Do you make up that sort of thing? Admittedly, XP isn't my main machine, but I have the misfortune of having to transfer data with some regularity, and I can assure you that I can just plug and unplug flash devices and it just works. There is no data loss if you wait for the blinkenlights to stop flashing, and there are no complaints or dialog boxes. And it's the same on my Linux machine (where you can configure automount to mount and unmount the flash drive as it is being used).
It's a marketing-desicion to keep the user tied to iTunes. Is this news to you?
Yes, this is news to me. It's not a nice thing for Apple to remove useful functionality from their iPod in order to tie people to their music store.
I remember when I copied some files to my USB-thumbdrive on XP. I copied the files, and simply removed the drive from the computer. When I got back home, I noticed that no files were in fact copied.
The same thing can happen to you on Macintosh, since nothing prevents you from unplugging the device too soon and you may not even see the dialog box after the fact. Furthermore, most of the time, the dialog box appears even though, in fact, the data is fine, and the dialog box gives you no way of fixing the problem anyway.
There is a reason why we "safely remove" removable media. We do it on Macs, Windows and Linux. What "other platforms" are you talking about, since I run in to this issue on all three platforms? All three systems tell me to "safely remove" the media, instead of simply unplugging it.
The correct thing to do is to tell the user when the device is in use and to finish with it as soon as possible. When there are no more operations pending, the user should be able to unplug it. For practical purposes, that's the way it works on Windows. That's also the way it works on many Linux systems (e.g., with automount and a short timeout). If your OS has reasonable write semantics, you can watch the blinking light on the flash drive to guess a good time to remove the device.
It is never correct to throw up big dialog boxes with error messages in a foolish attempt to punish the user when, in fact, no error occurred, and that's what Macintosh does.
This sort of thing isn't rocket science, and the Mac just gets it wrong. Apple has always thought they could control device removal perfectly, but they can't. What they need to do instead is improve the UI to reduce the risk of bad device removals and improve the OS to reduce the consequences of bad device removals.
On the other hand, where did you get the idea that other systems allow you to unplug removable media without unmounting?
Because they do and because it works.
Or let me rephrase that: just as on Windows or on Linux, on MacOS X you can unplug the device and in more than 99% of the times absolutely nothing bad is going to happen. But some systems (including MacOS X, and properly configured Windows 2000/XP or Linux) have the courtesy to remind you that it is not worth the risk, even though admittedly it is very, very low.
Your mental model of how this works is wrong. Users don't pull flash devices at random points in time, they pull flash devices when they think they are finished writing to them.
Well-designed systems indicate when it is not safe to pull the devices (both on-screen and through a light on the device), and when they are done writing, they make it clear that it is now safe to pull the device. So, you watch the indicator and pull the device when it's safe to do so. On the rare occasion you screw up, they give you an error message, but only if something bad actually happened.
Macintosh doesn't give you a clear indication of when it is unsafe to pull the device. In fact, it's so bad that you can't even unplug devices when the system is suspended; Macintosh doesn't seem to unmount removable devices on suspend. On the other hand, Macintosh always complains about pulling the device even if there were no bad consequences, apparently in some mistaken attempt to train the user. And Macintosh doesn't prevent bad device removals either, so the risk of data loss is at least as large on Macintosh.
What should Apple do? -- They should make sure at the system level that writes to flash devices are more synchronous with application-level operations. -- They should indicate on the desktop icon that a flash device is in use and has pending writes to it. -- They should make sure that they write to flash devices in such a way that their indicator lights blink as long as there are pending writes. -- They should get the OS into a state that it is safe to remove a device when all operations have completed on it. -- They should put up an error message only if device removal has actually caused an incomplete write, and it should be an informative message (i.e., not require acknowledgement). -- And they should change the file system code for flash devices so as to minimize the damage caused by early removal of flash devices (this may mean having two VFAT drivers--one for non-removable devices and one for removable devices). (Some of these things are not yet implemented in Windows or Linux either, but those other systems are closer.)
Well, let's hope that there weren't any pending operations in the filesystem when you pull your device.
I don't have to "hope", I know that there aren't any pending operations when I pull the device; that's what those big dialog boxes that are telling me that there are pending operations are for, as well as the blinking light on the device. When it is clear that there are no more pending operations, then I pull the device.
Especially if your device is formatted with a non-journaling filesystem! Say good-bye to data integrity.
There is no reason why there should be any more serious consequence of pulling out the device too early than a single truncated file. Furthermore, the correct thing to do is for the system to put up an informative popup saying "I didn't complete writing file XYZ or subsequent files; if you would like to continue, please reinsert the device. If you don't care, just don't worry about it."
I hope there haven't been too many times when you've lost data and blamed it on the hardware (or you haven't discovered it yet).
The only place I have ever lost data from removable devices was on the Macintosh because the Macintosh can't prevent me from removing the device either, but it doesn't actually give me a clear indication of when it is safe to remove. Even worse, afterwards, it doesn't even give me correct information about whether anything actually went wrong or what happend. It just whines and complains about what an ill-behaved user I am. This is really something where Apple screwed up big time, and they should fix it.
First of all, their product is really Open Source and it's the same version that is sold, not some dumbed down version.
First of all, that's not true. Troll Tech does not release all of the Qt implementations under GPL. Furthermore, when I buy a device running Qt/Embedded, I don't get the source code because Troll Tech sold the source code to the device manufacturer under the commercial license. That means that I can't actually modify how the thing is running on the device.
But more importantly, the project may be released under an open source license, but that doesn't make it an "open source project"; the power of open source and free software derives not from the license (which is a means to an end), but from the way those projects are run and what guarantees I get as a user and contributor. Open source projects accept and incorporate contributions from the community, and all developers on the project are on equal footing.
Moreover, Trolltech employees devote part of their paid time to work on other Open Source or Free Software projects. KDE is of course the prime example, but you don't have to go far: just think of EXA (X.org project).
That's philanthropy; it is philanthropy direct at open source, but it isn't relevant to the status of Qt.
And more importantly, they cannot stop releasing the GPLed version because of the Free Qt Foundation: if they don't release an Open Source version in a 12-month period, the last released version is automatically relicensed BSD.
