yet they're also releasing an open sourced distribution of Linux.
No, they are not: MadHatter contains some seriously proprietary bits, foremost Sun Java, and their Linux distribution probably does as well.
Sun has never been above using open source or free software when its suits them. SunOS and Solaris, after all, were based on a lot of BSD code. But, ultimately, Sun wants to own big chunks of intellectual property in what they are shipping.
I wouldn't want to support someone so wishy washy
I think Sun isn't wishy-washy at all: they know exactly what they are doing. And perhaps that is just the reason why you shouldn't support them.
Kudos to Sun for finally getting the desktop right.
Have you ever actually used any of Sun's user interfaces? SunView, OpenWindows, NeWS, Swing, OpenOffice? I don't think they have ever gotten user interfaces "right", and I seriously doubt they have gotten this one right.
Gnome by itself seems much more consistent and usable than any mix of Java, OpenOffice, and Gnome could ever be. Keep in mind that Java and OpenOffice don't even use the native Gnome toolkit or libraries.
So, let's see, Sun's proposition is that they will "enhance" the Linux desktop by throwing Java in there.
Now, Java is about as proprietary a platform as they come; just to download the specifications, you have to agree to a license that requires any implementation you base on it to pass a Sun conformance test (I'd like to link the license itself, but the Sun site uses some dynamic content that makes that difficult; just click on the download next to the documentation to see the license). If you dare to download the JRE or JDK, the license gets even scarier, with having to donate pretty much everything you do "based" on that to Sun.
Well, it's par for the course, I suppose: the same cast of characters from Sun Microsystems has been trying to replace open UNIX desktop GUIs with something proprietary before, most notably with NeWS. (As an aside, one of the main Java movers and shakers, Gosling, actually sparked the creation of the FSF and GNU Emacs by creating a proprietary version of Emacs.)
Maybe there would be some argument that this is a deal one had to live with if Sun actually had something to offer in terms of user interfaces. But look at what Sun's history of user interfaces: SunView, OpenWindows, and NeWS--not exactly stellar success stories, either in terms of technology or in terms of industry adoption. These days, Sun is shipping Swing, which manages to be bloated, slow, and thoroughly unintegrated with the Gnome or KDE desktops, and OpenOffice, which also manages to be bloated, slow, and thoroughly unintegrated with the Gnome or KDE desktops.
If Sun wants to ship MadHatter, that's their business: Gnome is open source and as long as they comply with the GPL/LGPL, they can do whatever they want. But I think the Linux desktop and UI needs help from Sun about as much as the US needs economic advice from North Vietnam.
For the longest time, many ham radio operators were sticking with old versions of Windows or even DOS. As a result, a lot of the software for interfacing to radios was DOS/Windows only and many companies were unwilling to publish protocols. Hopefully, that is going to change as more hams are moving to Linux.
Companies like Vonage really are only here for a transitional period: they give you a way to connect VoIP service to the regular phone network. That's not a long-term business model.
But for Internet-to-Internet calls, any attempt at regulation would be futile. In fact, there doesn't even have to be any kind of business involved in the middle.
States can, of course, tax IP traffic or Internet access, but regulations that try to distinguish between different uses of that traffic would be very hard and costly to enforce and trivial to work around.
SCO can have the GPL or BSD licenses declared invalid all they want--the code is still copyrighted. With the GPL or BSD in place, at least they have the excuse that "it's free anyway, we just got sloppy about satisfying a few conditions". But if they actually argue that they believe the GPL and BSD license are invalid, then they are committing willful copyright infringement of code they have no license at all for.
Perl packages have been packaged as Debian packages for quite some time--the package and all dependencies are automatically downloaded and they are automatically kept updated.
Shouldn't Netgear be liable for the costs they impose on the university? I mean, Netgear is a money-making enterprise and they chose to point an excessive number of clients at the service, they should pay for the IT staff, hardware, and bandwidth necessary to service their requests.
And, in addition to the monetary aspect, this would seem to fall under computer hacking statutes. I suspect that if this sort of thing had resulted from a student or small developer releasing some software, the FBI might be knocking on their door already, and protestations that it was an accident might not be listened to. I think Morris also claimed that his worm (the first Internet worm) escaped by accident, yet he got charged.
