That isn't it. That's a tool that does something similar to but not quite the same as what I'm asking for. And, regardless, lsof isn't used when building an error message to tell people on OS X why they can't dismount the disk.
What I want is for the dialog to say something like "You cannot dismount the volume "Workbench" because "Wordworth.app" has the file "Workbench:Docs/Thesis.ww" open." rather than "There was a problem dismounting Workbench: the volume is in use."
1. Property that is or can be inherited; an inheritance.
2. Something that is passed down from preceding generations; a tradition.
3. The status acquired by a person through birth; a birthright: a heritage of affluence and social position.
Which definition applies to "shared the same developer" and "some increadibly low level stuff is similar"? Which definition is even vaguely relevent to that single connection VMS and NT have?
I think that's downplaying things a little and upplaying others.
Microsoft had several attempts at revolutionary change. DOS to Xenix (stillborn.) DOS to DOS/Windows (third attempt worked.) DOS to OS/2 (failed.) 16 bit DOS/Windows to 32 bit DOS/Windows (success), DOS/Windows to NT (fifth attempt, after 3.1, 3.5x, 4, and 2000, successful.) There are probably others, but those five are particularly prominent and were, at some stage, active Microsoft policies. All of them pretty much changed PC based computing and PC based software development, to the extent that in each case, either Microsoft, the user, or the developer, threw out virtually everything they had (user skills, developer code (at least, those who didn't want their apps to look stupid and "legacy"), Microsoft core code) and had to start afresh.
Despite DOS being underneath Windows 3.x, do not write off the revolutionary nature of the change. DOS itself was barely used once Windows was running and developers didn't access it directly at all..
Apple has had a number of similar attempts which included some major disasters. For example, Apple II to Mac wasn't supposed to happen. Apple nearly bankrupted itself in the early eighties trying to get everyone moved to its next generation, all singing and dancing, Apple III platform. It spent a huge amount on marketing, dropped marketing of everything else, and in the end only survived because the II series continued to sell despite everything.
One could probably say that the Lisa was another attempt, but I'd put it in the stillborn class like Microsoft's Xenix. A lot of development work went into it, a lot of lessons were learnt, but Apple dropped the system before they really would otherwise have started to push it. Nonetheless, we could call it another genuine attempt at revolutionary change.
In some ways, Apple has benefitted from not even playing the game for the most part. After the tubulant eighties, Apple started a number of projects during the nineties that it subsequently dropped or pulled out of, but didn't actually bother to get to the final stages and start releasing real products that could fail. There were a few improvements on the hardware side (just as in the PC world, where the PC industry - in a remarkable show of communally doing the right thing, went 32 bit, switched to PCI, etc), but we saw nothing major in system software for around a decade from Apple.
I'm glad Apple are now changing that. They've released a lot of new stuff in the last four years. It's starting, gradually, to feel a little more exciting to be in this industry again.
The same, incidentally, is true of OS X. I've lost count of the number of times I've hit the little "Eject" sub-icon and nothing's happened, or a dialog reporting that an application is using the mounted disk image/Firewire drive/CD/etc has come up, and not even been able to find the offending app.
That, incidentally, is something that needs to be fixed. Locks need to have a clearly idenfiable chain of ownership so low level calls report what processes are locking the drive, and higher up apps can follow these processes to find, eventually, what application is actually doing this and report it to the user. That goes for GNU based operating systems, other Unixes, Mac OS X, Windows NT based systems, etc. It's 2005, we've been locking stuff for how long?
On a seperate note, I think the GGP is a troll. Various resources being locked means you don't have true pre-emptive multitasking? Riiiiiiiiight.
NT's "VMS" heritage is that it shared the same developer.
Please, let this meme go. While some increadibly low level stuff may be similar in both operating systems, from a user and programmer's point of view they resemble each other as much as a bicycle resembles a slice of cheese.
I've used and programmed for both. Those who claim they're the same seriously have no idea what they're talking about.
Apple is doing a great deal at the moment to keep a foot in every market, which has included expanding its software side and producing machines that have relatively low margins. It also has the iPod business, of course, which is "hardware" but not in the "competing with Dell" category. Call it peripherals, 'cos that's what the iPod is.
So, personally, I wouldn't rule out Intel OS X forever. It's not likely at the moment, certainly it's more than two years away. But if, for example, Apple's sales of PowerMacs, PowerBooks, and iMacs, were to tank, with only Mac minis, eMacs, and iBooks selling in any quantity, and if it can hold off Microsoft, Apple might find itself wondering if selling an operating system to OEMs is a practical way to make money again. The margins on its low-end machines would probably not make up for potentially selling many, many, times as many $50 boxes to Dell, HP, et al, and $200 boxes to consumers.
Apple may produce an Intel version, but the time has to be right. Right now, it isn't. It was in the late eighties and early nineties, and may have been in 2000-2002, but it isn't now.
Socialism is nothing of the sort. The co-operative movement predates communism and has always been considered socialist. Trade Unions are also socialist. The term socialist predates any governments calling themselves it, and the movement that blossomed this way has its origins in Robert Owen's paternalist attempts to reform mills and the working conditions of those who worked for them. The mills he reformed were those he owned.
The common thread in all of these cases is people working together, cooperating rather than competing, to improve living conditions, either on a local or global scale. In some cases, governments have been elected with "socialist" policies where the government takes the view that it is the will of the people and has a moral right, under socialist morality, to maintain certain services. But to extend, as you have, the assumption that this means that socialism "means" government monopolies is like arguing that flying by plane "means" crashing into the World Trade Center.
I've commented before I find it amusing that many self-styled Libertarians who use the "S" word as an insult are high up in movements like the Open Source movement. It's a shame people are so hung up on certain words, they'll jump through hoops to describe things they agree with in any other way possible.