Troll Tech chose those rules, they have to live with the consequences. It doesn't change the fact that (in my opinion) the right thing for them to do is stop dual-licensing and deliver an honest, commercial product.
And I don't see a problem with that consequence anyway. If Troll Tech really contributes value to Qt above and beyond what open source developer do, then within a short period of time, Troll Tech's version of Qt should be very competitive with the BSD version again.
From the brief description, I don't see anything in Flock that doesn't already exist as Firefox extensions. In fact, I find it hard to imagine a feature that couldn't be provided as a Firefox extensions, since Firefox is basically just a set of "extensions" running in a generic XUL runtime.
Does anybody know any technical reason why this needs to be a new browser, instead of being Firefox shipped with a bunch of pre-installed add-ons?
Ethiopia was a colony for only 7 years (1936-1943 under Italy). Maybe, just maybe, over 60 years after that 7-year period it's time to stop waiting for handouts and start to solve the problems themselves.
Maybe you should just stop speaking in such grand generalities and be specific. Who exactly do you think "is waiting for handouts"? Are you saying that the average Ethiopian is sitting at home wondering when the next check from their European sugar daddies is arriving? Are you suggesting that the average Ethiopian is lazier than the average American? Or what are you actually saying?
So your claim that it takes longer than half a century is just plain wrong. It takes one human generation to develop an industry (like in today's China) and 2 generations to generate wealth and luxury similar to western standards (like in Taiwan, Singapore, South Korea and Japan today).
China, Japan, and Korea were effectively developed nations already, with most of the social, political, administrative, and economic infrastructure in place.
As for Ethiopia, the fact that it was a monarchy until 1974 (with the interruption of Italian occupation), afterwards was ruled by a military junta, and has been mired in several wars and internal conflicts probably has something to do with its problems. Hard as that may be for you to understand, such political circumstances make it difficult for a nation to develop economically.
As both the US and Europe have shown, it can take centuries to develop and industrialize. And even deliberately planned economic development often fails; Germany has found it impossible to bring former East Germany up to West German standards and is resigned to taking decades longer to do so, and the US is full of regions and social groups that are bigger than entire developing nations and economically just as troubled.
Before the iPod Apple made a stir with OS X, which was certainly not just some stylized rehash of any PC stuff.
OS X was an updated version of NeXTStep. NeXTStep itself was a combination of Mach (CMU), Objective-C (Stepstone), Postscript (Adobe and Xerox), and Smalltalk (Xerox) technologies. So, Apple made a stir with it, but they didn't develop most of it.
The truth is since Steve returned with his NeXt compatriots Apple's been churning out lot of fantastic new products.
Yes, but... it's important not to confuse technical innovation with design and marketing. Apple occasionally brings out a technically innovative product, but a lot of their products are merely better designed and engineered versions of products pioneered by others.
Have you looked a Aperture which Apple announced yesterday? That product (unless there's some serious hidden bug in there somewhere) will totally rock the digital photography world. It's exactly the type of tool that everyones wanted and nobody made... Photoshop was resting on it market place domination and everyone else was trying to copy Photoshop.
Agreed--Aperture looks good, and Adobe was resting on their laurels. But with that statement, you yourself are saying that Apple isn't actually breaking new ground, but instead are implementing a checklist of features that photographers have been asking for (in fact, there have been several articles by photographers asking pretty specifically for someone to develop just this tool).
It will be interesting to see whether they can pull it off; Adobe has weathered other challenges. And although Apple took the initiative, you are likely going to see the GUI copied both by commercial and open source packages--and Apple really has no reason to complain about that.
The business opportunities for cross-platform development astound me when I sit here thinking about it. I'm just shocked no one else has picked up on this yet.
Well, given that good cross platform toolkits have been around for 20 years (some of them far better than Qt), obviously, there is something wrong with the concept.
Now: Support. Who here has supported a full linux network? Those of you who have never had the pleasure, let me tell you something: It's pure bliss.
When I'm an ISV and I sell software for Windows, Macintosh, and Linux, then I must support the software on all three platforms. That means phone staff, procedures, testing, upgrades, installers, tech writers, documentation, and a lot of other people and subprojects for each platform. When I hire all those people, I might as well hire someone to do a native port to OS X because the cost of doing that is negligible in comparison.
So, we just saved our Dr at least $6,000 is start up fees. What does this mean to us, the software developer?
Sounds like you are talking about in-house custom software development. Yes, that's a niche market where cross-platform toolkits sometimes make sense. Those kinds of applications require excellent tool support and very high level languages and toolkits. Custom software is usually done in Visual Basic, and now more in.NET. If you need cross-platform features, people use Real Basic or Java or some more specialized environments; Qt and C++ are completely the wrong choice.
On my iPod Mini, the songs are automatically synced the moment I plug it in the computer. No need to drag files around.
Yes, that works when you plug your own MP3 player into your own properly configured Macintosh running iTunes. That's one simple usage scenario.
But lots of people are using their MP3 players differently: they are moving them between many machines, they use different music management software, they copy clips from friends, etc. Those other players still synchronize with iTunes, but they do things like on-device playlist generation and bookmarking, and they are no harder to use than iPod either.
There is no technical or user interface reason why iPod functionality is so limited. Either Apple doesn't have the software development resources to add these functions, or it's a deliberate marketing decision by them to keep you tied to iTunes.
In the end, the iPods are nice looking, reasonably priced players that integrate very well with iTunes, but the fact remains: their on-board software is limited compared to what's out there.
How else would you like to do it?
I want to just unplug it when I'm done. That works on other platforms, why not on Macintosh? It's the obvious and simple thing to do. And don't tell me that it might "interrupt a write"; other platforms deal with that without problems.
Have you tried to remove a removable device on a PC (say, an USB memory stick)? You need to go to a tiny icon on the task bar, use a contextual menu, and two crazy screens show up with non-intuitive options, etc. Ridiculous.
Bullshit. On Windows, you just unplug the thing when you're done. And on a correctly set up Linux machine, you also just unplug the thing.
Colonialism ended less than half a century ago; it takes a lot longer to develop even under ideal conditions.
And conditions are hardly ideal: Africa's most natural exports are heavily disadvantaged by Western subsidies, and economic exploitation of Africa and propping up of undemocratic regimes in Africa by powerful nations continues to this day. Even our so-called economic aid is usually tied to specific purchases from the donor nations, so it isn't very effective, and what isn't tied up that way disappears in corrupt regimes, usually with knowledge of the Western donors.