Well, the stupidity of the way their NTP client was written (hard-coded IP address, 1sec queries, etc.) is pretty obvious. And of course, the Netgear technical support response ("we don't have time to respond to your problem, but maybe it has gone away on its own already") is typical of the company.
But what seems particularly stupid is that many companies would jump at the opportunity of having an excuse for obtaining detailed usage information about their customers. They would point these devices at their own servers (probably with 1h rather than 1s query intervals) and know exactly how many of their devices are in operation and what software versions they are running.
That is almost certainly what Microsoft and Apple are using their NTP servers for: keeping track of number of installed systems by the day and by IP address blocks (home users, Fortune-500, etc.).
So, we have to conclude that not only Netgear's programming is iffy and that their customer support is pretty awful, they also fail to behave like a rational US corporation by giving up the opportunity of keeping tabs on their customers. Well, maybe that makes them a little less bad, actually, and at least their gear is cheap and works most of the time...
Carbonate deposits would most likely be found in the basins, but those are also the areas most likely to be filled with dust, lava, and ice. On Earth, those low points with their deposits eventually get lifted up into mountain ranges, which is why they get exposed, but I believe on Mars, that doesn't happen very much. So, the result that we don't see a lot of carbonates isn't all that surprising even if there were moderately sized oceans on Mars a long time ago.
McBride: The Canopy Group [of Utah] is an investment company. Those are just ignorant statements about SCO's business. Hundreds of customers like and use SCO's Unix products.
Wow, hundreds of customers--McBride admits then that their UNIX business is nearly non-existent then. No wonder lawsuits are the only option they have left for making any money from their SysV code.
Company execs choose their publicity photographs according to the image they want to portray. You get autistic-nerdy Bill Gates, stylish and iconoclastic Jobs, etc. SCO is apparently going for the clean-cut MBA ex-jock appearance, about as far removed from technological competence or engineering as is possible. SCO not only fails to be about technology in practice, they don't even want to appear to be a technology company.
Someone has to "stage manage" the response to SCO's FUD machine, press releases, and outrageous claims, and IBM seems like they are in a good position to do it. After all, McBride and SCO is single-handedly "stage managing" an unwarranted and unfounded attack on open source, Linux, and the GPL. And unlike IBM's response, SCO's claims are pure theater (of the comedy variety), with no legal substance.
Unfortunately, in real life, I suspect the response of the open source community is not actually all that well coordinated, by IBM or anybody, which is why SCO has been able to get away with this nonsense for even this long.
I don't see any problems with corporate influence on Linux in general. The key issue, however, is to be smart and careful about licenses and to keep the agendas and strategies of the corporate contributors in mind. Some companies are trying to push software with proprietary licenses onto free software platforms, other companies pick free software licenses that are not in the best interests of free software.
Sun, for example, is making a big push for Sun Java-derived implementations on Linux, but large parts of that platform are only available as proprietary software from Sun. Making free software dependent on a proprietary implementation is probably not a good idea.
Qt is another example. It actually ships under the GPL, which is a free software license. But in the case of handhelds, Troll Tech's business interests have caused them to adopt a platform that excludes other toolkits from the same platform: they basically want the commercial Linux PDA market to themselves. Even on the desktop, FSF-endorsed projects like Gtk+ are covered by licenses like the LGPL because free software proponents believe that sometimes a more lenient license than the GPL is actually better for achieving the long term goals of the free software movement.
Even when corporate support of open source comes with no strings attached, it may sometimes still not be all that good. IBM has released a lot of software in open source form, software like LVM and JFS. For IBM that's a good deal because it lets them move their AIX customers to Linux. But LVM was rejected by Linux for inclusion into the Linux kernel, and ReiserFS and ext3 are better alternatives to JFS and far more widely used; a lot of corporate contributions just add redundant bulk. And JFS has become the object of SCO legal claims. While those are probably bogus, it shows another way in which accepting corporate contributions too indiscriminately could be a problem.
So, overall, I think corporations have a lot to contribute. But you always have to keep in mind that, even though they may be talking the talk, their interests are not aligned with the goals of the free software movement, or even the goals of the open source software movement.
