I prefer the idea of porting the NetBSD to Linux and dumping GNU dependencies.
It's a free world. Oh, wait...
Seriously though: sure, whatever jiggles your bits.
You do bring up an interesting point, however: NetBSD is, of course, generally considered more than just the kernel, the same way that "Red Hat Linux" is more than just the Linux kernel.
Using my nomenclature, NetBSD/NetBSD would be awkward, and, if OS vs. kernel were clear from context, NetBSD alone would suffice for either one. NetBSD/Linux, of course, would refer to the "standard" parts of NetBSD on a Linux kernel.
But, I have to ask, could you really get away from GNU completely and still have a relatively complete O/S if you did that?
There's a ton of home-brew linux OS's out there. Why not add your UYM$/OS to the fray?:)
Why not, indeed.
First, why reinvent the wheel?
Second, there may be good reasons to. Gentoo, for example, likely fills an interesting need: auditable, tuned distribution generation, for one. But, unless your personal distro is likely to have a wider audience, or your need is strong, why bother?
I do think the <packager> <raw source package set>/<kernel> naming convention is a good one, though I admit that <packager> and <raw source package set> might be redundant, together. I do see <raw source package set> evolving into a place holder for a standard for same, thus, "Red Hat GNU/Linux" would adhere to the same common GNU standard that "Debian GNU/Linux" and "Foo GNU/Linux" would. That is not the state of affairs today.
Furthermore, reducing the GNU moniker to a standards certification might not make RMS all that happy, unless the standard is one of all software being free. Unfortunately, that is not the purpose of the standardization in this context: one of features as opposed to freedom. It would also require the moniker to be trademarked, and used under license, presumably from the FSF, to have any teeth. Perhaps "GSB" (GNU Standard Base) would be a better monker in that case, leaving GNU free to be used by anyone chosing to express software freedom ideals.
As for the mayhem that would result from a plethora of distros using such a naming convention, I think such a fear is overblown: only a few would be popular, existing trademark law could be used to stem confusion, at least for commercial distros. Besides, the Sun hack of using.a.domain.name.as.part.of.a.distro.name.compo nent.mydomain.org could always be used, as ugly as that would get (hmm, the logical equivalent of CNAMES in product branding?/me smells a Bezos-style patent).
Yes, stability and ease of upgrade are more difficult than halfhazard addition of the latest bleeding edge components.
I had addressed another aspect of the kind of structured thought that has to go into those efforts, and that is portability, which also takes time and skill.
Will this make "GNU/Linux" more acceptable?
on
Debian NetBSD for Sparc
·
· Score: 4, Interesting
As greater mixing and matching of operating system (minus kernel) and kernel happens, perhaps designations like GNU/Linux will be more acceptable, for their technical precision.
However, then, it wouldn't be "Debian/NetBSD" but rather "Debian GNU/NetBSD" (Debian, at least, having already accepted the GNU/Linux moniker).
I really think there are practical advantages to distinguishing between 1) kernel, 2) "everything else", i.e. portable packages ported to run on that kernel, and with each other, and 3) who did the porting/distribution bundling.
Of course, when there is only one varient of one of those three components, i.e. the bits and pieces of what would be GNU, we tend to leave it out, as cumbersome, redundant, and unwieldly, RMS protests of the need to pay hommage (or at least use terminology that conveys GNU roots) notwithstanding. Thus, "Red Hat Linux": there is no other kind of "Linux" packaged by Red Hat (and since they did the bundling, they get to call it pretty much whatever they want (personally, I would have preferred "Up Yours, MS"/OS, but anyway...)), and no techical need for a "Red Hat GNU/Linux" designation.
RMS may want to see us embrace a phisosophical basis for free software, but, without the economic benefits that open source exemplifies, I doubt free software would have the contributions it does. Similarly, without a technical argument for "GNU/Linux", the moniker will likely not be popular. It would be nice, though, if the technical argument were there, so the philosophy and history could get some recognition and representation in common use.