Depends on the real world. To me this seems a slam dunk: For $2,000, I get an app that will run on multiple OSes, considerably broadening my potential customer base. And it'll only add a few bucks to the end product.
Cross-platform support at the toolkit level is uninteresting for most companies; if you are serious about cross-platform, then you have to put in place an entire infrastructure for supporting the new platform (help desk, tech writing, marketing, etc.). At that point, you might as well create a new native UI, which is quick and easy given modern tools, and which will be better quality than anything you do with a cross-platform toolkit.
Only a PHB would deny that kind of request, and they wouldn't even do that if you dropped the words, "Mac OSX".
Non-Macintosh software houses don't generally give a damn about Macintosh (not worth the extra trouble and expense), and Macintosh software houses are disdainful of Windows.
Overall, cross platform support is a small niche market.
To me, video on these devices is a gimmick. A useful application would be photo storage. But they turn out to be rather poor for photo storage, even though Apple offers them for that purpose: the camera/flash-to-iPod transfer is so slow and power hungry (according to reviews) that the battery runs out before transferring a single large card, and Apple doesn't even list any compatible CF readers. So, I ended up with another photo storage device.
The iPods are OK MP3 players, but even as MP3 players, they are mediocre in terms of features and functionality. Other MP3 players fully integrate disk, photo, and music player functions--you copy files to the device and they are accessible. With iPod, you must use a special application. And the fact that after all these years, Apple still hasn't figured out how to let you safely remove removable devices without dragging them into the trash can first is sad.
All of this is unfortunate, because the devices look nice, have nice screens, have a generally good UI, and have good sound quality.
You can make a sales pitch for how the extra $2000 don't matter and how good Qt is until you're blue in the face, the fact is people aren't buying it in large numbers. And if you think that you can present those kinds of considerations to justify $2000 in a real business, you haven't encountered real-world corporate budgeting.
As for Troll Tech's right to make money, nobody is disputing that; if they just offered Qt as a commercial, closed source product, we wouldn't even be discussing them here. Nobody has been complaining about the fact that XVT is closed source, for example. What I and others don't like is their marketing, which relies heavily on using the open source community to generate demand for their proprietary product.
Since most of those on slashdot are advocates of free software of course any company choosing to sell a product, try to be successful and employ people (which drives our economy) is going to be the bad guy.
Well, I can't speak for all of Slashdot. I have no problem with a company selling closed source software. What I fault companies for is if they use open source or free software as a marketing gimmick but are actually pursuing goals and strategies that run counter to the goals of open source and/or free software. I think that's what Troll Tech is doing: they are using KDE and the GPL as a marketing gimmick for a product that would otherwise have disappeared from the market long ago.
What I think Troll Tech ought to do is make all future releases of Qt closed source. KDE could go on with the GPL'ed version of Qt, so nobody would be harmed, and the conflict of interest in the development of Qt that the dual license represents would disappear.
And, whatever I think is the right thing to do, I guarantee you that this problem is not going away, no matter how much the KDE project may want it to. KDE could have taken over the Linux desktop long ago (it has really consistently been ahead of Gnome), but as long as this issue persists, that just is not going to happen, no matter how good KDE may get technically.
What "rights" do they retain, any other Free Software developer gives away?
Troll Tech retains all rights to the software; in particular, they retain the right to sell it commercially under non-free licenses. That is very different from most other GPL'ed projects, where the project accepts contributions under the GPL and nobody has the right to sell a non-free version.
Their copyrights?
That is exactly the problem: all of Qt is under their copyright. When the community contributes value to Qt (and they do), Troll Tech arranges for the resulting enhancements to Qt to fall under their own copyright or are shipped separately. That is very different from most other GPL'ed software projects (e.g., the Linux kernel, Emacs, etc.), where often dozens of different people hold the copyright on various parts of the source.
Multiple ownership of the source code of an open source project is good because it pretty much guarantees that the project will remain open and that the project will respond to the needs of its user community. When the entire codebase is owned by a single company, then the project will be driven by the commercial interests of that company, whether or not they also happen to ship a GPL'ed version.
I have a Zaurus SL-C1000 and 760, and also have had a 5500. I do understand what you're saying about Qt/Embedded (now just called qtopia), but it's horses for courses: an X11 handheld device isn't what everyone wants...
If Qtopia ran on the Zaurus through X11, it would be indistiguishable to users from the way it is running now. It may be that Troll Tech originally really did think that Qt/Embedded could be more efficient than Qt/X11, but they were wrong, and at this point, the only reason for Qt/Embedded's existence is that it is great for Troll Tech's business.
Qt/Embedded (now just called qtopia)
Qt/Embedded is the toolkit, Qtopia is the embedded toolkit plus a set of standard applications. (At least that was the case last I looked.)
[PdaXRom, X/Qt,...]
Those are workarounds for nerds, not something you can use for mainstream application delivery. If I want to deliver an open source application on Qtopia to mainstream users, I have to write it in Qt. And if I want to deliver a commercial application on Qtopia to mainstream users, I have to license Qt, overpriced and cumbersome that it is. And Troll Tech knows this; it's great for their business, it just happens to be lousy for the adoption of embedded Linux.
I disagree that the apps aren't optimised for small screens: to me, it's an excellent UI. Opie seems to look exactly the same, surely imitation is the sincerest form of flattery;-)
Well, I can't comment on Opie, and Opie isn't exactly mainstream anyway. The point is, we don't have a choice at this point: almost all embedded Linux devices run Qt, and you simply can't realistically develop and deploy other toolkits on it.
Yeah, releasing kick-ass software under the GPL is a clear threat to Free Software. Seriously: are you a retard?
So, your claim rests on your assumption that it's self-evident that dual licensing is good for free software. But it isn't self-evident to me, nor is it to lots of other developers. In fact, I think Troll Tech's licensing scheme is harmful for free software, and I have given you some reasons for that.
Your insults and flames don't substantiate your arguments. And given that the KDE project was running for a couple of years without even noticing that the GPL and QPL were incompatible and only avoided shut-down because Troll Tech finally gave in and changed their license, KDE proponents like you have no credibility when it comes to judging licensing issues--you got it badly wrong before, and it's pretty clear you are still getting it wrong this time around.
