They just have to word the laws right. Something along the lines of under no circumstances whatsoever shall patents be granted software, algorithms, business methods, or mathematical expressions, techniques or constructs.
Do they really need a gathering of lawyers to come up with that? It's not exactly rocket science.
This sounds more as if they're calling for a discussion on how to write pre-broken leglislation with full of carefully hidden loopholes. For something like that, I expect you'd get a lot of lawyers for a bash like that. Most of them with chequebooks to make sure the legislators were receptive to their clients views.
Not that I'm cynical or anything. I just don't see the difficulty in drafting such a law unless you're planning on adding "just kidding" on the end and hiding it under a pile of legalese
I think you need a better word than "infeasible". As it stands you seem to be saying that if the GPL only applies in cases that will never happen, and that if they somehow happen anyway, that can be used to prove that it wasn't really infeasible, and so the GPL didn't apply in any case.
I appreciate this is probably not what you meant, and I believe I agree with what I think you meant... but I'm not sure how to better articulate the idea, so it's difficult to be sure.
My PHP is a little rusty, but The analagous perl mechanism IIRC is "use" or "require".
As in your PHP example, you can effectively write "require" in terms of backticks, turning
require foo.pm
into
"$foo = `cat foo.pm`; eval "$foo";"
Now I'll concede that in such a case a case might be made (subject to my earlier qualifications) for the GPLification of foo.pm.
However -- and this is a big however -- the backtick mechanism is not the vector of contagion. The factor that may result in the invocation of the GPL is the inclusion of the contents of foo.pm into the script being executed. The backticks are being used to read the file being included, but the insertion of the code into the "address space" of the script is performed by the eval mechanism.
Backticks and function calls can be used to achieve similar results, and so I suppose it's tempting to consider that they do the same thing. However they work in profoundly different ways that make them quite distinct for GPL purposes.
Good point. I must confess, I hadn't realised that mere aggregation was being used in such a precise manner.
All the same, the backticks are still proprietory safe. Well, unless you are distributing someone else's GPL'd perl app and replacing one of their binary driver modules with a proprietory one of your own. Even then the FAQ suggests that intimate semantics and/or complex data structures would need to be involved.
I think the issue here is the way that language influences the way we think about things.
To use the term "intellectual property" suggests that ideas share certain properties in common with objects in the non-abstract, material world. In particular, it encourages us to consider that they tend to be owned by default unless specifically released into the public domain.
In fact the law is founded upon the opposite assumption: that ideas are form a commonwealth for all of humanity with creators being assigned a limited monopoloy over their implementation for a brief time as a reward for their efforts.
So I oppose intellectual property as a concept, and I also oppose the increasingly widespread abuse of those laws misleadingly referred to using that term.
None of this means I object to you making a living from your work. We may quibble over specifics of duration, but I accept unreservedly the priciple that you deserve some control over the knowledge brought forth by your labours.
My only objection is when you start mistaking that limited control for ownership. Ideas are the property of us all.
I would have to say intent. The intention of the backticks operator is to allow a perl script to run shell commands that may be completely unrelated to perl in any way whatsoever and to do so without modifing Perl itself in any manner at all.
if all you'd need was a clean interface to establish a separate work, then there would be no need for the LGPL at all, since the library would already be a separate work under the terms of the GPL.
We may be talking at cross purposes here. Running a program via perl's backtick operator isn't going to change the licence of the command invoked thereby. The GPL only kicks in if you modify and then distribute someone else's code.
On the other hand, suppose you were to take a GPLed application that used a mixture of perl and binary executables with the binaries called via backticks. Suppose further that you relace one of the original binaries with one of your own devising. Then, in that precise narrow circumstance, the application author might be able to call for your source to be released under the GPL.
If that's the scenario you meant, then yes, it could be considered a grey area.
As for libraries: they need a different licence because you implicitly distribute them when you distribute an application that has been compiled using them, and in order to distribute them, you are required to abide by the terms of the licence. It's a descision you take at the time of distribution, and nobody is forced into anything. Nevertheless, the LGPL makes this a non-issue in the majority of cases.
With backticks there is no distribution of anyone's code. All the backtick does is issue a command to the OS which may or may not be able to respond, and may respond with a variety of different executables depending on what packages are installed and the users PATH setting.
It may be that you had in mind some specific combination of authorship and distribution that I've failed to cover here. If so, I invite you to explain it in a bit more detailm since it's clearly gone over my head. If not, I think proprietory code is safe from the backticks operator, at least for this version of the GPL.
I'm not really sure whether your example, including proprietary code by calling the shell, is exactly the thing that can be called "mere aggregation".
Oh, really? If the backtick operator doesn't qualify as a mechanism of mere aggregation, then it's hard to imagine what could qualify.
The backtick operator can execute any executable on the system, to say nothing of arbitary shell scripts and sequences of shell commands. If backticks can "infect" code, so can bash. I don't think even the most fanatical free software zealot has ever pushed for that interpretation.
It seems to me that it's all a question of interfaces. If someone designs GPL software with a well defined interface, then you're free to write non GPL software that uses that interface. If there is no interface and you're changing the files in the original software, then the GPL applies.
Kernel modules are far more intimately connected to GPL software than an arbitatry executable is to Perl. And the Linux kernel development community seems count a healthy contingent of Free Software devotees amongst its number. Yet few if any people seem to consider that proprietory kernel modules are "tainted" by their association with the kernel. Quite the reverse in fact.
Nor does the compilation issue change matters. If I can write a compiler and use it to compile code that you have copyrighted. However there is no licence in existence under which that process of compilation grants me the the copyright to your code.