I expect that might come in niche markets: where Linux is combined with severly scaled down portions of GNU, and non-GNU software, particularly in the embedded market: look at BusyBox -- a combination of utilities in a single executible for space reasons. We are seeing attempts to standardize "Linux" for the desktop (LSB), as well as for the embedded space. I, for one, wouldn't mind seeing LSB (Linux Standard Base) become GSB (GNU Standard Base), with LSB dealing solely with a kernel standard, in this vein.
A parallel port (or even using the control line from a serial port) and an IR led will work, though you'll have to bit bang it, and may need external power for amplification (or steal it from a USB port, he he).
Alternately, Xantech makes RS232 to IR converters which are designed to translate a character code sent over an RS232 line (i.e. your computer's serial port) to an IR signal which can be broadcast. However, the unit is quite large, and definitely requires an external power supply.
Since you apparently didn't even bother to do a rudimentary Google search for "RS232 to IR", or you would know this already, doing that is left as an excercize.
When I say "should" I mean just that: a recommendation, in this case, of the ethics of a situation, no more and no less.
How this is perverted to be perceived as promulgation of a particular personal agenda is beyond me.
In this case, I encourage taking a position, one way or the other, and defending it. Nowehere do I argue that Linus is "wrong", bur rather wishy-washy.
Re:I'm sorry, I can't let that one past!
on
Linus on DRM
·
· Score: 2, Interesting
are you equating state funded healthcare with child labour and the holocaust?! If you are then... well it's just incredible. I'm speechless.
No, of course not. I'm simply using extreme examples to illustrate that there are ethical issues worth considering. Because of the disruptive effects of open/free software, it should be examined from an ethical context.
For the record, if you have state funded healthcare, the rich, not the poor, pay for it disproportionately. So you're never going to be in a position where your taxes take enough money to leave you vulnerable to 'medical needs not covered by the "free" program'
Funny. That's exactly how my father died: he paid so much in taxes he could not afford essential surgery (about US$25,000 to US$50,000 with a 30% mortality rate -- his situation was grave), nor even the insurance that would cover it. Canada, of course, had the excuse that there were no qualified Canadian doctors to perform it, so, "Sorry, you die.".
I'm told they do send some people to the U.S. for care, now, but many still die that could be saved if they had bought real insurance and not the government bogus crap that was shoved down their throats as part of their taxes.
Avoid state healthcare like the plague!
Re:Self defense vs. free software?
on
Linus on DRM
·
· Score: 1
First, I think that it's a reach to think that anyone involved with linux will actually be prosecuted/persecuted simply for developing an OS.
Honestly, I think so too.
But, stranger things have happened, and if one is going to stick one's neck out, one may as well stick it out all the way. Though, that is one prediction I wouldn't be happy to "I told ya so!" about if it came to pass.
Re:I'll have one OS, hold the politics.
on
Linus on DRM
·
· Score: 1
'Linus be a "fair weather" friend to the Linux community, quietly disappearing into obscurity? I hope not, but, sadly, I'm not sure. '
Watch that you don't confuse GNU with Linux as Stallman is ever quick to point out
I mean the community of users and developers around the Linux kernel.
However, GNU/Linux is an empowering operating system, that is disruptive to existing closed source interests, and can, no doubt (as can many things), be used to facilitiate hosting applications (particularly distributed, networked applications), that are favorable to "terrorism".
Linux is the heart of GNU/Linux, and Linus is the heart of Linux. Is it so far-fetched to think that the U.S. might wish to make an "example" out of him as "supporting terrorism", given the Constitution-shreading they've done since 9/11/2001?
If that happen to RMS, he'd not go down without kicking, screaming, generally making an annoying, attention-getting fuss, and making us think, "ya know, he's annoying, but doing that to him is wrong".