And if you hope that this issue will just go away, that people like me will just shut up and develop for Qt if you only insult us enough, and that and KDE will take over the world, well, just keep on hoping.
It makes no difference to me as a customer how much it costs to develop the library. There are plenty of good alternatives that I don't have to pay for (beyond what I already pay for the OS in any case): Swing, wxWidgets, Gtk+, Cocoa, MFC,.NET, etc. It's unfortunate for Troll Tech that they can't subsidize their development with other revenue streams, but that doesn't make me any more inclined to pay lots of money for their stuff.
Furthermore, if I'm going to pay $1800/developer, then I'm going to pay that money for improving an open source toolkit like wxWidgets or Gtk+, where I get specific improvements for my money and where I don't have to pay over and over again.
So wait, your biggest problem with trolltech is they don't have a price point for you?
No, the biggest problem is that Troll Tech is using open source licensing as a gimmick to push a commercial piece of software at an inflated price point. Without the open source marketing, Qt wouldn't have a prayer in the free market because it's overpriced for what it does.
And it is a commercial software project: it is run like one, it is sold under a commercial license, and Troll Tech retains all the rights.
It's their software, right? They have the right to make money off of their own work, right? And license it however they choose?
Yes, they have the right to offer it under whatever license they like. And we have the right to warn people against using it, and to expose what their approach to doing business means to FOSS in the long run.
In short: what is the "issue" with Qt? the fact that it licensed under the GPL. And why is that a problem? Because it makes it difficult to write proprietary software using Qt (you have to pay TT in order to do so).
The issue with Qt is not that it is covered by the GPL, the issue with Qt is that it is covered also by a commercial license. And the problem with that is that it makes it easy to write proprietary software using Qt--all you have to do is fork over some money to Troll Tech. Troll Tech's dual licensing scheme inhibits most of the things that make open source good: free sharing of ideas and improvements, communal development, and the ability of different forks to compete on merit alone.
Troll Tech is using open source as a marketing gimmick to sell proprietary software. They don't want open source to succeed because it would kill them as a commercial company; they just want open source to succeed enough to make their business model work.
What is the problem here? TrollTech offers their product under the GPL. They also offer it under a proprietary license. They don't force anyone to use their toolkit,
It's good that you recognize that you aren't forced to use it. So, I'm suggesting to you: don't use it.
Qt is licensed under the GPL. I really fail to see how they could "monopolize" anything. or are you worried what would happen if Linux "monopolized" the OS-market?
Software built on Linux doesn't need to fall under the GPL, but software built on Qt does.
And if Linux were dual licensed, I would be worried about it monopolizing the market. But Linux is covered only by the GPL, and that makes it OK.
So I shouldn't offer any bug-reports to the kernel-folks, because that might make the product a bit better, and some company might earn some money through it?
You should offer bug reports to the Linux kernel folks; the Linux kernel project is not dual-licensed. All contributors and all users of the Linux kernel use it under the same conditions.
People are complaining when some company offerws kick-ass software under the GPL?
Yes, people are complaining about that, and rightfully so. Not everything that is offered "free" (even "free-as-in-speech") is a good thing.
For example, Joe might give you free cocaine, and you might enjoy it, but that wouldn't make it a good thing. And Joe would give you free cocaine because he knows that there is a good chance that you get addicted to it and have to come back to him later and pay lots of money for it. That's roughly the kind of strategy Troll Tech uses with Qt.
GTK+ is controlled by the GTK+ team. GNOME is controlled by the GNOME-team. Mozilla and Firefox are controlled by the Mozilla Foundation. So why exactly are you singling out TrollTech and Qt here?
The Gtk+ team or Gnome foundation doesn't have any special rights over the Gtk+ code base--they can't make a commercial version. Troll Tech, however, has special rights over the Qt code base, they can make a commercial version, and they have. Troll Tech is using an open source license to promote their proprietary software.
Oh I see. the whole point of creating free software is to suck up to creators of proprietary software!
Apparently, that is what you believe, because that is exactly what proponents of Qt are doing: they are using an open source licensing gimmick to promote Troll Tech's business and help them make and sell proprietary software.
I heard that Microsoft is hiring. People like you would fit right in there!
Actually, it is your and Troll Tech's mindset that is disturbingly close to Microsoft's; you are promoting proprietary software, you are simply using open source to do it. I heard that Microsoft is hiring--why don't you apply yourself? You apparently put expediency and the appearance of freedom ahead of actual freedom and open source principles, so you should fit right in at Microsoft.
TT wrote the code, therefore they get to choose the license. If you find their license unacceptable [...] use something else, and quit your whining!
The fact that Troll Tech may choose whatever license they like doesn't place them above criticism: their policies are a threat to free software and open source software, and, of course, I am going to continue to criticise them for them.
And, of course, I'm using something else. In fact, I'm contributing to open source desktop and toolkit projects. Troll Tech must not be allowed to succeed.
I haven't seen the patent first hand, but often something obvious now, wasn't at that time. If it were then why wasn't someone else doing it already.
Simple: because the market demand wasn't there yet; pay-as-you-go only became interesting for wireless once companies had actually taken care of all the customers willing to go with what amounts to higher priced plans.
You should not be able to patent something that is merely an application of standard principles by someone of average skill in the profession. Pay-as-you-go billing clearly fits into that category, since it had been used with many other similar services (e.g., prepaid long-distance).
I can list many examples of this. The mouse, keyboard, screens, printers, windowing environment, The Internet, an Operating system and even a CPU and the IC chips, were at the time major conceptual steps forward. I can't tell you how hard it was to explain what the Internet or even a Network was to people in 1983, they just couldn't grasp it.
Well, which only goes to show that your judgement in these matters is impaired. The Internet was developed in the 1960's; by the 1980's, any patents covering the original ideas were reaching the end of their life, and the concepts and technologies were widely understood by people in the profession. In fact, the Internet and what people do with it really hasn't changed all that much in the last 20 years, except that a whole lot more people/companies are using it and that it has gotten a lot faster.
And I suspect that many of the other items you list probably weren't invented in the form and by the people you think invented them either.