Well, in fairness, he makes a bunch of value judgements that seem a little out-of-whack in a only-a-Mac-basher-would-put-it-that-way way. Like the suggestion that XP SP2 represented a greater upgrade than Panther->Tiger, which he repeats several times.
Now, I'm of the opinion that a fundamental change in the filesystem (the vastly upgraded metadata system that allows the kind of dynamic searching et al described, coupled with yet another GUI look, in addition to upgrades of less prominent functionality elsewhere, represents a bigger upgrade than, say, Windows 2000 to Windows XP, which in many senses was 2000 with yet another GUI look and lots of minor improvements.
Indeed, it seems a festival of grudges, from the discredited claim that Dashboard is a rip-off of Konfabulator (Thurrot even mentions the counter evidence himself, but not in the context of discrediting the claim, instead in terms of discrediting the need for any special status for these widgets), to an attack on Apple's over-the-top marketing that manages to be just as over-the-top as Steve Jobs on Jolt:
It will not change the way you use your computer at all, and instead uses the exact same mouse and windows interface we've had since the first Mac debuted in 1984.
Er, what?
From my point of view, it's a wierd review. It is, of course, aimed at an audience that is rabidly pro-Microsoft just as much as a review of some Longhorn-type thing on MacRumors.com would be as grudge filled.
It's certainly interesting, it's a good demonstration that platform fanaticism is still very much with us, and very much full-duplex. It's interesting less for what it says about Tiger than what it says about Thurrot and his audience.
And they'd be able to interoperate with Linus Torvalds how exactly if they wrote an incompatable SCM system?
The bottleneck here is that Linux is currently maintained by Linus Torvalds. Linux is currently the kernel for the GNU operating system. It has mindshare and it's going to be hard to replace. If Linus continues to use proprietary products for maintaining Linux as managements and submissions become more and more complex, it is going to become increasingly difficult for the Free Software community to contribute.
Right now, the options aren't a ground up new SCM system (of which many already exist). The options are:
Persuade Torvalds to move away from the dark side. Ask him to stop using Bitkeeper. Point out the many ethical problems other contributors have been having with him doing this. This has been pretty much the strategy of the last three years, and far from actually making a difference, has actually made none whatsoever. Linus prefers, as he's demonstrated here, to attack his contributors - attack the very people who have made Linux usable, made it interoperate with Windows networks for example - rather than address their concerns.
Reverse engineer Bitkeeper, using the information to write compatable clients.
Dump Torvalds. Fork the kernel. That's radical, and it's going to be hard to get people on board with a single fork.
I'm not going to rule out (3) from happening eventually, but having failed with the first, we're now seeing people resort to the second, which despite the bloody-minded behaviour of the BK people, is probably going to be successful.
I can't blame them, and I can't see anything remotely unethical about what they're doing. I am staggered that anyone sees reverse engineering in the hysterical terms you do. Far from it being unethical, the question should be why these protocols need to be reverse engineered in the first place. Why aren't they documented?
In any case, I think this is nonsense. If the problem had been the lack of a gratis client, and the wish to make one where none existed, then your allegation might make sense. But there was a gratis client, indeed it was the gratis client that made the entire issue come to a head.
All that happened was that Torvalds deliberately picked a proprietary SCM system, and other developers - who happen to have strong opinions on the ethics of using Free (as in liberty) vs Proprietary software - wanted to interoperate with Linus.
Nobody was trying to undermine BK's business model. They were trying to interoperate with a virtual collegue (Linus) without using proprietary software. If they wanted to undermine BK's business model, they could have continued to use the gratis client provided by the Bitkeeper authors.
Driver support, incidentally, existed for the vast majority of hardware that uses reverse engineered drivers in Linux. The drivers were for Windows. Why is not using Windows more ethical than not using Bitkeeper?
I realise it's (obviously!) not an ethical issue for you to use proprietary software. People are entitled to have different views on the subject. But it takes willful ignorance, contempt for those who think differently, not to mention a complete loss of logic, to ignore the fact that this was a Free (as in liberty) software issue, not about trying to get software without paying for it.
Take a look at the filesystems supported by Linux, together with device drivers with little apparent official documentation, and see how much of the kernel is actually dependent on reverse engineering.
And Tridgell didn't reverse engineer something that already had a capable Free Software client. You know, the PC actually has a "capable and popular" operating system. So, presumably, it's absolutely wrong for the Linux developers to continue to use reverse engineering to develop a Free Software alternative. Right?
Wrong. This is about freedom. The Bitkeeper people had no business discouraging those who want to use free tools yet who want to interoperate with those who lack the same standards from writing free tools to spec. It was bad enough that the protocols were undocumented and proprietary to begin with. It's worse that this kind of vengeful stance was taken against third parties for daring to have an association with someone trying to create those free tools.
Im hardly shocked that Linus came out with a stance that pretty much noone expected him to take, and I have great respect for him doing it. He doesnt really care much for the FOSS philosophy, and that is entirely his right to do so, although I am shocked by the number of people who expected Linus to have a similiar outlook as RMS or Alan Cox.
I'm absolutely amazed. He may not be the loudest proponent of FOSS in the world, but he's at least made himself look like such a proponent, and he has relied upon the very people he attacks doing exactly what he's attacking them for doing to make his kernel usable and what it is today.
Your respect for him may have risen, mine has dropped. I was prepared to handle the fact he adopted BK in the first place because, well, people do often see themselves as pragmatic when making decisions that essentially defy good practices. He should have learnt something from this lesson, but essentially it looks like he's merely digging himself into a deeper hole while yelling "You all suck!" at those trying to get him out.