The Caldera license Parens cites as allowing the use of code in Linux does no such thing, according the FSF. It is similar to the original BSD license, which is NOT GPL-compatible, according to FSF, because of the advertising clause.
Well, then either the authors of the GPL'ed or of the BSD'ed code could make an issue out of this. If they don't make an issue out of it, it's not an issue, and SCO has no legal standing to make it an issue.
He's at Microsoft research. Research labs at wealthy companies tend to do a lot of stuff that's completely unrelated to products. Furthermore, the kind of analysis he does has a long history and it would be surprising if MS didn't have some people working on it.
However, he doesn't seem to be too smart:
It turns out that two-thirds of all threads in Usenet, in 2002, had a whopping two messages. And two-thirds of all authors are the people who write a message, post once one day, and never again.
They probably use a throwaway E-mail address and account. People do it all the time. Why put up with spam for every informative question?
USENET is just not a good source of data for this sort of research; corporate E-mail archives are considerably better. But I suppose Microsoft doesn't keep those anymore, right?:-)
the current implentations of X on a framebuffer with common toolkits it was a tiny bit slower.
Keep in mind that XFree86 is a fairly mediocre implementation and fails to take full advantage of acceleration on a lot of hardware. It's also tuned for desktop machines. Furthermore, many common toolkits for X11 these days (Qt, Gtk+, FLTK, wxWindows) are actually cross-platform toolkits and/or assume a GDI-like model, so they are far less efficient than a high-quality native X11 toolkit would be.
Saying that implementing something for X makes it work for every toolkit is only partially accurate though. Your extention will continue to be X specific, if a user wants to run a PicoGUI or DirectFB application it will have to be rewritten for those environments.
That's because PicoGUI and DirectFB are not toolkits (meaning X11 toolkits), they are entire window systems. Implementing something for X11 also won't give the capability to Windows or MacOS. But implementing something for X11 makes it available to all X11 toolkits, and it often transparently improves all X11 toolkits.
The extention also may not be available automatically to all applications depending on how it was written.
X11 has standard interfaces for everything I mentioned. If a toolkit or application doesn't use them, it's a problem with the toolkit or application. You can lead a horse to water, but you can't make it dring.
Antialiasing support was implemented as an X extension, but applications did not automatically begin using it once it was written.
That was a deliberate choice: it is simply wrong to replace non-anti-aliased rendering with anti-aliased rendering.
In OPIE the developers also had to implement antialiasing for their toolkit, and they did it using the same Freetype engine that X uses for its fonts.
I think X11 kind of threw in the towel on antialiased fonts: I think the renderer should be on the server, not in the client. But between the fact that most users of this feature are Windows refugees who simply assume that fonts are client-side, that TrueType has some serious patent issues, and that Unicode can't get their act together, they didn't have much of a choice but do go this route for now. Still, even with the current implementation, acceleration for rendering is shared, and they can do it right in a few years when the dust has settled.
However, you simply cannot transparently replace a non-anti-aliased system with an anti-aliased system, except for really trivial graphics. So, no matter what they did, clients and applications needed to change. But now that the protocols are in place, any client can take advantage of it on any X11 server implementation that supports it.
The X11 standard clearly isn't perfect, and XFree86 is not the greatest implementation of it. But I think overall, a toolkit together with an X11 server is still the best option for getting a GUI onto almost any device.
I think the biggest room for improvement would be in the toolkit area: the default toolkit for a PDA should probably be designed and implemented from scratch. It shouldn't try to have a cross-platform backend, and it shouldn't try to emulate a desktop API or look-and-feel. Small screens really are different. Using X11 for the back-end makes such efforts quite easy and allows a lot of reuse of coding effort from existing projects (acceleration, window management, rendering, etc.). And using X11 also allows the occasional oddball application to be ported more easily without a total rewrite of the GUI for those who need/want it.
The IBM counter suit I'm not sure about, there patent portfolio is a weapon that could just as easily be turned on us.
IBM's patent portfolio does make me uneasy as well, but software patents are a fact of life. From a practical point of view, IBM is probably one of the best companies in this regard because they actually distribute lots of GPL'ed software. As I understand the GPL, IBM cannot make patent infringement claims related to GPL'ed code they themselves have, at some point, distributed, and unlike other companies, they distribute quite a bit of GPL'ed code.