Slashdot Mirror
Do you make up that sort of thing? Admittedly, XP isn't my main machine, but I have the misfortune of having to transfer data with some regularity, and I can assure you that I can just plug and unplug flash devices and it just works. There is no data loss if you wait for the blinkenlights to stop flashing, and there are no complaints or dialog boxes. And it's the same on my Linux machine (where you can configure automount to mount and unmount the flash drive as it is being used).
It's a marketing-desicion to keep the user tied to iTunes. Is this news to you?
Yes, this is news to me. It's not a nice thing for Apple to remove useful functionality from their iPod in order to tie people to their music store.
I remember when I copied some files to my USB-thumbdrive on XP. I copied the files, and simply removed the drive from the computer. When I got back home, I noticed that no files were in fact copied.
The same thing can happen to you on Macintosh, since nothing prevents you from unplugging the device too soon and you may not even see the dialog box after the fact. Furthermore, most of the time, the dialog box appears even though, in fact, the data is fine, and the dialog box gives you no way of fixing the problem anyway.
There is a reason why we "safely remove" removable media. We do it on Macs, Windows and Linux. What "other platforms" are you talking about, since I run in to this issue on all three platforms? All three systems tell me to "safely remove" the media, instead of simply unplugging it.
The correct thing to do is to tell the user when the device is in use and to finish with it as soon as possible. When there are no more operations pending, the user should be able to unplug it. For practical purposes, that's the way it works on Windows. That's also the way it works on many Linux systems (e.g., with automount and a short timeout). If your OS has reasonable write semantics, you can watch the blinking light on the flash drive to guess a good time to remove the device.
It is never correct to throw up big dialog boxes with error messages in a foolish attempt to punish the user when, in fact, no error occurred, and that's what Macintosh does.
This sort of thing isn't rocket science, and the Mac just gets it wrong. Apple has always thought they could control device removal perfectly, but they can't. What they need to do instead is improve the UI to reduce the risk of bad device removals and improve the OS to reduce the consequences of bad device removals.
On the other hand, where did you get the idea that other systems allow you to unplug removable media without unmounting?
Because they do and because it works.
Or let me rephrase that: just as on Windows or on Linux, on MacOS X you can unplug the device and in more than 99% of the times absolutely nothing bad is going to happen. But some systems (including MacOS X, and properly configured Windows 2000/XP or Linux) have the courtesy to remind you that it is not worth the risk, even though admittedly it is very, very low.
Your mental model of how this works is wrong. Users don't pull flash devices at random points in time, they pull flash devices when they think they are finished writing to them.
Well-designed systems indicate when it is not safe to pull the devices (both on-screen and through a light on the device), and when they are done writing, they make it clear that it is now safe to pull the device. So, you watch the indicator and pull the device when it's safe to do so. On the rare occasion you screw up, they give you an error message, but only if something bad actually happened.
Macintosh doesn't give you a clear indication of when it is unsafe to pull the device. In fact, it's so bad that you can't even unplug devices when the system is suspended; Macintosh doesn't seem to unmount removable devices on suspend. On the other hand, Macintosh always complains about pulling the device even if there were no bad consequences, apparently in some mistaken attempt to train the user. And Macintosh doesn't prevent bad device removals either, so the risk of data loss is at least as large on Macintosh.
What should Apple do?
-- They should make sure at the system level that writes to flash devices are more synchronous with application-level operations.
-- They should indicate on the desktop icon that a flash device is in use and has pending writes to it.
-- They should make sure that they write to flash devices in such a way that their indicator lights blink as long as there are pending writes.
-- They should get the OS into a state that it is safe to remove a device when all operations have completed on it.
-- They should put up an error message only if device removal has actually caused an incomplete write, and it should be an informative message (i.e., not require acknowledgement).
-- And they should change the file system code for flash devices so as to minimize the damage caused by early removal of flash devices (this may mean having two VFAT drivers--one for non-removable devices and one for removable devices).
(Some of these things are not yet implemented in Windows or Linux either, but those other systems are closer.)
Well, let's hope that there weren't any pending operations in the filesystem when you pull your device.
I don't have to "hope", I know that there aren't any pending operations when I pull the device; that's what those big dialog boxes that are telling me that there are pending operations are for, as well as the blinking light on the device. When it is clear that there are no more pending operations, then I pull the device.
Especially if your device is formatted with a non-journaling filesystem! Say good-bye to data integrity.
There is no reason why there should be any more serious consequence of pulling out the device too early than a single truncated file. Furthermore, the correct thing to do is for the system to put up an informative popup saying "I didn't complete writing file XYZ or subsequent files; if you would like to continue, please reinsert the device. If you don't care, just don't worry about it."
I hope there haven't been too many times when you've lost data and blamed it on the hardware (or you haven't discovered it yet).
The only place I have ever lost data from removable devices was on the Macintosh because the Macintosh can't prevent me from removing the device either, but it doesn't actually give me a clear indication of when it is safe to remove. Even worse, afterwards, it doesn't even give me correct information about whether anything actually went wrong or what happend. It just whines and complains about what an ill-behaved user I am. This is really something where Apple screwed up big time, and they should fix it.
First of all, their product is really Open Source and it's the same version that is sold, not some dumbed down version.
First of all, that's not true. Troll Tech does not release all of the Qt implementations under GPL. Furthermore, when I buy a device running Qt/Embedded, I don't get the source code because Troll Tech sold the source code to the device manufacturer under the commercial license. That means that I can't actually modify how the thing is running on the device.
But more importantly, the project may be released under an open source license, but that doesn't make it an "open source project"; the power of open source and free software derives not from the license (which is a means to an end), but from the way those projects are run and what guarantees I get as a user and contributor. Open source projects accept and incorporate contributions from the community, and all developers on the project are on equal footing.
Moreover, Trolltech employees devote part of their paid time to work on other Open Source or Free Software projects. KDE is of course the prime example, but you don't have to go far: just think of EXA (X.org project).
That's philanthropy; it is philanthropy direct at open source, but it isn't relevant to the status of Qt.
And more importantly, they cannot stop releasing the GPLed version because of the Free Qt Foundation: if they don't release an Open Source version in a 12-month period, the last released version is automatically relicensed BSD.