I can see Linus trying to "pull off a Ghandi" in the same predicament, in the best of circumstances, but, he doesn't come off even as confrontational enough for that, and would probably silently "let go" of Linux if the state guns were at his head (or, more likely, the heads of his wife and kids).
Yes, I am a cynic, and somewhat paranoid. Obviously, I hope that such a scenario does not unfold. But, I'd feel a lot more comfortable if Linus showed a bit more spine at times rather than constantly letting the water roll off his back, like a duck. Ducks get shot in hunting season.
For a guy who claims to be oh-so-concerned with ethics, your position on piracy is... well, let's just say it's *surprising.*
How so?
First, ethics and law often diverge.
While I recognize the notions of copyright, trademark, and patent, I also note that the copyright and patent systems in the U.S. are highly broken. Thus, as far as the U.S. Constitution is concerned (and admitedly, it is but one legal framework), certain forms of *piracy* may very well be legal. I know I am a strong defender of fair use, and object to excessive copyright terms. Whether that makes me a law-breaker because I (a) back up my CDs and DVDs, and (b) use DeCSS to facilitate distribution of movies I have licensed, depends on how one interprets the law, and when -- perhaps the DMCA will, one day, be thrown out as unconsititutional. If that happens, present illegal violations of it would, legally, never happen (one of the rare cases where a law can be erased retroactively). But, this does not change the ethical notion of "not stealing", just because the question of what is property arises, and how, other than by contract, is property recognized (hint: state, i.e. mob, fiat).
For the record, I do not "pirate" movies or music, and have paid for every videocassette, DVD, and CD in my collection.
However, the ethics of ownership of thoughts and ideas are not clear: if "piracy" is so rampant, should it not be O.K.? If you want to keep something secret, well,... keep it secret. In the absence of explicit contract that certain actions are a breech, ethically, those actions (to a contractarian, anyway), are neither moral, nor immoral, they are amoral.
Copyright and patents, are legal fabrications that make non-tangible ideas ownable, presumably to stimulate their creation and reward. A tricky balance between innovation and exploitation is necessary, and the present state of this balance drives, among other things, "pirating" behaviour.
So, no, I do not encourage, nor do I practice, what would be considered (by most -- the law might disagree on technical details, vis. use of DeCSS while in the U.S.A.) "piracy". However, I note that "piracy" is a highly disruptive social and economic force, and therefore software that can support it is, by extention, powerful, as a social mover. Powerful things should be examined in their ethical context.
On the other side of the coin, the whole internet can be viewed as leveraging very powerful software to effect societal change. While generally considered a constructive application of software, it too has a "dark side" that deserves ethical review (spam, porn, particularly exploitive (i.e. kiddie) porn).
My main objection to Linus is that he comes off as a little to blase with regard to potentially ethical aspects of Linux. I am not suggesting that he take a binary yes/no stance with regard to things like "DRM support in the kernel" (whatever that might mean), but rather express his views about DRM with a bit more than a virtual shrug.
However, like a firearm, software can be used for constructive as well as disruptive purposes, and to a surprising degree (though, the disruptive ones tend to be the ones making the news).
Thus, software raises significant ethical issues.
I think Linus shrugs them off a little too lightly. It isn't so much a DRM vs. "DRM verbotten" issue. It's more along the lines of "If DRM matters, what kinds of DRM would we think are O.K., and what aren't? And, are there things that we can do to encourage or facilitate the 'O.K.' kinds?". Linus appears to take the "not my problem" stance.
Would we consider support for, say, signed or encrypted modules in the kernel as a good thing? Perhaps, espescially if it could be excluded, so individuals could chose. Now, that's so defeatable as to not be strong DRM, but could still be leveraged in embedded applications in a crippling fashion -- remember most users don't build custom kernels, and what's easy for us isn't for the masses.
Okay, *you* may happen to feel that way, but you also aren't in the driver's seat. Canada has a very powerful ally that wants the war in Iraq to go through, and a huge number of citizens that don't.