But.NET has certain limitations that, open square bracket two tees close square bracket, ultimately undermine the entire system.
Look, what.NET does is hardly new. It's been done before countless times. Java is one example, smallTalk and FORTH are others. I've used FORTH. I've programmed in it. It's a good language. But like all of these systems, the IDE is constrained, ultimately, by the technology in use.
.NET works by abstracting essentially the Windows paradigms. That, however, is its cross as much as its camel. What we're looking at here is the 2005 equivalent of the UCSD p-System. And, well, the problem is this:
When I first found out about the UCSD p-System I thought "Great! Portability, a decent solid language (and Pascal was, for the time), a standardized front-end, etc." You could run the damned thing on anything that had a disk drive and a few kilobytes of RAM. But that's where things went horribly wrong.
At the time this came out, both were essentially rare. Your C64 was as likely hooked up to a tape deck as a floppy drive. Worse, the C64 had a whole bunch of crap that the p-System couldn't use. It had colour. It had graphics. It had sprites. It had sound. And to add that kind of functionality to p-System would have simply prevented it from running on many other platforms. Commodore, as you know, ended up going bankrupt.
This wasn't altogether surprising. I recall hearing about a project called "360" in the mid-sixties that attempted to create a single platform for mainframes. The idea was that the mainframes were all compatable, so you could port your software from one to the other pretty much by putting the cards in the other machine's bin. I bet you're not reading this. This was radical at the time, but it came at a price - the top end system was limited in functionality to that of the bottom end. Within fifteen years, the idea was dead, the systems lived on but co-existing with a range of incompatable systems from the same supplier. Within 25 years, the company concerned was teetering on the verge of bankrupsy.
I think you can see where I'm going with this, so let me get to the point..NET is limited by its need to run on all current hardware and support virtually nothing original or innovative. And what goes right to the heard of the IDE. Without tools that are synergistic with the underlying technologies, programmers cannot develop great software..NET hides the underlying technologies, limiting the scope of software development and denying programmers exposure to the very things that could inspire them.
This is ultimately why.NET will probably win out. With superior IDEs including the excellent Visual Studio.NET environment for software development, a consistant, well designed, underlying architecture, and fantastic expansion potential, I think we'll see.NET go from strength to strength. The work by people like Miguel de Icaza will help too, as programmers in other environments will be exposed to the latest Microsoft technologies, and will almost certainly be chomping at the bit to get on board.
But that can only happen with a well designed IDE. Unless software developers are comfortable when developing for.NET, they'll avoid it. But, looking at what they've done so far, and looking at their experience in this area, I don't think Microsoft has a thing to worry about.
I[tt] surprises me it's out this early to be honest, what I heard is that Tiger's been delayed indefinitely, due to a major bug and a political issue. Apparently in tests, it took over 20 minutes to copy a 17Mb file from one folder to another. The copy, apparently, causes Netscape to hang, and puts a lot of strain on editors like BBEdit lite.
As if to add insult to injury, Steve Jobs has apparently dealt with this inevitable delay by ordering that the BSD-underpinnings of the operating system be stripped out and replaced with GNU or Unix code, citing an internal study that shows BSD developers in complete disarray, with usage - based upon Usenet postings - showing a definite decline in BSD support. Perhaps if Apple ported Mac OS X to Intel, they wouldn't be suffering these problems.
On an unrelated note, I've just heard on Talk Radio that the famed computer hacker Stephen Wozniak has died, aged 54. Whether you enjoyed his Integer BASIC interpreter, you can't deny his contribution to modern computing. Truly an American icon.
In fact it would appear that your definition of neutral is almost synonymous with "GPL-compatible".
No, my definition of neutral is neutral. The GPL doesn't favour any project or any developer. Your suggestion incidentally suggests you didn't bother to read anything I wrote, as the discussion is about finding a license to be compatable with. As I said, the issue right now is that there's no other candidate on the table. The GPL is the only neutral license that's widely used and provides for no further restrictions being added. You're welcome to write an alternative.
A better choice for a truly neutral default license, that imposes as few restrictions as possible on the end user, would be the BSD license.
No, it wouldn't. As I said above, the BSD license doesn't allow for reciprocal licenses, it allows a licensee to impose more restrictions on the use of their code. It is therefore unacceptable as a license to be compatable with. Otherwise all code goes one way - you can develop new code under the BSD license and someone can MPL it, but not vice-versa. You can't take code from an MPL project and make it BSD.
Right now, the discussion is about legal practicality. We need a way to be able to combine code from different projects. If those projects are a mix of CDDL, MPL, et at, we can't do that. The only way to get this to work is if the projects are compatable in that their licenses are compatable with a common, neutral, license, that is the code licensed under the project-specific license can be relicensed under some other license. This is not possible, and will never be possible, with an overly liberal license like BSD or X11. Think practicality.
I appreciate this is a difficult subject to understand, especially on Slashdot where "The GPL is 3v1l! BSD is free!!!11!!" is taken as a non-ideological statement, but I want to give you a specific example that might make it easier to see what the issue is:
Suppose I want to take some Solaris code and put it in Darwin. The former is licensed under Sun's CDDL. The latter under Apple's APSL.
If we were to go to Apple and Sun, and asked them to modify their licenses so I could do this, legally, what do you think they'd want?
To modify their licenses so it's possible to take their code and relicense it under a neutral copyleft, such as the GPL, preventing their code from being used in proprietary projects made by their competitors?
To modify their licenses so it's possible to take their code and relicense it under a neutral non-copyleft, such as the X11 license, allowing their code from being used in proprietary projects made by their competitors?
