilities in Windows' TCP/IP stack vs. Linux's, and so forth. Obviously it won't always be easy to determine categories, but it would be better than the methodology in this report.
Umm, right.
You mean like they charge money for Internet Explorer now (Netscape)? And Media Player (Real)? And the
Developing IE costs money. Microsoft passes that cost on to consumers. For further evidence, consider the rising price of Windows, and by that I mean, look at the price range for the various versions of Vista. How many people really pay for the lowest-end version and not one of the costlier ones (even if bundled into the cost of a new computer)?
Unless after a few years of giving away their products for free, VMware goes bankrupt. Microsoft, on the other hand, remains extremely wealthy due to Windows and Office revenue, and then proceeds to charge money for Microsoft virtualization products. This is a good thing only for Microsoft, and they've done it before.