All SCO has to do is to demonstrate clearly which parts of the Linux kernel they actually hold copyright to and haven't yet licensed under an open source license themselves. Once they do, I'll be happy to come into compliance by removing the offending code from my Linux source tree--based even on SCO's most outrageous claims, I'm not using any of those features.
Overall, I seriously doubt that any judge is going to award damages or even consider something willful copyright infringement until SCO actually identifies the offending code. And then SCO's legal strategy simply falls apart because as soon as they demonstrate any copyright infringement, the offending code will be removed from the kernel instantly.
Sontag warned users that ignoring SCO's requests to license the code in advance of a court case could be costly. "Those who have chosen to ignore the license are more in a situation of potential willful infringement," Sontag said.
The only "willful infringement" here is SCO's infringement of copyrighted Linux source code. SCO has already admitted that they believe that they are not in posession of a valid license to distribute Linux--they claim that the GPL is invalid. Furthermore, SCO has stated publicly that Linux contains non-GPL'ed code, yet they continue to distribute it in violation of the license under which they obtained that code.
On the other hand, Qt/embedded applications are usually a bit more responsive to user input [compared to X applications].
I think that depends entirely on the toolkit; there is nothing in X11 that makes it intrinsically slower than Qt/Embedded.
They make better use of available screen space ( Although themeable X toolkits take care of this). The biggest downside I've seen to X is the multiple toolkits.
People keep making that argument, and I just don't think it holds. If there is a benefit to having a single toolkit, then people will start using only a single toolkit.
In the real world, as soon as a platform becomes successful, most toolkits are ported to it anyway: Gtk+ on Windows, wxWindows on X11, OSX, and Windows, Fltk on pretty much everything, etc. If Qt/Embedded were to take off, all of those would get a Qt/Embedded back-end anyway. The fact that nobody has bothered so far is just a consequence of the fact that Linux PDAs just haven't taken off.
Now whenever the user decides they want to install an app, they have to worry about whether or not they have the space to install the toolkit depenencies.
Yes, and if the install is too big, they just won't use the software. The same is true if I decide, for example, to base some piece of handheld software on Java, Python, or Mono. You can give programmers the means of creating small applications by pre-packaging a "preferred" toolkit with the PDA, but you can't keep developers from screwing themselves or their users.
Everyone is already going to have to significantly rework their applications for the PDA because of the size constraints of both storage and the display.
Sure, but if the PDA runs the same toolkit I am already familiar with from my desktop development, I don't have to learn an entirely new toolkit. I can continue to use the same development and scripting tools (wxWindows, PyGtk, FltkLua, etc.). And often the trickiest part is performance critical low-level graphics code, which I can write to native X11 APIs independent of the toolkit or screen size, but have to completely rewrite for Qt/Embedded.
Furthermore, there are lots of toolkit-independent pieces of functionality that X11 standardizes. For example, I can add a new handwriting recognizer or input method to an X11 server and use it with any toolkit. Work on anti-aliasing, pen input, hardware acceleration, video decoding, and other server functionality is also toolkit indepenent and benefits everybody. In contrast, with an approach like Qt/Embedded, everything has to be redone from scratch for every toolkit.
You seem to be assuming I'm American - I'm not. [...] and why on earth do you get the idea I believe that "people should act against the wrongdoing of foreign governments
Sorry, I got my threads mixed up. My general comments stand though: worry about problems in your own country. Just about every nation has its domestic problems.
As for China, I think people who make comments like yours really need to think things through a bit more. While it is clear that China does not have freedom of expression, that is not the same as having a terror regime that kills people randomly as you seemed to imply. Furthermore, it's far from clear to me that there is anything anybody can do--hastening political change in China might have a greater risk than letting China evolve by itself over this century. It took Europe and the US centuries to transition to democracy, and the transition was quite bloody and protracted. Your knee-jerk condemnation of China just seems unwise to me. Give it some thought...
Slashdot Mirror
yet they're also releasing an open sourced distribution of Linux.
No, they are not: MadHatter contains some seriously proprietary bits, foremost Sun Java, and their Linux distribution probably does as well.