Elizabeth Boukis, spokeswoman for Sony Electronics, says the work is speculative. "There were not any experiments done," she says. "This particular patent was a prophetic invention. It was based on an inspiration that this may someday be the direction that technology will take us."
You see? They're still trying to control ideas. No material product involved
It's a logical progression. MS have already made progress pushing the idea of ideas as property. Now they just want to take it further and establish the notion that thinking about their ideas (presumably in an unlicenced manner) is theft.
The technology to tell what someone is thinking may never exist. All the same, the notion could prove useful to MS. Just let a generation grow up that will accept the notion of thoughts as property. You could "own" the ideas that constitute an operating system, say, and licence how people could and could not think them. And since discussion could be argued as proof of thought, they could make it a criminal offence to say unkind things about windows. Hey presto, no more bad reviews. Even private conversations would be actionable. Better yet, this being "theft", it would be a criminal case rather than a legal one.
Of course, we could expect this to require a certain amount of testing the courts, and probably some bespoke legislation. That shouldn't pose an insurmountable problem: imagine if politicians and political parties could licence ideas in their campaigns and dictate how they could be discussed.
All this is impossible at the moment: any such case would be laughed out of court. So the first step to changing that would be to raise a generation of kids that wouldn't laugh their socks off at the idea.
Obviously this is all IMHO. I am not privy to Microsoft long range strategy sessions, and, consistent as it would be with their usual business practices, it's entirely possible that they have never considered any such scenario.
And as James Turner points out, this is an ad hominem attack: unable to refute PJs arguments, supporters of SCO have taken to trying to slur PJ herself in order to discredit her data and her arguments.
Therefore I ask: are you aware of any factual inaccuracies on Groklaw? If so, did you make PJ aware of them? If so, did she amend or retract the article, if not did she give any justifcation.
And what, precisely was the disputed data?
PJ has co-orindated the collection of a body of high quality data. If you have evidence that contradicts that on groklaw, then I expect PJ will be glad to set the record straight. If you have better arguments, state them! She will probably print those too.
But if your best argument involves the lady's religion, her alleged employer, her work experience then you have nothing to say at all.
Well, it looks like we've been arguing at cross purposes then.
I had a look at the Get The Facts site. In its current incarnation at least, it's a lot less one sided than I'd been led to believe. On the other hand there's no shortage of headlines with MS marketing chief saying that "windows wins every time" and citing the site to support his assertion, so maybe its not too surprising that I got the wrong end of the stick.
It isn't all that different with Microsoft. Naturally there is only one supplier for the OS itself, but the direct relationship is usually with a systems vendor like Siemens, Dell, IBM, HP, etc, who each have the ability to influence Microsoft. There is also a lot of variety in device drivers, applications, development tools, etc. OS crashes/hangs, for example, are almost always caused by device drivers, so if the PC vendor doesn't offer reliable drivers, you can switch.
Try getting a version of windows without explorer:)
Being serious, the tight integration of the MS desktop seems to be the source of most of their security woes. Features that would have been really cool in an offline world turn out to be a liability in the context of the internet.
I'd like a version of windows that ran on top of a dos prompt again, so I could run a different window manager. That'll never happen. one reason MS went nuts about netscape was that they feared netscape would evolve into a desktop environment and they'd lose that level of control.
I'd like a registry that didn't encrypt half its information. Hell, I'd like to go back to config files I could read and edit and which never filled up or got corrupted. I'd like... well you get the picture. I've also wandered out of the context of support.
At the end of the day though, it's the same story. With Microsoft you get what you're given. With Linux, you choose. Not having those options is part of the opportunity cost of choosing MS.
Well, to me, at least:)
As for the OEMs, I wouldn't rate small buisness' chances of motivating Dell as much higher that MS.
On the other hand there are open source drivers for a lot of devices now. Intel themselves released the drivers for the IPW2100 wifi chips. Others have been reverse engineered, some now work better than the vendor supplied drivers. You have to do a bit of reearch first, but that should be part of the planning process.
Renting can actually be better if the money that would have been invested in the house can be put to use in a way that generates a higher return.
Yup, and a lot of large businesses rent property for corporate hospitality because they don't want to have to set up department specially to maintain the premises.
But if you could buy outright, I'd guess that you would. A business venture might come along to tempt you to chance the capital elsewhere, but property is still the best long term investment you're going to make. Like Mark Twain said "Buy land, they've stopped making it."
I still believe that long term, windows represents a poor choice for anyone. I understand why you don't share that view. Still, if I had to give a lng term recommendation, it would always be open source.
If the benefit of both groups of software is assumed to be exactly same (a very questionable assumption), then what matters is the economic cost. The cost of licensing fees for commercial software is obviously higher than for free software, but it isn't the only cost! There is also the cost of installing and configuring the software (either an accounting cost of paying someone to do it, or the opportunity cost of doing it instead of something else).
mmm... I've always granted you that MS apps (or apple, IBM or anyone else) can sometimes be the best choice for some particular circumstances. The trouble is that opportunity cost is intimately bound up with the goals and circumstances of the company or individual. It seems difficult to generalise given the inherent variation in opportunity cost, which in turn appears to be one of your points.
And yet, Microsoft do generalise in their TCO argument. They say that Linux is more expensive than windows -- end of story. The only way to make such an analysis, even a fraudulent one such I believe this to be, is do disregard the opportunity cost and only consider the accounting cost.
So I'm not sure I understand how you can defend their postion. Surely either they are wrong for not factoring in opportunity cost (which can't be done) or else analyses based purely on accounting cost are admissable, in which case my rebuttal is valid.
Of course, this assumes intellectual honesty on the part of the MS marketing team. I'll come back to that.