If you said (2), you might want to look at both the CDDL and APSL. Both are essentially licenses that prevent anyone other than the "Initial developer" (ie Apple and Sun in the case of Darwin and Solaris) from putting code in proprietary projects. Why would they suddenly be happy with a loophole you could drive a train through being introduced into their licenses?
Have you stopped to consider that it is the GPL that is incompatible with the other licenses, not the other way around?
Apologies for sounding pedantic here, but this is incorrect. The GPL cannot be incompatable with the CDDL and MPL without the MPL and CDDL being incompatable with it. It's obviously silly to describe one as incompatable with the other but not the other way around. The MPL and CDDL most certainly are incompatable with the GPL.
If you badly worded the above, and meant the emphasis should be on the GPL in terms of where the problems are coming, my entire posting addresses that: The GPL is neutral, it requires the imposition of no further restrictions, and therefore is one of the few players in town that qualifies as a legitimate baseline FOSS license for licenses to aspire compability with. The same is not true of the CDDL, MPL, or non-copyleft licenses such as the BSD and X11 licenses. The former are not project neutral. The latter do not prevent more conditions being attached, which essentially makes them useless from an interoperability point of view.
For example, you can mix and match code from CDDL, BSD and MPL all you want, there are no restrictions in that regard.
No, you can't.
The MPL is not compatable with the CDDL and vice versa. You cannot use code from an MPL project in a CDDL project. In fact, the licenses get legally problematic even when you try to combine code licensed under the same license from two different projects sourced by two different "Initial Developers". So not only is the MPL and CDDL incompatable with the GPL and each other, but frequently not even themselves!
When you throw GPL in, you have to license the entire kit and kaboodle under GPL.
There really isn't a lot of difference in the case of the MPL and CDDL. Both require that any relicensing be done under a license that imposes the same restrictions and grants. Essentially they're saying "You can relicense this under any license you want, as long as it's this one, or a rewrite." Most additional rights granted are, in any case, a re-statement of copyright law.
The real problem with both is that both try to elevate the initial developer to a special status. While it might be understandable, it's by design going to create problems getting multiple projects to cooperate. As soon as you say "This license does something special if {something that's project specific}", then you're into incompatable licenses.
Example - a company works with an OSS project and contributes code back on a regular basis. They have an opportunity to provide the project with support for a new deice, but the information they can do this with is covered under an agreement that prevents them releasing the code for this device. Under BSD or CDDL, no problem, they are not required to provide the code. Under the GPL, they cannot do this - development is stifled. Of course, the root of the problem is the locking down of the required information - but often we are talking buisness and competition issues i.e the real world.
In the situation you describe, the "fault" doesn't lie with the GPL. There's nothing stopping the developers - as a group, all of them who have contributed code - from saying "Yes, we'll do this closed source thing for this particular module", because the developers own the copyrights, and can agree, as a group, to dual license. This, of course, doesn't mean the developers will, it may be - shock horror - that the developers are opposed to their work being used in proprietary code, which leaves the company that wants to restrict the information used for the device driver the major stumbling block.
The CDDL, incidentally, doesn't allow this any more than the GPL does with the possible exception (and the CDDL is actually vague here, which makes matters worse) of the "initial developer" who gets some special rights. Both the MPL and CDDL are so-called "weak
The GPL is a neutral copy-left. It doesn't give special rights to the "initial developer". It's well understood, there's a large body of code licensed under it, and whether you agree with the politics behind it or not, it's a good baseline for those who want a copy-left license.
Why is this important? Because most developers want compatability with some common license. If the MPL and CDDL were compatable with the GPL, then there would be ways of mixing code licensed under one with the other.
Not having compatability harms Freedom. It means that you cannot "use" code under circumstances that most Free Software advocates, and probably most Open Source advocates, would consider reasonable and a requirement for software to be considered Free or Open Source.
Yes, it doesn't have to the GPL. But there are no other real candidates. The BSD license doesn't do it because it has no requirement that further conditions be eliminated, so while BSD code can go in an MPL project, the fact the CDDL is compatable with the BSD license doesn't mean CDDL code can go into an MPL project.
Other licenses are, for the most part, more obscure and are no better than the GPL, or are not neutral. Some are even incompatable with themselves as they grant special privileges to an "Initial Developer", which means two projects under such licenses with different initial developers cannot share code.
So for the time being, GPL compatability needs, as far as many people are concerned, to be the baseline test of whether a license is a good one, regardless of whether - on the surface - it is "too political".
If you hate the fact that software can be called Free or Open Source and yet you can't use code from one project in another, you should be opposed to incompatable license proliferation. Incompatable licenses are killing Free Software.
FWIW, anyone who needs a fairly reasonable list of licenses will find a reasonable collection, with related discussion, here.
The OSI has a list of "approved" licenses here, but it's literally just a list (no explanation of the differences or analysis of their implications), and I get the impression the OSI is regretting approving so many of them, given most of them are incompatable with one another (some of them can be incompatable with themselves, depending on the precise circumstances!)
While this doesn't contradict what I said, I should probably point out it's incorrect. It was designed by IBM as part of an effort to give the IBM PC some form of networking, and Microsoft worked with them on this.
I was using Microsoft networking back in the late eighties, on a network of something called a Research Machines Nimbus, at school. Oddly enough, we ran Windows 1.0 over it, with hilarious results (ie it was slow and crashed a lot.)
Anyway, the point is Microsoft's SMB protocols pre-date Windows. Windows interoperated with them without any problems, they were just DOS drives, after all.
You young'uns! You don't know how hard it was then! We used to have to wire coax to the back of PCs to get out Ethernet networks, kid!
Because building a model may be prohibitively expensive. What if you had a great idea on how to improve a satellite? Are you really going to be able to afford a satellite to show a basic working implementation? Or how about the subsequent launch into orbit?