Sun has never been above using open source or free software when its suits them. SunOS and Solaris, after all, were based on a lot of BSD code. But, ultimately, Sun wants to own big chunks of intellectual property in what they are shipping.
I wouldn't want to support someone so wishy washy
I think Sun isn't wishy-washy at all: they know exactly what they are doing. And perhaps that is just the reason why you shouldn't support them.
Kudos to Sun for finally getting the desktop right.
Have you ever actually used any of Sun's user interfaces? SunView, OpenWindows, NeWS, Swing, OpenOffice? I don't think they have ever gotten user interfaces "right", and I seriously doubt they have gotten this one right.
Gnome by itself seems much more consistent and usable than any mix of Java, OpenOffice, and Gnome could ever be. Keep in mind that Java and OpenOffice don't even use the native Gnome toolkit or libraries.
So, let's see, Sun's proposition is that they will "enhance" the Linux desktop by throwing Java in there.
Now, Java is about as proprietary a platform as they come; just to download the specifications, you have to agree to a license that requires any implementation you base on it to pass a Sun conformance test (I'd like to link the license itself, but the Sun site uses some dynamic content that makes that difficult; just click on the download next to the documentation to see the license). If you dare to download the JRE or JDK, the license gets even scarier, with having to donate pretty much everything you do "based" on that to Sun.
Well, it's par for the course, I suppose: the same cast of characters from Sun Microsystems has been trying to replace open UNIX desktop GUIs with something proprietary before, most notably with NeWS. (As an aside, one of the main Java movers and shakers, Gosling, actually sparked the creation of the FSF and GNU Emacs by creating a proprietary version of Emacs.)
Maybe there would be some argument that this is a deal one had to live with if Sun actually had something to offer in terms of user interfaces. But look at what Sun's history of user interfaces: SunView, OpenWindows, and NeWS--not exactly stellar success stories, either in terms of technology or in terms of industry adoption. These days, Sun is shipping Swing, which manages to be bloated, slow, and thoroughly unintegrated with the Gnome or KDE desktops, and OpenOffice, which also manages to be bloated, slow, and thoroughly unintegrated with the Gnome or KDE desktops.
If Sun wants to ship MadHatter, that's their business: Gnome is open source and as long as they comply with the GPL/LGPL, they can do whatever they want. But I think the Linux desktop and UI needs help from Sun about as much as the US needs economic advice from North Vietnam.
For the longest time, many ham radio operators were sticking with old versions of Windows or even DOS. As a result, a lot of the software for interfacing to radios was DOS/Windows only and many companies were unwilling to publish protocols. Hopefully, that is going to change as more hams are moving to Linux.
Companies like Vonage really are only here for a transitional period: they give you a way to connect VoIP service to the regular phone network. That's not a long-term business model.
But for Internet-to-Internet calls, any attempt at regulation would be futile. In fact, there doesn't even have to be any kind of business involved in the middle.
States can, of course, tax IP traffic or Internet access, but regulations that try to distinguish between different uses of that traffic would be very hard and costly to enforce and trivial to work around.
SCO can have the GPL or BSD licenses declared invalid all they want--the code is still copyrighted. With the GPL or BSD in place, at least they have the excuse that "it's free anyway, we just got sloppy about satisfying a few conditions". But if they actually argue that they believe the GPL and BSD license are invalid, then they are committing willful copyright infringement of code they have no license at all for.
Perl packages have been packaged as Debian packages for quite some time--the package and all dependencies are automatically downloaded and they are automatically kept updated.
Shouldn't Netgear be liable for the costs they impose on the university? I mean, Netgear is a money-making enterprise and they chose to point an excessive number of clients at the service, they should pay for the IT staff, hardware, and bandwidth necessary to service their requests.
And, in addition to the monetary aspect, this would seem to fall under computer hacking statutes. I suspect that if this sort of thing had resulted from a student or small developer releasing some software, the FBI might be knocking on their door already, and protestations that it was an accident might not be listened to. I think Morris also claimed that his worm (the first Internet worm) escaped by accident, yet he got charged.