Timescale 2:
[...]
In the overall context of owning and running a system for, eg, three years, licensing fees are typically a very small part of the cost, with such things as staff salaries being far more important.
I've skipped over the issue of oportunity cost here, because they'd essentially be a rehash of those given for Timescale One
The staff salaries issue is a part of the argument that I've not so far addressed. MS contend that the cost of linux support personelle is higher because the skills are rarer. The opposing argument, which I cannot personally verify, is that you need fewer linux support staff because the system is prone to fewer systemic errors than windows.
What I can verify is that demand for linux personell is on the increase. As a contractor, I've seen demand for linux jobs increase over the last two years from "never hear of it" to "I understand you have linux experience".
Where there is demand, people will position themselves to satisfy it. And as the body of linux literate staff rises, so the costs of employing them will go down. Given all that, it doesn't seem reasonable to assume that staff costs for linux, if they are in fact higher, will remain that way.
As for electricity and machine room space, I think we can safely regard them as being equal for TCO purposes. I can't think of a single application that windows can do on a generic beige box but which requires Big Iron to do under linux. (Or vice versa, though I'm inclined to believe that scenario more readily:))
Timescale 3:
Once again, it is not only the hardware. There are costs that vary based on the OS (I'm not arguing here for one or the other, only that they vary), such as staff costs. The assumption that the benefit of different OSes is precisely the same is also not tenable, so a higher cost does not necessarily mean a lower net benefit.
Benefits are like opportunity cost - highly relative and effecively proof against generalisation.
If MS are trying to include one-off retraining costs as part of the normal cost of a system running Linux, over its lifetime, I don't think they'll fool many people. If they're rather saying the recurring costs of running a Windows system are lower than those of running a Linux system (or even marginally higher in the case of existing Windows customers where retraining costs would be very high), people will listen
If you can't provide a coherent argument as to why your argument is rational, and have to start dancing around notions of an 'arbiter of rationality' for the planet, I'd say that speaks for itself. If your argument is rational, you should be able to defend it.
See? That's better already. Instead of proclaiming my irrational bias as fact, you've provided a coherent argument as to why you feel this might be the case. Perhaps we could proceed in this new spirit of mutuality?
As for supporting my argument, I thought I had already laid out my position. Maybe I could have made myself clearer. Let's try a thought experiment.
Suppose you are a small business and you have a hardware upgrade planned. You can choose MS or Linux. In the short term Linux can be available for free, while MS require new OS licences, and new licences for the Offcie suite. Linux wins if only this timescale is considered. Even MS do not dispuite this. Let's call this Timescale One.
Timescale Two considers the total lifespan of the equipment. in this analysis we get to consider the costs of retraining staff and support people to use linux, possible migration issues, maybe a support contract with one of the commercial distros. Of course, these many of these same costs may apply for a migration, say, from 98SE to XP, and the cost of MS software licences still needs to be considered. It should also be noted that by ceasing to support a given OS, MS can pressure businesses to upgrading the version of windows in between planned hardare upgrades. This will again incurr licence costs for winodw, office and possible migration costs for the business as a whole. Similarly, if they have in house coding expertese, MS can inflict expense upon them by changing the implementation language as they recently did to VB. All that said, it is possible for MS to win over Timescale Two. This I do not dispute.
Timescale Three is ongoing starting with the next upgrade. A linux shop has no new costs other than hardware. Maybe a new deal with a distro if they've changed architecures and want to use binary packages, but that's not essential. After three years as a Linux house they have the in house experience such that the retraining expense does not need to be repeated, nor the costs of migration. In all likelihood they don't have to change their software at all.
With MS software however, the company pays the same costs as it did on the previous iteration. It retains the option to migrate to linux under the conditions already discussed, although the longer they remain with MS the more painful this migration may become due to increased entaglement of company data in proprietary formats and patchy supports for open standards. Nevertheless, the option remains open.
With subsequent iterations the situation is the same. The MS shop incurrs the same costs again and again, while the linux shop does not. On Timescale Three, Linux wins.
Timescale Two is very convenient for Microsoft because it allows them to misrepresent the one-time-only costs of migrating to linux as a repeating overhead similar in magnitude to those incurred by licencing MS software. However, as I hope I have persuaded you, it distorts the picture over the longer term. If you want to consider the matter on terms of opportunity cost, then the opportunity cost of choosing MS is being tied into another cycle of Microsofts predatory pricing and manipulation.
It may well be the case that standard accounting practice only considers Timescale Two. I dare say that is very useful in most cases. However business is above all pragmatic, if for no other reason than that business that are not rarely succeed. If a buisiness practice ceases to reflect accurately the reality of the market, that practice gets modified or junked in place of something that works better.
The open source movement is a new thing. It has never before been possible to create a commonwealth of resources that could be freely distributed at no disadvantage to anyone. A
You clearly think Linux is a better option for you than Windows. The question is, do you believe that Windows is, or can ever be, a better option that Linux? If so, you're agreeing with me. If not, it's quite obvious you're irrationally biased. Those of us who take the rational approach evaluate each PC purchase on its merits (net benefits over the expected lifetime), just like purchases of anything else.
As long as you stick to "ever can be" then we've never been in disagreement. Except maybe in so far as my belief that linux wins hands down over the long term. Since that's a scenario you refuse to consider, however, we can hardly disagree on it. I do have dificulties with your definition of irrational bias, at least to the extent that you seem to define it as not holding your opinion.