If you're not going to spend any money, what costs do you have to recover from creating the invention?
I don't see this as a major problem. The point about patents is that they're there to encourage people to invest the time and money into making something new and real. What you're proposing - giving monopolies to people who will not spend the money - actually discourages innovation as it prevents or makes more expensive the doing of the real work to make something theoretical a reality.
...and those bastards who write the dictionaries are in on it too. It's a conspiracy I tell you, a conspiracy! Why I bet all the documents that were published before the RIAA was even created that used these terms are plants!
I don't want to suggest that the publishing of private medical records is justified, but I think your contract argument is incorrect here: if she wasn't paid and wasn't going to be paid (it was a threat, after all), then her employer was already in breach of contract. Presumably she has a right to consider the contract null and void at that point.
If we agree that you'll not kick me in the balls, and I'll not either, are you going to feel restrained from doing just that if I kick you in the balls anyway? Would you? Huh? Would you? Come here and prove it.
I'm not sure what brought that on. I think I'll go and have a little nap somewhere. Maybe some chamomile tea.
Took the train from Hartford CT to New York (Penn Station) in 1998. Took a commuter train from the outskirts of Boston into Boston (I forget the name of both locations, the station in Boston wasn't the major one with the bus station next to it, it was on the other side of town or something.) in 2000.
Neither were recent, so things may have deteriorated rapidly since then, but both were extremely comfortable then, if a little slow. Comparable to a prison bus? Well, I'll take you at your word that you're the one with the experience there (I'm fortunate and honest enough to never have been convicted of anything worse than a minor speeding offense), but compared to any Continental, Northwest, Delta, British Airways, or Virgin Atlantic flights I've taken, it was the height of luxury.
As for driving being only stressful and unpleasant because I "allow" it to be, I wonder if you have a clue what you're talking about? I can't exactly read a book while I drive. Or get up and go to the restroom. Or get some food and drinks. Or go for a stroll. I have to concentrate 100% on the road ahead, and I have to do so in a relatively cramped little box.
And since surveys of potential Amtrak customers support my point, I'll say I was speaking for them, not everybody as you assert.
You made a sweeping claim about Amtrak which doesn't hold up in my experience. Given you didn't limit your claim to "potential (but not actual) Amtrak customers" then, I think it was reasonable for me to suggest your comment was supposed to be universal.
I see you don't like Amtrak. Not everyone agrees with you.
Riding the train is a wonderfully relaxing way to travel. Just the ability to get up and walk around makes it an order of magnitude better than the alternatives. The decent sized seats, flat ride, lack of restrictions on when and where you can eat and drink, go for pee, etc, puts it into a class of its own. It's a massive shame that it's so restricted in the US.
Slashdot Mirror
What I want is for the dialog to say something like "You cannot dismount the volume "Workbench" because "Wordworth.app" has the file "Workbench:Docs/Thesis.ww" open." rather than "There was a problem dismounting Workbench: the volume is in use."
It's dumb. Stop claiming it.
Microsoft had several attempts at revolutionary change. DOS to Xenix (stillborn.) DOS to DOS/Windows (third attempt worked.) DOS to OS/2 (failed.) 16 bit DOS/Windows to 32 bit DOS/Windows (success), DOS/Windows to NT (fifth attempt, after 3.1, 3.5x, 4, and 2000, successful.) There are probably others, but those five are particularly prominent and were, at some stage, active Microsoft policies. All of them pretty much changed PC based computing and PC based software development, to the extent that in each case, either Microsoft, the user, or the developer, threw out virtually everything they had (user skills, developer code (at least, those who didn't want their apps to look stupid and "legacy"), Microsoft core code) and had to start afresh.
Despite DOS being underneath Windows 3.x, do not write off the revolutionary nature of the change. DOS itself was barely used once Windows was running and developers didn't access it directly at all..
Apple has had a number of similar attempts which included some major disasters. For example, Apple II to Mac wasn't supposed to happen. Apple nearly bankrupted itself in the early eighties trying to get everyone moved to its next generation, all singing and dancing, Apple III platform. It spent a huge amount on marketing, dropped marketing of everything else, and in the end only survived because the II series continued to sell despite everything.
One could probably say that the Lisa was another attempt, but I'd put it in the stillborn class like Microsoft's Xenix. A lot of development work went into it, a lot of lessons were learnt, but Apple dropped the system before they really would otherwise have started to push it. Nonetheless, we could call it another genuine attempt at revolutionary change.
In some ways, Apple has benefitted from not even playing the game for the most part. After the tubulant eighties, Apple started a number of projects during the nineties that it subsequently dropped or pulled out of, but didn't actually bother to get to the final stages and start releasing real products that could fail. There were a few improvements on the hardware side (just as in the PC world, where the PC industry - in a remarkable show of communally doing the right thing, went 32 bit, switched to PCI, etc), but we saw nothing major in system software for around a decade from Apple.
I'm glad Apple are now changing that. They've released a lot of new stuff in the last four years. It's starting, gradually, to feel a little more exciting to be in this industry again.
That, incidentally, is something that needs to be fixed. Locks need to have a clearly idenfiable chain of ownership so low level calls report what processes are locking the drive, and higher up apps can follow these processes to find, eventually, what application is actually doing this and report it to the user. That goes for GNU based operating systems, other Unixes, Mac OS X, Windows NT based systems, etc. It's 2005, we've been locking stuff for how long?
On a seperate note, I think the GGP is a troll. Various resources being locked means you don't have true pre-emptive multitasking? Riiiiiiiiight.
Please, let this meme go. While some increadibly low level stuff may be similar in both operating systems, from a user and programmer's point of view they resemble each other as much as a bicycle resembles a slice of cheese.