Well, the stupidity of the way their NTP client was written (hard-coded IP address, 1sec queries, etc.) is pretty obvious. And of course, the Netgear technical support response ("we don't have time to respond to your problem, but maybe it has gone away on its own already") is typical of the company.
But what seems particularly stupid is that many companies would jump at the opportunity of having an excuse for obtaining detailed usage information about their customers. They would point these devices at their own servers (probably with 1h rather than 1s query intervals) and know exactly how many of their devices are in operation and what software versions they are running.
That is almost certainly what Microsoft and Apple are using their NTP servers for: keeping track of number of installed systems by the day and by IP address blocks (home users, Fortune-500, etc.).
So, we have to conclude that not only Netgear's programming is iffy and that their customer support is pretty awful, they also fail to behave like a rational US corporation by giving up the opportunity of keeping tabs on their customers. Well, maybe that makes them a little less bad, actually, and at least their gear is cheap and works most of the time...
Carbonate deposits would most likely be found in the basins, but those are also the areas most likely to be filled with dust, lava, and ice. On Earth, those low points with their deposits eventually get lifted up into mountain ranges, which is why they get exposed, but I believe on Mars, that doesn't happen very much. So, the result that we don't see a lot of carbonates isn't all that surprising even if there were moderately sized oceans on Mars a long time ago.
Wow, hundreds of customers--McBride admits then that their UNIX business is nearly non-existent then. No wonder lawsuits are the only option they have left for making any money from their SysV code.
Company execs choose their publicity photographs according to the image they want to portray. You get autistic-nerdy Bill Gates, stylish and iconoclastic Jobs, etc. SCO is apparently going for the clean-cut MBA ex-jock appearance, about as far removed from technological competence or engineering as is possible. SCO not only fails to be about technology in practice, they don't even want to appear to be a technology company.
Someone has to "stage manage" the response to SCO's FUD machine, press releases, and outrageous claims, and IBM seems like they are in a good position to do it. After all, McBride and SCO is single-handedly "stage managing" an unwarranted and unfounded attack on open source, Linux, and the GPL. And unlike IBM's response, SCO's claims are pure theater (of the comedy variety), with no legal substance.
Unfortunately, in real life, I suspect the response of the open source community is not actually all that well coordinated, by IBM or anybody, which is why SCO has been able to get away with this nonsense for even this long.
I don't see any problems with corporate influence on Linux in general. The key issue, however, is to be smart and careful about licenses and to keep the agendas and strategies of the corporate contributors in mind. Some companies are trying to push software with proprietary licenses onto free software platforms, other companies pick free software licenses that are not in the best interests of free software.
Sun, for example, is making a big push for Sun Java-derived implementations on Linux, but large parts of that platform are only available as proprietary software from Sun. Making free software dependent on a proprietary implementation is probably not a good idea.
Qt is another example. It actually ships under the GPL, which is a free software license. But in the case of handhelds, Troll Tech's business interests have caused them to adopt a platform that excludes other toolkits from the same platform: they basically want the commercial Linux PDA market to themselves. Even on the desktop, FSF-endorsed projects like Gtk+ are covered by licenses like the LGPL because free software proponents believe that sometimes a more lenient license than the GPL is actually better for achieving the long term goals of the free software movement.
Even when corporate support of open source comes with no strings attached, it may sometimes still not be all that good. IBM has released a lot of software in open source form, software like LVM and JFS. For IBM that's a good deal because it lets them move their AIX customers to Linux. But LVM was rejected by Linux for inclusion into the Linux kernel, and ReiserFS and ext3 are better alternatives to JFS and far more widely used; a lot of corporate contributions just add redundant bulk. And JFS has become the object of SCO legal claims. While those are probably bogus, it shows another way in which accepting corporate contributions too indiscriminately could be a problem.
So, overall, I think corporations have a lot to contribute. But you always have to keep in mind that, even though they may be talking the talk, their interests are not aligned with the goals of the free software movement, or even the goals of the open source software movement.
I liked the look of the actress who played the doomed queen [...] Much more convincing than Yul Brenner, and a darn sight better looking.
Well, unless Nefertiti was a drag queen, it is perhaps not all that surprising that Yul Brynner didn't make a convincing Nefertiti.
like Gnome.