Where we do disagree is the question of whether the TCO argument, as presented by Microsoft, holds water. Specifically, I doubt that the total lifetime of one piece of equipment is the most rational timescale to evaluate those costs. I might take your point regarding the limits of accurate forcasting, if you hadn't been at such pains to stress that financial costs are far from the only consideration. In any case, this does not seem to be a prediction that requires six sigma confidence.
You've provided nothing to support the implicit suggestion that this would impact calculations of return on investment one way or the other.
MS have a vested interest in forcing you to pay for the same software over and over. No such pressure exists for linux. With MS you're going to pay for each upgrade over and above the nominal cost. Such has been my experience, as discussed previously in this thread. I might just as well ask what evidence you have to suggest that you can accurately predict benefits over a single machine lifetime. Unless you have a working crystal ball, it's all guesswork in the end.
The major advantage to considering TCO over a single machine lifetime would seem to be that it's the only time frame that accords MS's TCO argument even a shred of dignity. I expect we'll have to disagree on that. I trust we can do so without my being labeled as biased and irrational.
But if not, we can always have a discussion about who precisely it was that appointed you the sole arbiter of rationality for Planet Earth, can't we?
On the one side we have a company whose investors are accustomed to tremendous growth; one that finds the market for its product approaching saturation, and which has a track record of using its market dominance to force customers into unneeded and largely cosmetic changes.
On the other hand we have a body of enterprise grade software, with no corporate imperative to make you pay anything at all.
And you say you can't accurately predict the comparative costs beyond the lifetime of your current machine? It doesn't sound like something that requires superhuman foresight to me.
What were you saying earlier about "irrational bias?"
I'm not going to use my current PC for my whole life, and nor am I going to use Windows XP for my whole life, so I'm interested in the value over the period during which I'll be using them
But you are going to be using computers for your entire life, or so it seems reasonable to assume. Why are you so interested in the cost of using MS products over the lifetime of the hardware, and yet so disinterested in considering the same cost over a longer period? It sees inconsistent somehow.
MS have said, in effect: "it's not sensible to simply consider the purchase price, because buying linux may prove more expensive over the lifetime of the hardware". I can see the logic in that, but why not take it all the way: "It's not sensible to consider simply the lifetime of the machine, since buying microsoft products may prove much more expensive over the course of your lifetime".
If the logic was valid coming from MS, it should hold up when I use it too. Otherwise you're just trying to insist on a context that best supports your argument. (And in that case I have two counter examples - one longer term, and one shorter).
You know, the UK parliament has a nasty habit of trying out new laws with possibly unforseen consequences on Scotland before they inflict them on the rest of us. So the scots got the poll tax first, and they also got 24 hour boozing. They have a separate legal system which makes it possible to do this, you see.
So given that Blair his cronies are so thoroughly under the thumb of Bush and Co., do you ever wonder who might be using the UK as a test case?
Actually I prefer to think of it as covering lifetime of the person or enterprise using the software. Forced upgrade cycle, broken backwards compatibilty, software bloat... all these things cost me money. Not just on this purchase, but on the next, when I find myself locked in by MS proprietory formats, and the next when they change things around so all my perfectly functional peripherals need to be replaced, and the next, when I have to buy yet another version of office. Don't even get me started on their attempts to inflict a software leasing scheme on their customers.
Consider the TCO of MS software across your entire lifetime. That's you personally, not your current piece of hardware. That's the long term view. To consider the matter puerly in terms of a single piece of equipment, likey obsolete in a year or two, that is what I call short termist.
I don't think the malware issue is that cut and dried. Consider the firfox spoofing attack that used i18n charsets to create bogus links. There was a fix avalable as an extension for that within 24 hours IIRC. The workaround fix was posted almost straight away and a new version which wasn't vulnerable was available inside a week or so. It takes a month or two to get MS to even admit a problem exists. There are some (supposedly) they have stated they will never fix.
You may never get malware problems, technically savvy as you are. On the other hand, that does rather weaken your argument as windows as the platform of the ordinary user. On todays net your brand new XP box will get owned long before it finishes downloading the first service pack.
In general though, and as you say, we seem to be in broad agreement on a lot of points. I'm certainly no mad zealot to insist that everyone should use linux. All the same, I'd have thought the forced upgrade problem was a compelling argument for anyone. Maybe you have to be bitten by it a few times before you really feel its force.
Maybe we should have this discussion again, post longhorn perhaps, and see if you still feel the same way;)
The trouble with TCO is that Microsoft have brought the argument into disrepute. On the one hand there's their sponsor ship of skewed "independent" reports. On the other, their laughably one-sided and poorly named Get The Facts campaign. If you have to compare RedHat's most expensive corporate package running on a top of the line against OEM XP bundled on a mass market Dell, then you needn't expect to be taken any too seriously.
If a fair and even handed consideration of TCO made a compelling case for Microsoft, then I expect that MS would have used just such an argument. MS is not, after all, known for employing idiots. What they did was present a one-sided and distorted picture masquerading as a fair and even handed treatment. This strongly suggests that they couldn't make a convincing case without the use of deception. If the facts are on your side in the first place, then why risk calling into question the integrity of your argument?
As for those managers who are making rational and well informed descisions, well they seem to be switching to Linux. And the best indication that this is so lies is Microsoft's recent descision to support linux under virtualisation. This is not a thing they would do if they unless they feared losing sales to Linux.
For home use, I think part of the problem is that we get used to the inconveniences of the system we use reguarly. The big one on Windows would be malware. In your original post you said that , when troubleshooting friends' systems, you invariably found that poor third party device drivers were to blame. I too often troubleshoot friends' systems and by far and away the most common problem cause of problems I find is malware, followed by registry errors.
We tend to become accustomed to such aggravations over time, and come to accept them as part of the business of using a computer. On the other hand we change to a new system and all the aggravations are new and strange and scary, so we perceive them as being that much bigger than they really are.