Troll Tech chose those rules, they have to live with the consequences. It doesn't change the fact that (in my opinion) the right thing for them to do is stop dual-licensing and deliver an honest, commercial product.
And I don't see a problem with that consequence anyway. If Troll Tech really contributes value to Qt above and beyond what open source developer do, then within a short period of time, Troll Tech's version of Qt should be very competitive with the BSD version again.
From the brief description, I don't see anything in Flock that doesn't already exist as Firefox extensions. In fact, I find it hard to imagine a feature that couldn't be provided as a Firefox extensions, since Firefox is basically just a set of "extensions" running in a generic XUL runtime.
Does anybody know any technical reason why this needs to be a new browser, instead of being Firefox shipped with a bunch of pre-installed add-ons?
Ethiopia was a colony for only 7 years (1936-1943 under Italy). Maybe, just maybe, over 60 years after that 7-year period it's time to stop waiting for handouts and start to solve the problems themselves.
Maybe you should just stop speaking in such grand generalities and be specific. Who exactly do you think "is waiting for handouts"? Are you saying that the average Ethiopian is sitting at home wondering when the next check from their European sugar daddies is arriving? Are you suggesting that the average Ethiopian is lazier than the average American? Or what are you actually saying?
So your claim that it takes longer than half a century is just plain wrong. It takes one human generation to develop an industry (like in today's China) and 2 generations to generate wealth and luxury similar to western standards (like in Taiwan, Singapore, South Korea and Japan today).
China, Japan, and Korea were effectively developed nations already, with most of the social, political, administrative, and economic infrastructure in place.
As for Ethiopia, the fact that it was a monarchy until 1974 (with the interruption of Italian occupation), afterwards was ruled by a military junta, and has been mired in several wars and internal conflicts probably has something to do with its problems. Hard as that may be for you to understand, such political circumstances make it difficult for a nation to develop economically.
As both the US and Europe have shown, it can take centuries to develop and industrialize. And even deliberately planned economic development often fails; Germany has found it impossible to bring former East Germany up to West German standards and is resigned to taking decades longer to do so, and the US is full of regions and social groups that are bigger than entire developing nations and economically just as troubled.
Before the iPod Apple made a stir with OS X, which was certainly not just some stylized rehash of any PC stuff.
OS X was an updated version of NeXTStep. NeXTStep itself was a combination of Mach (CMU), Objective-C (Stepstone), Postscript (Adobe and Xerox), and Smalltalk (Xerox) technologies. So, Apple made a stir with it, but they didn't develop most of it.
The truth is since Steve returned with his NeXt compatriots Apple's been churning out lot of fantastic new products.
Yes, but... it's important not to confuse technical innovation with design and marketing. Apple occasionally brings out a technically innovative product, but a lot of their products are merely better designed and engineered versions of products pioneered by others.
Have you looked a Aperture which Apple announced yesterday? That product (unless there's some serious hidden bug in there somewhere) will totally rock the digital photography world. It's exactly the type of tool that everyones wanted and nobody made... Photoshop was resting on it market place domination and everyone else was trying to copy Photoshop.
Agreed--Aperture looks good, and Adobe was resting on their laurels. But with that statement, you yourself are saying that Apple isn't actually breaking new ground, but instead are implementing a checklist of features that photographers have been asking for (in fact, there have been several articles by photographers asking pretty specifically for someone to develop just this tool).
It will be interesting to see whether they can pull it off; Adobe has weathered other challenges. And although Apple took the initiative, you are likely going to see the GUI copied both by commercial and open source packages--and Apple really has no reason to complain about that.
The business opportunities for cross-platform development astound me when I sit here thinking about it. I'm just shocked no one else has picked up on this yet.
.NET. If you need cross-platform features, people use Real Basic or Java or some more specialized environments; Qt and C++ are completely the wrong choice.
Well, given that good cross platform toolkits have been around for 20 years (some of them far better than Qt), obviously, there is something wrong with the concept.
Now: Support. Who here has supported a full linux network? Those of you who have never had the pleasure, let me tell you something: It's pure bliss.
When I'm an ISV and I sell software for Windows, Macintosh, and Linux, then I must support the software on all three platforms. That means phone staff, procedures, testing, upgrades, installers, tech writers, documentation, and a lot of other people and subprojects for each platform. When I hire all those people, I might as well hire someone to do a native port to OS X because the cost of doing that is negligible in comparison.
So, we just saved our Dr at least $6,000 is start up fees. What does this mean to us, the software developer?
Sounds like you are talking about in-house custom software development. Yes, that's a niche market where cross-platform toolkits sometimes make sense. Those kinds of applications require excellent tool support and very high level languages and toolkits. Custom software is usually done in Visual Basic, and now more in
On my iPod Mini, the songs are automatically synced the moment I plug it in the computer. No need to drag files around.
Yes, that works when you plug your own MP3 player into your own properly configured Macintosh running iTunes. That's one simple usage scenario.
But lots of people are using their MP3 players differently: they are moving them between many machines, they use different music management software, they copy clips from friends, etc. Those other players still synchronize with iTunes, but they do things like on-device playlist generation and bookmarking, and they are no harder to use than iPod either.
There is no technical or user interface reason why iPod functionality is so limited. Either Apple doesn't have the software development resources to add these functions, or it's a deliberate marketing decision by them to keep you tied to iTunes.
In the end, the iPods are nice looking, reasonably priced players that integrate very well with iTunes, but the fact remains: their on-board software is limited compared to what's out there.
How else would you like to do it?
I want to just unplug it when I'm done. That works on other platforms, why not on Macintosh? It's the obvious and simple thing to do. And don't tell me that it might "interrupt a write"; other platforms deal with that without problems.
Have you tried to remove a removable device on a PC (say, an USB memory stick)? You need to go to a tiny icon on the task bar, use a contextual menu, and two crazy screens show up with non-intuitive options, etc. Ridiculous.
Bullshit. On Windows, you just unplug the thing when you're done. And on a correctly set up Linux machine, you also just unplug the thing.
Why can't African states bootstrap?
Colonialism ended less than half a century ago; it takes a lot longer to develop even under ideal conditions.
And conditions are hardly ideal: Africa's most natural exports are heavily disadvantaged by Western subsidies, and economic exploitation of Africa and propping up of undemocratic regimes in Africa by powerful nations continues to this day. Even our so-called economic aid is usually tied to specific purchases from the donor nations, so it isn't very effective, and what isn't tied up that way disappears in corrupt regimes, usually with knowledge of the Western donors.