Such an overdesigned system can withstand quite a lot of punishment. So the idea of a worm bringing down signalling is laughable at best.
The hardware is overdesigned. The software is evidently underdesigned if it can be brought down by a Windows virus.
The Caldera license Parens cites as allowing the use of code in Linux does no such thing, according the FSF. It is similar to the original BSD license, which is NOT GPL-compatible, according to FSF, because of the advertising clause.
Well, then either the authors of the GPL'ed or of the BSD'ed code could make an issue out of this. If they don't make an issue out of it, it's not an issue, and SCO has no legal standing to make it an issue.
He's at Microsoft research. Research labs at wealthy companies tend to do a lot of stuff that's completely unrelated to products. Furthermore, the kind of analysis he does has a long history and it would be surprising if MS didn't have some people working on it.
:-)
However, he doesn't seem to be too smart:
It turns out that two-thirds of all threads in Usenet, in 2002, had a whopping two messages. And two-thirds of all authors are the people who write a message, post once one day, and never again.
They probably use a throwaway E-mail address and account. People do it all the time. Why put up with spam for every informative question?
USENET is just not a good source of data for this sort of research; corporate E-mail archives are considerably better. But I suppose Microsoft doesn't keep those anymore, right?
the current implentations of X on a framebuffer with common toolkits it was a tiny bit slower.
Keep in mind that XFree86 is a fairly mediocre implementation and fails to take full advantage of acceleration on a lot of hardware. It's also tuned for desktop machines. Furthermore, many common toolkits for X11 these days (Qt, Gtk+, FLTK, wxWindows) are actually cross-platform toolkits and/or assume a GDI-like model, so they are far less efficient than a high-quality native X11 toolkit would be.
Saying that implementing something for X makes it work for every toolkit is only partially accurate though. Your extention will continue to be X specific, if a user wants to run a PicoGUI or DirectFB application it will have to be rewritten for those environments.
That's because PicoGUI and DirectFB are not toolkits (meaning X11 toolkits), they are entire window systems. Implementing something for X11 also won't give the capability to Windows or MacOS. But implementing something for X11 makes it available to all X11 toolkits, and it often transparently improves all X11 toolkits.
The extention also may not be available automatically to all applications depending on how it was written.
X11 has standard interfaces for everything I mentioned. If a toolkit or application doesn't use them, it's a problem with the toolkit or application. You can lead a horse to water, but you can't make it dring.
Antialiasing support was implemented as an X extension, but applications did not automatically begin using it once it was written.
That was a deliberate choice: it is simply wrong to replace non-anti-aliased rendering with anti-aliased rendering.
In OPIE the developers also had to implement antialiasing for their toolkit, and they did it using the same Freetype engine that X uses for its fonts.
I think X11 kind of threw in the towel on antialiased fonts: I think the renderer should be on the server, not in the client. But between the fact that most users of this feature are Windows refugees who simply assume that fonts are client-side, that TrueType has some serious patent issues, and that Unicode can't get their act together, they didn't have much of a choice but do go this route for now. Still, even with the current implementation, acceleration for rendering is shared, and they can do it right in a few years when the dust has settled.
However, you simply cannot transparently replace a non-anti-aliased system with an anti-aliased system, except for really trivial graphics. So, no matter what they did, clients and applications needed to change. But now that the protocols are in place, any client can take advantage of it on any X11 server implementation that supports it.
The X11 standard clearly isn't perfect, and XFree86 is not the greatest implementation of it. But I think overall, a toolkit together with an X11 server is still the best option for getting a GUI onto almost any device.
I think the biggest room for improvement would be in the toolkit area: the default toolkit for a PDA should probably be designed and implemented from scratch. It shouldn't try to have a cross-platform backend, and it shouldn't try to emulate a desktop API or look-and-feel. Small screens really are different. Using X11 for the back-end makes such efforts quite easy and allows a lot of reuse of coding effort from existing projects (acceleration, window management, rendering, etc.). And using X11 also allows the occasional oddball application to be ported more easily without a total rewrite of the GUI for those who need/want it.
The IBM counter suit I'm not sure about, there patent portfolio is a weapon that could just as easily be turned on us.