And some of the aggaravtions we are accustomed to are severe. My father needed a new machine. So he bought a new e-machines box. About £250 it cost him. Then we got it back and found out that his canner would not work. Nor would his printer. And the copy of MS Office that he already owned refused to install on the new machine. And the retail copies of office didn't include publisher which he needed. Another £200 right there. Add it all up and that cheap new computer was looking to set him back more than a grand! This is an aggravation MS would like us to become accustomed to; to accept as part of the computing experience. I don't accept that.
In the case of my father, I set him up with linux. I had to do some searching to make the scanner work, and I never did get the framebuffer going so I could hide the scrolling boot messages behind a slice of eye candy. Everything else just worked.
That's a thousand quid versus nothing at all. Furthermore, since XP came bundled with the hardware, we can consider the cost of the OS in both cases to be zero. We saved 800 quid by removing XP. There's your total cost of ownership as far as I'm concerned.
At the end of the day what MS want are revenue streams. You'll never stop paying them, because their marketing strategy is based upon forcing you to pay and pay and pay again. Linux is under no such pressure. With linux's increasing adoption, linux drivers are getting better quickly and the shortage of linux competant support staff is rapidly being filled. MS on the other hand is never going to stop trying to force you into buying that next upgrade. And so long as that remains true, choosing windows based on TCO is at best an inheretly short-term strategy. Choosing linux on the other hand may well work in the short term, and wins hands down over time.
Slashdot Mirror
That's what I mean by a "just kidding" clause. If they wrote it clear and simple without qualifiers there'd be no loophole.
That and a right royal bollocking for certain members of the euro patent office. Maybe a rewritten set of guidelines for patent office employees.
It still seems like something that could be solved transparently and without the need for a lawyerfest.
Do they really need a gathering of lawyers to come up with that? It's not exactly rocket science.
This sounds more as if they're calling for a discussion on how to write pre-broken leglislation with full of carefully hidden loopholes. For something like that, I expect you'd get a lot of lawyers for a bash like that. Most of them with chequebooks to make sure the legislators were receptive to their clients views.
Not that I'm cynical or anything. I just don't see the difficulty in drafting such a law unless you're planning on adding "just kidding" on the end and hiding it under a pile of legalese
If it's such a small mattter, you'll hardly hesitate.
I appreciate this is probably not what you meant, and I believe I agree with what I think you meant... but I'm not sure how to better articulate the idea, so it's difficult to be sure.
As in your PHP example, you can effectively write "require" in terms of backticks, turning
intoNow I'll concede that in such a case a case might be made (subject to my earlier qualifications) for the GPLification of foo.pm.
However -- and this is a big however -- the backtick mechanism is not the vector of contagion. The factor that may result in the invocation of the GPL is the inclusion of the contents of foo.pm into the script being executed. The backticks are being used to read the file being included, but the insertion of the code into the "address space" of the script is performed by the eval mechanism.
Backticks and function calls can be used to achieve similar results, and so I suppose it's tempting to consider that they do the same thing. However they work in profoundly different ways that make them quite distinct for GPL purposes.
All the same, the backticks are still proprietory safe. Well, unless you are distributing someone else's GPL'd perl app and replacing one of their binary driver modules with a proprietory one of your own. Even then the FAQ suggests that intimate semantics and/or complex data structures would need to be involved.
To use the term "intellectual property" suggests that ideas share certain properties in common with objects in the non-abstract, material world. In particular, it encourages us to consider that they tend to be owned by default unless specifically released into the public domain.
In fact the law is founded upon the opposite assumption: that ideas are form a commonwealth for all of humanity with creators being assigned a limited monopoloy over their implementation for a brief time as a reward for their efforts.
So I oppose intellectual property as a concept, and I also oppose the increasingly widespread abuse of those laws misleadingly referred to using that term.
None of this means I object to you making a living from your work. We may quibble over specifics of duration, but I accept unreservedly the priciple that you deserve some control over the knowledge brought forth by your labours.
My only objection is when you start mistaking that limited control for ownership. Ideas are the property of us all.
Does that satisfy your objection?
I would have to say intent. The intention of the backticks operator is to allow a perl script to run shell commands that may be completely unrelated to perl in any way whatsoever and to do so without modifing Perl itself in any manner at all.
We may be talking at cross purposes here. Running a program via perl's backtick operator isn't going to change the licence of the command invoked thereby. The GPL only kicks in if you modify and then distribute someone else's code.
On the other hand, suppose you were to take a GPLed application that used a mixture of perl and binary executables with the binaries called via backticks. Suppose further that you relace one of the original binaries with one of your own devising. Then, in that precise narrow circumstance, the application author might be able to call for your source to be released under the GPL.
If that's the scenario you meant, then yes, it could be considered a grey area.
As for libraries: they need a different licence because you implicitly distribute them when you distribute an application that has been compiled using them, and in order to distribute them, you are required to abide by the terms of the licence. It's a descision you take at the time of distribution, and nobody is forced into anything. Nevertheless, the LGPL makes this a non-issue in the majority of cases.
With backticks there is no distribution of anyone's code. All the backtick does is issue a command to the OS which may or may not be able to respond, and may respond with a variety of different executables depending on what packages are installed and the users PATH setting.
It may be that you had in mind some specific combination of authorship and distribution that I've failed to cover here. If so, I invite you to explain it in a bit more detailm since it's clearly gone over my head. If not, I think proprietory code is safe from the backticks operator, at least for this version of the GPL.
Oh, really? If the backtick operator doesn't qualify as a mechanism of mere aggregation, then it's hard to imagine what could qualify.