I've used and programmed for both. Those who claim they're the same seriously have no idea what they're talking about.
Apple is doing a great deal at the moment to keep a foot in every market, which has included expanding its software side and producing machines that have relatively low margins. It also has the iPod business, of course, which is "hardware" but not in the "competing with Dell" category. Call it peripherals, 'cos that's what the iPod is.
So, personally, I wouldn't rule out Intel OS X forever. It's not likely at the moment, certainly it's more than two years away. But if, for example, Apple's sales of PowerMacs, PowerBooks, and iMacs, were to tank, with only Mac minis, eMacs, and iBooks selling in any quantity, and if it can hold off Microsoft, Apple might find itself wondering if selling an operating system to OEMs is a practical way to make money again. The margins on its low-end machines would probably not make up for potentially selling many, many, times as many $50 boxes to Dell, HP, et al, and $200 boxes to consumers.
Apple may produce an Intel version, but the time has to be right. Right now, it isn't. It was in the late eighties and early nineties, and may have been in 2000-2002, but it isn't now.
The common thread in all of these cases is people working together, cooperating rather than competing, to improve living conditions, either on a local or global scale. In some cases, governments have been elected with "socialist" policies where the government takes the view that it is the will of the people and has a moral right, under socialist morality, to maintain certain services. But to extend, as you have, the assumption that this means that socialism "means" government monopolies is like arguing that flying by plane "means" crashing into the World Trade Center.
I've commented before I find it amusing that many self-styled Libertarians who use the "S" word as an insult are high up in movements like the Open Source movement. It's a shame people are so hung up on certain words, they'll jump through hoops to describe things they agree with in any other way possible.
Now, I'm of the opinion that a fundamental change in the filesystem (the vastly upgraded metadata system that allows the kind of dynamic searching et al described, coupled with yet another GUI look, in addition to upgrades of less prominent functionality elsewhere, represents a bigger upgrade than, say, Windows 2000 to Windows XP, which in many senses was 2000 with yet another GUI look and lots of minor improvements.
Indeed, it seems a festival of grudges, from the discredited claim that Dashboard is a rip-off of Konfabulator (Thurrot even mentions the counter evidence himself, but not in the context of discrediting the claim, instead in terms of discrediting the need for any special status for these widgets), to an attack on Apple's over-the-top marketing that manages to be just as over-the-top as Steve Jobs on Jolt:
Er, what?From my point of view, it's a wierd review. It is, of course, aimed at an audience that is rabidly pro-Microsoft just as much as a review of some Longhorn-type thing on MacRumors.com would be as grudge filled. It's certainly interesting, it's a good demonstration that platform fanaticism is still very much with us, and very much full-duplex. It's interesting less for what it says about Tiger than what it says about Thurrot and his audience.
I once applied for that job, but my application was rejected because I forgot to include the barcode from the newspaper containing the listing.
The bottleneck here is that Linux is currently maintained by Linus Torvalds. Linux is currently the kernel for the GNU operating system. It has mindshare and it's going to be hard to replace. If Linus continues to use proprietary products for maintaining Linux as managements and submissions become more and more complex, it is going to become increasingly difficult for the Free Software community to contribute.
Right now, the options aren't a ground up new SCM system (of which many already exist). The options are:
- Persuade Torvalds to move away from the dark side. Ask him to stop using Bitkeeper. Point out the many ethical problems other contributors have been having with him doing this. This has been pretty much the strategy of the last three years, and far from actually making a difference, has actually made none whatsoever. Linus prefers, as he's demonstrated here, to attack his contributors - attack the very people who have made Linux usable, made it interoperate with Windows networks for example - rather than address their concerns.
- Reverse engineer Bitkeeper, using the information to write compatable clients.
- Dump Torvalds. Fork the kernel. That's radical, and it's going to be hard to get people on board with a single fork.
I'm not going to rule out (3) from happening eventually, but having failed with the first, we're now seeing people resort to the second, which despite the bloody-minded behaviour of the BK people, is probably going to be successful.I can't blame them, and I can't see anything remotely unethical about what they're doing. I am staggered that anyone sees reverse engineering in the hysterical terms you do. Far from it being unethical, the question should be why these protocols need to be reverse engineered in the first place. Why aren't they documented?
In any case, I think this is nonsense. If the problem had been the lack of a gratis client, and the wish to make one where none existed, then your allegation might make sense. But there was a gratis client, indeed it was the gratis client that made the entire issue come to a head.
All that happened was that Torvalds deliberately picked a proprietary SCM system, and other developers - who happen to have strong opinions on the ethics of using Free (as in liberty) vs Proprietary software - wanted to interoperate with Linus.
Nobody was trying to undermine BK's business model. They were trying to interoperate with a virtual collegue (Linus) without using proprietary software. If they wanted to undermine BK's business model, they could have continued to use the gratis client provided by the Bitkeeper authors.
Driver support, incidentally, existed for the vast majority of hardware that uses reverse engineered drivers in Linux. The drivers were for Windows. Why is not using Windows more ethical than not using Bitkeeper?
I realise it's (obviously!) not an ethical issue for you to use proprietary software. People are entitled to have different views on the subject. But it takes willful ignorance, contempt for those who think differently, not to mention a complete loss of logic, to ignore the fact that this was a Free (as in liberty) software issue, not about trying to get software without paying for it.
It's named the 360 because it's as big as the mainframe of the same name.
And Tridgell didn't reverse engineer something that already had a capable Free Software client. You know, the PC actually has a "capable and popular" operating system. So, presumably, it's absolutely wrong for the Linux developers to continue to use reverse engineering to develop a Free Software alternative. Right?