Depends on the real world. To me this seems a slam dunk: For $2,000, I get an app that will run on multiple OSes, considerably broadening my potential customer base. And it'll only add a few bucks to the end product.
Cross-platform support at the toolkit level is uninteresting for most companies; if you are serious about cross-platform, then you have to put in place an entire infrastructure for supporting the new platform (help desk, tech writing, marketing, etc.). At that point, you might as well create a new native UI, which is quick and easy given modern tools, and which will be better quality than anything you do with a cross-platform toolkit.
Only a PHB would deny that kind of request, and they wouldn't even do that if you dropped the words, "Mac OSX".
Non-Macintosh software houses don't generally give a damn about Macintosh (not worth the extra trouble and expense), and Macintosh software houses are disdainful of Windows.
Overall, cross platform support is a small niche market.
To me, video on these devices is a gimmick. A useful application would be photo storage. But they turn out to be rather poor for photo storage, even though Apple offers them for that purpose: the camera/flash-to-iPod transfer is so slow and power hungry (according to reviews) that the battery runs out before transferring a single large card, and Apple doesn't even list any compatible CF readers. So, I ended up with another photo storage device.
The iPods are OK MP3 players, but even as MP3 players, they are mediocre in terms of features and functionality. Other MP3 players fully integrate disk, photo, and music player functions--you copy files to the device and they are accessible. With iPod, you must use a special application. And the fact that after all these years, Apple still hasn't figured out how to let you safely remove removable devices without dragging them into the trash can first is sad.
All of this is unfortunate, because the devices look nice, have nice screens, have a generally good UI, and have good sound quality.
You can make a sales pitch for how the extra $2000 don't matter and how good Qt is until you're blue in the face, the fact is people aren't buying it in large numbers. And if you think that you can present those kinds of considerations to justify $2000 in a real business, you haven't encountered real-world corporate budgeting.
As for Troll Tech's right to make money, nobody is disputing that; if they just offered Qt as a commercial, closed source product, we wouldn't even be discussing them here. Nobody has been complaining about the fact that XVT is closed source, for example. What I and others don't like is their marketing, which relies heavily on using the open source community to generate demand for their proprietary product.
Since most of those on slashdot are advocates of free software of course any company choosing to sell a product, try to be successful and employ people (which drives our economy) is going to be the bad guy.
Well, I can't speak for all of Slashdot. I have no problem with a company selling closed source software. What I fault companies for is if they use open source or free software as a marketing gimmick but are actually pursuing goals and strategies that run counter to the goals of open source and/or free software. I think that's what Troll Tech is doing: they are using KDE and the GPL as a marketing gimmick for a product that would otherwise have disappeared from the market long ago.
What I think Troll Tech ought to do is make all future releases of Qt closed source. KDE could go on with the GPL'ed version of Qt, so nobody would be harmed, and the conflict of interest in the development of Qt that the dual license represents would disappear.
And, whatever I think is the right thing to do, I guarantee you that this problem is not going away, no matter how much the KDE project may want it to. KDE could have taken over the Linux desktop long ago (it has really consistently been ahead of Gnome), but as long as this issue persists, that just is not going to happen, no matter how good KDE may get technically.
What "rights" do they retain, any other Free Software developer gives away?
Troll Tech retains all rights to the software; in particular, they retain the right to sell it commercially under non-free licenses. That is very different from most other GPL'ed projects, where the project accepts contributions under the GPL and nobody has the right to sell a non-free version.
Their copyrights?
That is exactly the problem: all of Qt is under their copyright. When the community contributes value to Qt (and they do), Troll Tech arranges for the resulting enhancements to Qt to fall under their own copyright or are shipped separately. That is very different from most other GPL'ed software projects (e.g., the Linux kernel, Emacs, etc.), where often dozens of different people hold the copyright on various parts of the source.
Multiple ownership of the source code of an open source project is good because it pretty much guarantees that the project will remain open and that the project will respond to the needs of its user community. When the entire codebase is owned by a single company, then the project will be driven by the commercial interests of that company, whether or not they also happen to ship a GPL'ed version.
I have a Zaurus SL-C1000 and 760, and also have had a 5500. I do understand what you're saying about Qt/Embedded (now just called qtopia), but it's horses for courses: an X11 handheld device isn't what everyone wants...
...]
;-)
If Qtopia ran on the Zaurus through X11, it would be indistiguishable to users from the way it is running now. It may be that Troll Tech originally really did think that Qt/Embedded could be more efficient than Qt/X11, but they were wrong, and at this point, the only reason for Qt/Embedded's existence is that it is great for Troll Tech's business.
Qt/Embedded (now just called qtopia)
Qt/Embedded is the toolkit, Qtopia is the embedded toolkit plus a set of standard applications. (At least that was the case last I looked.)
[PdaXRom, X/Qt,
Those are workarounds for nerds, not something you can use for mainstream application delivery. If I want to deliver an open source application on Qtopia to mainstream users, I have to write it in Qt. And if I want to deliver a commercial application on Qtopia to mainstream users, I have to license Qt, overpriced and cumbersome that it is. And Troll Tech knows this; it's great for their business, it just happens to be lousy for the adoption of embedded Linux.
I disagree that the apps aren't optimised for small screens: to me, it's an excellent UI. Opie seems to look exactly the same, surely imitation is the sincerest form of flattery
Well, I can't comment on Opie, and Opie isn't exactly mainstream anyway. The point is, we don't have a choice at this point: almost all embedded Linux devices run Qt, and you simply can't realistically develop and deploy other toolkits on it.
Yeah, releasing kick-ass software under the GPL is a clear threat to Free Software. Seriously: are you a retard?
So, your claim rests on your assumption that it's self-evident that dual licensing is good for free software. But it isn't self-evident to me, nor is it to lots of other developers. In fact, I think Troll Tech's licensing scheme is harmful for free software, and I have given you some reasons for that.
Your insults and flames don't substantiate your arguments. And given that the KDE project was running for a couple of years without even noticing that the GPL and QPL were incompatible and only avoided shut-down because Troll Tech finally gave in and changed their license, KDE proponents like you have no credibility when it comes to judging licensing issues--you got it badly wrong before, and it's pretty clear you are still getting it wrong this time around.