The backtick operator can execute any executable on the system, to say nothing of arbitary shell scripts and sequences of shell commands. If backticks can "infect" code, so can bash. I don't think even the most fanatical free software zealot has ever pushed for that interpretation.
It seems to me that it's all a question of interfaces. If someone designs GPL software with a well defined interface, then you're free to write non GPL software that uses that interface. If there is no interface and you're changing the files in the original software, then the GPL applies.
Kernel modules are far more intimately connected to GPL software than an arbitatry executable is to Perl. And the Linux kernel development community seems count a healthy contingent of Free Software devotees amongst its number. Yet few if any people seem to consider that proprietory kernel modules are "tainted" by their association with the kernel. Quite the reverse in fact.
Nor does the compilation issue change matters. If I can write a compiler and use it to compile code that you have copyrighted. However there is no licence in existence under which that process of compilation grants me the the copyright to your code.
You see? They're still trying to control ideas. No material product involved
The technology to tell what someone is thinking may never exist. All the same, the notion could prove useful to MS. Just let a generation grow up that will accept the notion of thoughts as property. You could "own" the ideas that constitute an operating system, say, and licence how people could and could not think them. And since discussion could be argued as proof of thought, they could make it a criminal offence to say unkind things about windows. Hey presto, no more bad reviews. Even private conversations would be actionable. Better yet, this being "theft", it would be a criminal case rather than a legal one.
Of course, we could expect this to require a certain amount of testing the courts, and probably some bespoke legislation. That shouldn't pose an insurmountable problem: imagine if politicians and political parties could licence ideas in their campaigns and dictate how they could be discussed.
All this is impossible at the moment: any such case would be laughed out of court. So the first step to changing that would be to raise a generation of kids that wouldn't laugh their socks off at the idea.
Obviously this is all IMHO. I am not privy to Microsoft long range strategy sessions, and, consistent as it would be with their usual business practices, it's entirely possible that they have never considered any such scenario.
Nevertheless, I'm not laughing.
Therefore I ask: are you aware of any factual inaccuracies on Groklaw? If so, did you make PJ aware of them? If so, did she amend or retract the article, if not did she give any justifcation.
And what, precisely was the disputed data?
PJ has co-orindated the collection of a body of high quality data. If you have evidence that contradicts that on groklaw, then I expect PJ will be glad to set the record straight. If you have better arguments, state them! She will probably print those too.
But if your best argument involves the lady's religion, her alleged employer, her work experience then you have nothing to say at all.
Well, it looks like we've been arguing at cross purposes then.
I had a look at the Get The Facts site. In its current incarnation at least, it's a lot less one sided than I'd been led to believe. On the other hand there's no shortage of headlines with MS marketing chief saying that "windows wins every time" and citing the site to support his assertion, so maybe its not too surprising that I got the wrong end of the stick.
It isn't all that different with Microsoft. Naturally there is only one supplier for the OS itself, but the direct relationship is usually with a systems vendor like Siemens, Dell, IBM, HP, etc, who each have the ability to influence Microsoft. There is also a lot of variety in device drivers, applications, development tools, etc. OS crashes/hangs, for example, are almost always caused by device drivers, so if the PC vendor doesn't offer reliable drivers, you can switch.
Try getting a version of windows without explorer :)
Being serious, the tight integration of the MS desktop seems to be the source of most of their security woes. Features that would have been really cool in an offline world turn out to be a liability in the context of the internet.
I'd like a version of windows that ran on top of a dos prompt again, so I could run a different window manager. That'll never happen. one reason MS went nuts about netscape was that they feared netscape would evolve into a desktop environment and they'd lose that level of control.
I'd like a registry that didn't encrypt half its information. Hell, I'd like to go back to config files I could read and edit and which never filled up or got corrupted. I'd like ... well you get the picture. I've also wandered out of the context of support.
At the end of the day though, it's the same story. With Microsoft you get what you're given. With Linux, you choose. Not having those options is part of the opportunity cost of choosing MS. Well, to me, at least :)
As for the OEMs, I wouldn't rate small buisness' chances of motivating Dell as much higher that MS. On the other hand there are open source drivers for a lot of devices now. Intel themselves released the drivers for the IPW2100 wifi chips. Others have been reverse engineered, some now work better than the vendor supplied drivers. You have to do a bit of reearch first, but that should be part of the planning process.
Renting can actually be better if the money that would have been invested in the house can be put to use in a way that generates a higher return. Yup, and a lot of large businesses rent property for corporate hospitality because they don't want to have to set up department specially to maintain the premises.
But if you could buy outright, I'd guess that you would. A business venture might come along to tempt you to chance the capital elsewhere, but property is still the best long term investment you're going to make. Like Mark Twain said "Buy land, they've stopped making it."
I still believe that long term, windows represents a poor choice for anyone. I understand why you don't share that view. Still, if I had to give a lng term recommendation, it would always be open source.
mmm... I've always granted you that MS apps (or apple, IBM or anyone else) can sometimes be the best choice for some particular circumstances. The trouble is that opportunity cost is intimately bound up with the goals and circumstances of the company or individual. It seems difficult to generalise given the inherent variation in opportunity cost, which in turn appears to be one of your points.
And yet, Microsoft do generalise in their TCO argument. They say that Linux is more expensive than windows -- end of story. The only way to make such an analysis, even a fraudulent one such I believe this to be, is do disregard the opportunity cost and only consider the accounting cost.
So I'm not sure I understand how you can defend their postion. Surely either they are wrong for not factoring in opportunity cost (which can't be done) or else analyses based purely on accounting cost are admissable, in which case my rebuttal is valid.