Wrong. This is about freedom. The Bitkeeper people had no business discouraging those who want to use free tools yet who want to interoperate with those who lack the same standards from writing free tools to spec. It was bad enough that the protocols were undocumented and proprietary to begin with. It's worse that this kind of vengeful stance was taken against third parties for daring to have an association with someone trying to create those free tools.
I'm absolutely amazed. He may not be the loudest proponent of FOSS in the world, but he's at least made himself look like such a proponent, and he has relied upon the very people he attacks doing exactly what he's attacking them for doing to make his kernel usable and what it is today.Your respect for him may have risen, mine has dropped. I was prepared to handle the fact he adopted BK in the first place because, well, people do often see themselves as pragmatic when making decisions that essentially defy good practices. He should have learnt something from this lesson, but essentially it looks like he's merely digging himself into a deeper hole while yelling "You all suck!" at those trying to get him out.
Look, what .NET does is hardly new. It's been done before countless times. Java is one example, smallTalk and FORTH are others. I've used FORTH. I've programmed in it. It's a good language. But like all of these systems, the IDE is constrained, ultimately, by the technology in use.
When I first found out about the UCSD p-System I thought "Great! Portability, a decent solid language (and Pascal was, for the time), a standardized front-end, etc." You could run the damned thing on anything that had a disk drive and a few kilobytes of RAM. But that's where things went horribly wrong.
At the time this came out, both were essentially rare. Your C64 was as likely hooked up to a tape deck as a floppy drive. Worse, the C64 had a whole bunch of crap that the p-System couldn't use. It had colour. It had graphics. It had sprites. It had sound. And to add that kind of functionality to p-System would have simply prevented it from running on many other platforms. Commodore, as you know, ended up going bankrupt.
This wasn't altogether surprising. I recall hearing about a project called "360" in the mid-sixties that attempted to create a single platform for mainframes. The idea was that the mainframes were all compatable, so you could port your software from one to the other pretty much by putting the cards in the other machine's bin. I bet you're not reading this. This was radical at the time, but it came at a price - the top end system was limited in functionality to that of the bottom end. Within fifteen years, the idea was dead, the systems lived on but co-existing with a range of incompatable systems from the same supplier. Within 25 years, the company concerned was teetering on the verge of bankrupsy.
I think you can see where I'm going with this, so let me get to the point. .NET is limited by its need to run on all current hardware and support virtually nothing original or innovative. And what goes right to the heard of the IDE. Without tools that are synergistic with the underlying technologies, programmers cannot develop great software. .NET hides the underlying technologies, limiting the scope of software development and denying programmers exposure to the very things that could inspire them.
This is ultimately why .NET will probably win out. With superior IDEs including the excellent Visual Studio.NET environment for software development, a consistant, well designed, underlying architecture, and fantastic expansion potential, I think we'll see .NET go from strength to strength. The work by people like Miguel de Icaza will help too, as programmers in other environments will be exposed to the latest Microsoft technologies, and will almost certainly be chomping at the bit to get on board.
But that can only happen with a well designed IDE. Unless software developers are comfortable when developing for .NET, they'll avoid it. But, looking at what they've done so far, and looking at their experience in this area, I don't think Microsoft has a thing to worry about.
As if to add insult to injury, Steve Jobs has apparently dealt with this inevitable delay by ordering that the BSD-underpinnings of the operating system be stripped out and replaced with GNU or Unix code, citing an internal study that shows BSD developers in complete disarray, with usage - based upon Usenet postings - showing a definite decline in BSD support. Perhaps if Apple ported Mac OS X to Intel, they wouldn't be suffering these problems.
On an unrelated note, I've just heard on Talk Radio that the famed computer hacker Stephen Wozniak has died, aged 54. Whether you enjoyed his Integer BASIC interpreter, you can't deny his contribution to modern computing. Truly an American icon.
Right now, the discussion is about legal practicality. We need a way to be able to combine code from different projects. If those projects are a mix of CDDL, MPL, et at, we can't do that. The only way to get this to work is if the projects are compatable in that their licenses are compatable with a common, neutral, license, that is the code licensed under the project-specific license can be relicensed under some other license. This is not possible, and will never be possible, with an overly liberal license like BSD or X11. Think practicality.
I appreciate this is a difficult subject to understand, especially on Slashdot where "The GPL is 3v1l! BSD is free!!!11!!" is taken as a non-ideological statement, but I want to give you a specific example that might make it easier to see what the issue is:
Suppose I want to take some Solaris code and put it in Darwin. The former is licensed under Sun's CDDL. The latter under Apple's APSL.
If we were to go to Apple and Sun, and asked them to modify their licenses so I could do this, legally, what do you think they'd want?
- To modify their licenses so it's possible to take their code and relicense it under a neutral copyleft, such as the GPL, preventing their code from being used in proprietary projects made by their competitors?
- To modify their licenses so it's possible to take their code and relicense it under a neutral non-copyleft, such as the X11 license, allowing their code from being used in proprietary projects made by their competitors?
If you said (2), you might want to look at both the CDDL and APSL. Both are essentially licenses that prevent anyone other than the "Initial developer" (ie Apple and Sun in the case of Darwin and Solaris) from putting code in proprietary projects. Why would they suddenly be happy with a loophole you could drive a train through being introduced into their licenses?Apologies for sounding pedantic here, but this is incorrect. The GPL cannot be incompatable with the CDDL and MPL without the MPL and CDDL being incompatable with it. It's obviously silly to describe one as incompatable with the other but not the other way around. The MPL and CDDL most certainly are incompatable with the GPL.