And if you hope that this issue will just go away, that people like me will just shut up and develop for Qt if you only insult us enough, and that and KDE will take over the world, well, just keep on hoping.
It makes no difference to me as a customer how much it costs to develop the library. There are plenty of good alternatives that I don't have to pay for (beyond what I already pay for the OS in any case): Swing, wxWidgets, Gtk+, Cocoa, MFC, .NET, etc. It's unfortunate for Troll Tech that they can't subsidize their development with other revenue streams, but that doesn't make me any more inclined to pay lots of money for their stuff.
Furthermore, if I'm going to pay $1800/developer, then I'm going to pay that money for improving an open source toolkit like wxWidgets or Gtk+, where I get specific improvements for my money and where I don't have to pay over and over again.
So wait, your biggest problem with trolltech is they don't have a price point for you?
No, the biggest problem is that Troll Tech is using open source licensing as a gimmick to push a commercial piece of software at an inflated price point. Without the open source marketing, Qt wouldn't have a prayer in the free market because it's overpriced for what it does.
And it is a commercial software project: it is run like one, it is sold under a commercial license, and Troll Tech retains all the rights.
It's their software, right? They have the right to make money off of their own work, right? And license it however they choose?
Yes, they have the right to offer it under whatever license they like. And we have the right to warn people against using it, and to expose what their approach to doing business means to FOSS in the long run.
In short: what is the "issue" with Qt? the fact that it licensed under the GPL. And why is that a problem? Because it makes it difficult to write proprietary software using Qt (you have to pay TT in order to do so).
The issue with Qt is not that it is covered by the GPL, the issue with Qt is that it is covered also by a commercial license. And the problem with that is that it makes it easy to write proprietary software using Qt--all you have to do is fork over some money to Troll Tech. Troll Tech's dual licensing scheme inhibits most of the things that make open source good: free sharing of ideas and improvements, communal development, and the ability of different forks to compete on merit alone.
Troll Tech is using open source as a marketing gimmick to sell proprietary software. They don't want open source to succeed because it would kill them as a commercial company; they just want open source to succeed enough to make their business model work.
What is the problem here? TrollTech offers their product under the GPL. They also offer it under a proprietary license. They don't force anyone to use their toolkit,
It's good that you recognize that you aren't forced to use it. So, I'm suggesting to you: don't use it.
Qt is licensed under the GPL. I really fail to see how they could "monopolize" anything. or are you worried what would happen if Linux "monopolized" the OS-market?
Software built on Linux doesn't need to fall under the GPL, but software built on Qt does.
And if Linux were dual licensed, I would be worried about it monopolizing the market. But Linux is covered only by the GPL, and that makes it OK.
So I shouldn't offer any bug-reports to the kernel-folks, because that might make the product a bit better, and some company might earn some money through it?
You should offer bug reports to the Linux kernel folks; the Linux kernel project is not dual-licensed. All contributors and all users of the Linux kernel use it under the same conditions.
People are complaining when some company offerws kick-ass software under the GPL?
Yes, people are complaining about that, and rightfully so. Not everything that is offered "free" (even "free-as-in-speech") is a good thing.
For example, Joe might give you free cocaine, and you might enjoy it, but that wouldn't make it a good thing. And Joe would give you free cocaine because he knows that there is a good chance that you get addicted to it and have to come back to him later and pay lots of money for it. That's roughly the kind of strategy Troll Tech uses with Qt.
GTK+ is controlled by the GTK+ team. GNOME is controlled by the GNOME-team. Mozilla and Firefox are controlled by the Mozilla Foundation. So why exactly are you singling out TrollTech and Qt here?
The Gtk+ team or Gnome foundation doesn't have any special rights over the Gtk+ code base--they can't make a commercial version. Troll Tech, however, has special rights over the Qt code base, they can make a commercial version, and they have. Troll Tech is using an open source license to promote their proprietary software.
Oh I see. the whole point of creating free software is to suck up to creators of proprietary software!
Apparently, that is what you believe, because that is exactly what proponents of Qt are doing: they are using an open source licensing gimmick to promote Troll Tech's business and help them make and sell proprietary software.
I heard that Microsoft is hiring. People like you would fit right in there!
Actually, it is your and Troll Tech's mindset that is disturbingly close to Microsoft's; you are promoting proprietary software, you are simply using open source to do it. I heard that Microsoft is hiring--why don't you apply yourself? You apparently put expediency and the appearance of freedom ahead of actual freedom and open source principles, so you should fit right in at Microsoft.
TT wrote the code, therefore they get to choose the license. If you find their license unacceptable [...] use something else, and quit your whining!
The fact that Troll Tech may choose whatever license they like doesn't place them above criticism: their policies are a threat to free software and open source software, and, of course, I am going to continue to criticise them for them.
And, of course, I'm using something else. In fact, I'm contributing to open source desktop and toolkit projects. Troll Tech must not be allowed to succeed.
I haven't seen the patent first hand, but often something obvious now, wasn't at that time. If it were then why wasn't someone else doing it already.
Simple: because the market demand wasn't there yet; pay-as-you-go only became interesting for wireless once companies had actually taken care of all the customers willing to go with what amounts to higher priced plans.
You should not be able to patent something that is merely an application of standard principles by someone of average skill in the profession. Pay-as-you-go billing clearly fits into that category, since it had been used with many other similar services (e.g., prepaid long-distance).
I can list many examples of this. The mouse, keyboard, screens, printers, windowing environment, The Internet, an Operating system and even a CPU and the IC chips, were at the time major conceptual steps forward. I can't tell you how hard it was to explain what the Internet or even a Network was to people in 1983, they just couldn't grasp it.
Well, which only goes to show that your judgement in these matters is impaired. The Internet was developed in the 1960's; by the 1980's, any patents covering the original ideas were reaching the end of their life, and the concepts and technologies were widely understood by people in the profession. In fact, the Internet and what people do with it really hasn't changed all that much in the last 20 years, except that a whole lot more people/companies are using it and that it has gotten a lot faster.
And I suspect that many of the other items you list probably weren't invented in the form and by the people you think invented them either.