IBM's patent portfolio does make me uneasy as well, but software patents are a fact of life. From a practical point of view, IBM is probably one of the best companies in this regard because they actually distribute lots of GPL'ed software. As I understand the GPL, IBM cannot make patent infringement claims related to GPL'ed code they themselves have, at some point, distributed, and unlike other companies, they distribute quite a bit of GPL'ed code.
All SCO has to do is to demonstrate clearly which parts of the Linux kernel they actually hold copyright to and haven't yet licensed under an open source license themselves. Once they do, I'll be happy to come into compliance by removing the offending code from my Linux source tree--based even on SCO's most outrageous claims, I'm not using any of those features.
Overall, I seriously doubt that any judge is going to award damages or even consider something willful copyright infringement until SCO actually identifies the offending code. And then SCO's legal strategy simply falls apart because as soon as they demonstrate any copyright infringement, the offending code will be removed from the kernel instantly.
Sontag warned users that ignoring SCO's requests to license the code in advance of a court case could be costly. "Those who have chosen to ignore the license are more in a situation of potential willful infringement," Sontag said.
The only "willful infringement" here is SCO's infringement of copyrighted Linux source code. SCO has already admitted that they believe that they are not in posession of a valid license to distribute Linux--they claim that the GPL is invalid. Furthermore, SCO has stated publicly that Linux contains non-GPL'ed code, yet they continue to distribute it in violation of the license under which they obtained that code.
On the other hand, Qt/embedded applications are usually a bit more responsive to user input [compared to X applications].
I think that depends entirely on the toolkit; there is nothing in X11 that makes it intrinsically slower than Qt/Embedded.
They make better use of available screen space ( Although themeable X toolkits take care of this). The biggest downside I've seen to X is the multiple toolkits.
People keep making that argument, and I just don't think it holds. If there is a benefit to having a single toolkit, then people will start using only a single toolkit.
In the real world, as soon as a platform becomes successful, most toolkits are ported to it anyway: Gtk+ on Windows, wxWindows on X11, OSX, and Windows, Fltk on pretty much everything, etc. If Qt/Embedded were to take off, all of those would get a Qt/Embedded back-end anyway. The fact that nobody has bothered so far is just a consequence of the fact that Linux PDAs just haven't taken off.
Now whenever the user decides they want to install an app, they have to worry about whether or not they have the space to install the toolkit depenencies.
Yes, and if the install is too big, they just won't use the software. The same is true if I decide, for example, to base some piece of handheld software on Java, Python, or Mono. You can give programmers the means of creating small applications by pre-packaging a "preferred" toolkit with the PDA, but you can't keep developers from screwing themselves or their users.
Everyone is already going to have to significantly rework their applications for the PDA because of the size constraints of both storage and the display.
Sure, but if the PDA runs the same toolkit I am already familiar with from my desktop development, I don't have to learn an entirely new toolkit. I can continue to use the same development and scripting tools (wxWindows, PyGtk, FltkLua, etc.). And often the trickiest part is performance critical low-level graphics code, which I can write to native X11 APIs independent of the toolkit or screen size, but have to completely rewrite for Qt/Embedded.
Furthermore, there are lots of toolkit-independent pieces of functionality that X11 standardizes. For example, I can add a new handwriting recognizer or input method to an X11 server and use it with any toolkit. Work on anti-aliasing, pen input, hardware acceleration, video decoding, and other server functionality is also toolkit indepenent and benefits everybody. In contrast, with an approach like Qt/Embedded, everything has to be redone from scratch for every toolkit.
You seem to be assuming I'm American - I'm not. [...] and why on earth do you get the idea I believe that "people should act against the wrongdoing of foreign governments
Sorry, I got my threads mixed up. My general comments stand though: worry about problems in your own country. Just about every nation has its domestic problems.
As for China, I think people who make comments like yours really need to think things through a bit more. While it is clear that China does not have freedom of expression, that is not the same as having a terror regime that kills people randomly as you seemed to imply. Furthermore, it's far from clear to me that there is anything anybody can do--hastening political change in China might have a greater risk than letting China evolve by itself over this century. It took Europe and the US centuries to transition to democracy, and the transition was quite bloody and protracted. Your knee-jerk condemnation of China just seems unwise to me. Give it some thought...