If you badly worded the above, and meant the emphasis should be on the GPL in terms of where the problems are coming, my entire posting addresses that: The GPL is neutral, it requires the imposition of no further restrictions, and therefore is one of the few players in town that qualifies as a legitimate baseline FOSS license for licenses to aspire compability with. The same is not true of the CDDL, MPL, or non-copyleft licenses such as the BSD and X11 licenses. The former are not project neutral. The latter do not prevent more conditions being attached, which essentially makes them useless from an interoperability point of view.
No, you can't.
The MPL is not compatable with the CDDL and vice versa. You cannot use code from an MPL project in a CDDL project. In fact, the licenses get legally problematic even when you try to combine code licensed under the same license from two different projects sourced by two different "Initial Developers". So not only is the MPL and CDDL incompatable with the GPL and each other, but frequently not even themselves!
There really isn't a lot of difference in the case of the MPL and CDDL. Both require that any relicensing be done under a license that imposes the same restrictions and grants. Essentially they're saying "You can relicense this under any license you want, as long as it's this one, or a rewrite." Most additional rights granted are, in any case, a re-statement of copyright law.
The real problem with both is that both try to elevate the initial developer to a special status. While it might be understandable, it's by design going to create problems getting multiple projects to cooperate. As soon as you say "This license does something special if {something that's project specific}", then you're into incompatable licenses.
In the situation you describe, the "fault" doesn't lie with the GPL. There's nothing stopping the developers - as a group, all of them who have contributed code - from saying "Yes, we'll do this closed source thing for this particular module", because the developers own the copyrights, and can agree, as a group, to dual license. This, of course, doesn't mean the developers will, it may be - shock horror - that the developers are opposed to their work being used in proprietary code, which leaves the company that wants to restrict the information used for the device driver the major stumbling block.
The CDDL, incidentally, doesn't allow this any more than the GPL does with the possible exception (and the CDDL is actually vague here, which makes matters worse) of the "initial developer" who gets some special rights. Both the MPL and CDDL are so-called "weak
Why is this important? Because most developers want compatability with some common license. If the MPL and CDDL were compatable with the GPL, then there would be ways of mixing code licensed under one with the other.
Not having compatability harms Freedom. It means that you cannot "use" code under circumstances that most Free Software advocates, and probably most Open Source advocates, would consider reasonable and a requirement for software to be considered Free or Open Source.
Yes, it doesn't have to the GPL. But there are no other real candidates. The BSD license doesn't do it because it has no requirement that further conditions be eliminated, so while BSD code can go in an MPL project, the fact the CDDL is compatable with the BSD license doesn't mean CDDL code can go into an MPL project.
Other licenses are, for the most part, more obscure and are no better than the GPL, or are not neutral. Some are even incompatable with themselves as they grant special privileges to an "Initial Developer", which means two projects under such licenses with different initial developers cannot share code.
So for the time being, GPL compatability needs, as far as many people are concerned, to be the baseline test of whether a license is a good one, regardless of whether - on the surface - it is "too political".
If you hate the fact that software can be called Free or Open Source and yet you can't use code from one project in another, you should be opposed to incompatable license proliferation. Incompatable licenses are killing Free Software.
The OSI has a list of "approved" licenses here, but it's literally just a list (no explanation of the differences or analysis of their implications), and I get the impression the OSI is regretting approving so many of them, given most of them are incompatable with one another (some of them can be incompatable with themselves, depending on the precise circumstances!)
Wikipedia, more in-depth discussion here.
I was using Microsoft networking back in the late eighties, on a network of something called a Research Machines Nimbus, at school. Oddly enough, we ran Windows 1.0 over it, with hilarious results (ie it was slow and crashed a lot.)
Anyway, the point is Microsoft's SMB protocols pre-date Windows. Windows interoperated with them without any problems, they were just DOS drives, after all.
You young'uns! You don't know how hard it was then! We used to have to wire coax to the back of PCs to get out Ethernet networks, kid!
</TONE>
I don't see this as a major problem. The point about patents is that they're there to encourage people to invest the time and money into making something new and real. What you're proposing - giving monopolies to people who will not spend the money - actually discourages innovation as it prevents or makes more expensive the doing of the real work to make something theoretical a reality.
...and those bastards who write the dictionaries are in on it too. It's a conspiracy I tell you, a conspiracy! Why I bet all the documents that were published before the RIAA was even created that used these terms are plants!
If we agree that you'll not kick me in the balls, and I'll not either, are you going to feel restrained from doing just that if I kick you in the balls anyway? Would you? Huh? Would you? Come here and prove it.
I'm not sure what brought that on. I think I'll go and have a little nap somewhere. Maybe some chamomile tea.
Neither were recent, so things may have deteriorated rapidly since then, but both were extremely comfortable then, if a little slow. Comparable to a prison bus? Well, I'll take you at your word that you're the one with the experience there (I'm fortunate and honest enough to never have been convicted of anything worse than a minor speeding offense), but compared to any Continental, Northwest, Delta, British Airways, or Virgin Atlantic flights I've taken, it was the height of luxury.
As for driving being only stressful and unpleasant because I "allow" it to be, I wonder if you have a clue what you're talking about? I can't exactly read a book while I drive. Or get up and go to the restroom. Or get some food and drinks. Or go for a stroll. I have to concentrate 100% on the road ahead, and I have to do so in a relatively cramped little box.
You made a sweeping claim about Amtrak which doesn't hold up in my experience. Given you didn't limit your claim to "potential (but not actual) Amtrak customers" then, I think it was reasonable for me to suggest your comment was supposed to be universal.I see you don't like Amtrak. Not everyone agrees with you.
Riding the train is a wonderfully relaxing way to travel. Just the ability to get up and walk around makes it an order of magnitude better than the alternatives. The decent sized seats, flat ride, lack of restrictions on when and where you can eat and drink, go for pee, etc, puts it into a class of its own. It's a massive shame that it's so restricted in the US.