Of course, this assumes intellectual honesty on the part of the MS marketing team. I'll come back to that.
Timescale 2:
[...]
In the overall context of owning and running a system for, eg, three years, licensing fees are typically a very small part of the cost, with such things as staff salaries being far more important.
I've skipped over the issue of oportunity cost here, because they'd essentially be a rehash of those given for Timescale One
The staff salaries issue is a part of the argument that I've not so far addressed. MS contend that the cost of linux support personelle is higher because the skills are rarer. The opposing argument, which I cannot personally verify, is that you need fewer linux support staff because the system is prone to fewer systemic errors than windows.
What I can verify is that demand for linux personell is on the increase. As a contractor, I've seen demand for linux jobs increase over the last two years from "never hear of it" to "I understand you have linux experience".
Where there is demand, people will position themselves to satisfy it. And as the body of linux literate staff rises, so the costs of employing them will go down. Given all that, it doesn't seem reasonable to assume that staff costs for linux, if they are in fact higher, will remain that way.
As for electricity and machine room space, I think we can safely regard them as being equal for TCO purposes. I can't think of a single application that windows can do on a generic beige box but which requires Big Iron to do under linux. (Or vice versa, though I'm inclined to believe that scenario more readily :))
Timescale 3:
Once again, it is not only the hardware. There are costs that vary based on the OS (I'm not arguing here for one or the other, only that they vary), such as staff costs. The assumption that the benefit of different OSes is precisely the same is also not tenable, so a higher cost does not necessarily mean a lower net benefit.
Benefits are like opportunity cost - highly relative and effecively proof against generalisation.
If MS are trying to include one-off retraining costs as part of the normal cost of a system running Linux, over its lifetime, I don't think they'll fool many people. If they're rather saying the recurring costs of running a Windows system are lower than those of running a Linux system (or even marginally higher in the case of existing Windows customers where retraining costs would be very high), people will listen
See? That's better already. Instead of proclaiming my irrational bias as fact, you've provided a coherent argument as to why you feel this might be the case. Perhaps we could proceed in this new spirit of mutuality?
As for supporting my argument, I thought I had already laid out my position. Maybe I could have made myself clearer. Let's try a thought experiment.
Suppose you are a small business and you have a hardware upgrade planned. You can choose MS or Linux. In the short term Linux can be available for free, while MS require new OS licences, and new licences for the Offcie suite. Linux wins if only this timescale is considered. Even MS do not dispuite this. Let's call this Timescale One.
Timescale Two considers the total lifespan of the equipment. in this analysis we get to consider the costs of retraining staff and support people to use linux, possible migration issues, maybe a support contract with one of the commercial distros. Of course, these many of these same costs may apply for a migration, say, from 98SE to XP, and the cost of MS software licences still needs to be considered. It should also be noted that by ceasing to support a given OS, MS can pressure businesses to upgrading the version of windows in between planned hardare upgrades. This will again incurr licence costs for winodw, office and possible migration costs for the business as a whole. Similarly, if they have in house coding expertese, MS can inflict expense upon them by changing the implementation language as they recently did to VB. All that said, it is possible for MS to win over Timescale Two. This I do not dispute.
Timescale Three is ongoing starting with the next upgrade. A linux shop has no new costs other than hardware. Maybe a new deal with a distro if they've changed architecures and want to use binary packages, but that's not essential. After three years as a Linux house they have the in house experience such that the retraining expense does not need to be repeated, nor the costs of migration. In all likelihood they don't have to change their software at all.
With MS software however, the company pays the same costs as it did on the previous iteration. It retains the option to migrate to linux under the conditions already discussed, although the longer they remain with MS the more painful this migration may become due to increased entaglement of company data in proprietary formats and patchy supports for open standards. Nevertheless, the option remains open.
With subsequent iterations the situation is the same. The MS shop incurrs the same costs again and again, while the linux shop does not. On Timescale Three, Linux wins.
Timescale Two is very convenient for Microsoft because it allows them to misrepresent the one-time-only costs of migrating to linux as a repeating overhead similar in magnitude to those incurred by licencing MS software. However, as I hope I have persuaded you, it distorts the picture over the longer term. If you want to consider the matter on terms of opportunity cost, then the opportunity cost of choosing MS is being tied into another cycle of Microsofts predatory pricing and manipulation.
It may well be the case that standard accounting practice only considers Timescale Two. I dare say that is very useful in most cases. However business is above all pragmatic, if for no other reason than that business that are not rarely succeed. If a buisiness practice ceases to reflect accurately the reality of the market, that practice gets modified or junked in place of something that works better.
The open source movement is a new thing. It has never before been possible to create a commonwealth of resources that could be freely distributed at no disadvantage to anyone. A
As long as you stick to "ever can be" then we've never been in disagreement. Except maybe in so far as my belief that linux wins hands down over the long term. Since that's a scenario you refuse to consider, however, we can hardly disagree on it. I do have dificulties with your definition of irrational bias, at least to the extent that you seem to define it as not holding your opinion.
Where we do disagree is the question of whether the TCO argument, as presented by Microsoft, holds water. Specifically, I doubt that the total lifetime of one piece of equipment is the most rational timescale to evaluate those costs. I might take your point regarding the limits of accurate forcasting, if you hadn't been at such pains to stress that financial costs are far from the only consideration. In any case, this does not seem to be a prediction that requires six sigma confidence.
You've provided nothing to support the implicit suggestion that this would impact calculations of return on investment one way or the other.
MS have a vested interest in forcing you to pay for the same software over and over. No such pressure exists for linux. With MS you're going to pay for each upgrade over and above the nominal cost. Such has been my experience, as discussed previously in this thread. I might just as well ask what evidence you have to suggest that you can accurately predict benefits over a single machine lifetime. Unless you have a working crystal ball, it's all guesswork in the end.
The major advantage to considering TCO over a single machine lifetime would seem to be that it's the only time frame that accords MS's TCO argument even a shred of dignity. I expect we'll have to disagree on that. I trust we can do so without my being labeled as biased and irrational.
But if not, we can always have a discussion about who precisely it was that appointed you the sole arbiter of rationality for Planet Earth, can't we?
On the one side we have a company whose investors are accustomed to tremendous growth; one that finds the market for its product approaching saturation, and which has a track record of using its market dominance to force customers into unneeded and largely cosmetic changes.
On the other hand we have a body of enterprise grade software, with no corporate imperative to make you pay anything at all.
And you say you can't accurately predict the comparative costs beyond the lifetime of your current machine? It doesn't sound like something that requires superhuman foresight to me.
What were you saying earlier about "irrational bias?"
But you are going to be using computers for your entire life, or so it seems reasonable to assume. Why are you so interested in the cost of using MS products over the lifetime of the hardware, and yet so disinterested in considering the same cost over a longer period? It sees inconsistent somehow.
MS have said, in effect: "it's not sensible to simply consider the purchase price, because buying linux may prove more expensive over the lifetime of the hardware". I can see the logic in that, but why not take it all the way: "It's not sensible to consider simply the lifetime of the machine, since buying microsoft products may prove much more expensive over the course of your lifetime".
If the logic was valid coming from MS, it should hold up when I use it too. Otherwise you're just trying to insist on a context that best supports your argument. (And in that case I have two counter examples - one longer term, and one shorter).
You know, the UK parliament has a nasty habit of trying out new laws with possibly unforseen consequences on Scotland before they inflict them on the rest of us. So the scots got the poll tax first, and they also got 24 hour boozing. They have a separate legal system which makes it possible to do this, you see.
So given that Blair his cronies are so thoroughly under the thumb of Bush and Co., do you ever wonder who might be using the UK as a test case?
I'll just go and put my tin foil hat omn...
Consider the TCO of MS software across your entire lifetime. That's you personally, not your current piece of hardware. That's the long term view. To consider the matter puerly in terms of a single piece of equipment, likey obsolete in a year or two, that is what I call short termist.
You may never get malware problems, technically savvy as you are. On the other hand, that does rather weaken your argument as windows as the platform of the ordinary user. On todays net your brand new XP box will get owned long before it finishes downloading the first service pack.
In general though, and as you say, we seem to be in broad agreement on a lot of points. I'm certainly no mad zealot to insist that everyone should use linux. All the same, I'd have thought the forced upgrade problem was a compelling argument for anyone. Maybe you have to be bitten by it a few times before you really feel its force.
Maybe we should have this discussion again, post longhorn perhaps, and see if you still feel the same way ;)
heheheh. I don't really have to explain that it's a figure of speech, do I?
At least they're not touting this as an anti-terrorist measure. I guess that's no longer as credible as it used to be.
If a fair and even handed consideration of TCO made a compelling case for Microsoft, then I expect that MS would have used just such an argument. MS is not, after all, known for employing idiots. What they did was present a one-sided and distorted picture masquerading as a fair and even handed treatment. This strongly suggests that they couldn't make a convincing case without the use of deception. If the facts are on your side in the first place, then why risk calling into question the integrity of your argument?
As for those managers who are making rational and well informed descisions, well they seem to be switching to Linux. And the best indication that this is so lies is Microsoft's recent descision to support linux under virtualisation. This is not a thing they would do if they unless they feared losing sales to Linux.
For home use, I think part of the problem is that we get used to the inconveniences of the system we use reguarly. The big one on Windows would be malware. In your original post you said that , when troubleshooting friends' systems, you invariably found that poor third party device drivers were to blame. I too often troubleshoot friends' systems and by far and away the most common problem cause of problems I find is malware, followed by registry errors.
We tend to become accustomed to such aggravations over time, and come to accept them as part of the business of using a computer. On the other hand we change to a new system and all the aggravations are new and strange and scary, so we perceive them as being that much bigger than they really are.
And some of the aggaravtions we are accustomed to are severe. My father needed a new machine. So he bought a new e-machines box. About £250 it cost him. Then we got it back and found out that his canner would not work. Nor would his printer. And the copy of MS Office that he already owned refused to install on the new machine. And the retail copies of office didn't include publisher which he needed. Another £200 right there. Add it all up and that cheap new computer was looking to set him back more than a grand! This is an aggravation MS would like us to become accustomed to; to accept as part of the computing experience. I don't accept that.
In the case of my father, I set him up with linux. I had to do some searching to make the scanner work, and I never did get the framebuffer going so I could hide the scrolling boot messages behind a slice of eye candy. Everything else just worked.
That's a thousand quid versus nothing at all. Furthermore, since XP came bundled with the hardware, we can consider the cost of the OS in both cases to be zero. We saved 800 quid by removing XP. There's your total cost of ownership as far as I'm concerned.
At the end of the day what MS want are revenue streams. You'll never stop paying them, because their marketing strategy is based upon forcing you to pay and pay and pay again. Linux is under no such pressure. With linux's increasing adoption, linux drivers are getting better quickly and the shortage of linux competant support staff is rapidly being filled. MS on the other hand is never going to stop trying to force you into buying that next upgrade. And so long as that remains true, choosing windows based on TCO is at best an inheretly short-term strategy. Choosing linux on the other hand may well work in the short term, and wins hands down over time.