Actually, it's the other way around: many citizens here are strong supporters of U.S. action, Canadians have served on the front lines (two have died, actually), yet the government is "not involved".
think I speak for all of Slashdot here when I say:
Fortunately, you don't,... and what was that about proportion?
This is software we're talking about, for chrissakes! Not sweatshop labor, not Nazi death camps, not... whatever that third thing you said was supposed to mean.
Ethical issues surround us. Some of them are nasty, hence the extreme examples I provide. Ergo, one should have an interest in ethics. Lacking an interest even in the ethics surrounding one's own creation suggests an apathy of a degree that even extreme issues don't matter.
Furthermore, history, and recent events in the U.S., suggest that the difference between extreme measures, and apparently innocent things like software is not as large as one might like, making taking an ethical stance even nore compelling (one way or the other).
(You seem to be saying that doctors shouldn't be paid for their services.
No, I am saying that forcefully taking money from one to pay doctors that serve another, to the point where the individuals so impoverished can no longer purchase the medical care they need, is wrong.
When you talk about software in the same breath as Mengele, you're demonstrating your lack of a sense of proportion.
If you check, I didn't, but simply used the example of Mengele as substantiation of an earlier claim.
Software is becoming an extremely powerful force on this planet: either empowering, or controlling. I do not think it is extreme to compare the ramifications of how popular software evolves to very bad situations, given that powerful things can be used for good or evil.
RMS is just another communist who believes people like Linus are obligated to sacrifice their work for the masses, rather than having a choice. Linus had a choice and decided to freely share his information. This is based on modern philosophy since perverted by RMS. By your arguement BSD shouldn't exist. It's not under GPL.
Communists (the practical as opposed to the philosophical kind) practice pursuasion from the end of a gun. RMS does no such thing, and thus does give others the choice to agree or disagree with him.
My complaint is not that Linus doesn't kowtow to RMS, in fact I disagree with many of RMS' views. My complaint is that Linus ducks the hard questions.
It follows the letter of the law, but tramples all over the spirit.
Sounds like a good hack of the legal system to me (though your example would constitute trespassing), and fitting for a hacker.
I do suppose, though, that when (not if) Linus and RMS get thrown in jail for "manufacturing tools that support terrorism", Linus, though far less political, will have far greater support for his release. [... they always jail the intellectuals...]
A search on "Nazi" AND "Joseph Mengele" yields, among other things, this lovely snippet:
A prominent Nazi scientist such as von Verschuer certainly had first-hand knowledge of the Final Solution policy that had recently been formalized in Berlin by the top members of the Nazi hierarchy. He would have also correspondingly been aware of Nazi plans to construct enormous concentration camps across Europe, and that such camps held untold opportunities for in vivo experiments, living genetic research to be conducted on human subjects. Within a year after being posted to Berlin, Dr. Josef Mengele received a new assignment. In May of 1943, Mengele departed from Berlin for his next assignment: the Nazi concentration camp at Auschwitz, Poland.
Until RMS starts deriding Linus with the business end of a firearm or other form of force, he's free to air his views as he wishes.
You, of course, are free to ignore him, and dislike him for it.
Linus is in a position to lend support to RMS' views and thus it is to be expected that RMS will apply pressure to try to bring such support about. Linus, of course, can resist such pressures as he chooses, and it is somewhat to his credit that he does.
However, it is not clear that he does this because of ideological differences with RMS, or because he has no political views about anything.
Bluntly put, is he a Finn, or a Swiss national in disguise? (or worse, a Canadian wanna-be).
Re:Okaaaaay
on
Linus on DRM
·
· Score: 3, Insightful
Not that hard, if you actually put your brain to the task.:) There is a core set of behavior that is damaging to society as a whole, whether the behavior is practiced by politicians or criminals (the same class!) or whatever. Murder is in this class of behavior. This behavior, when let run wild (as in the favorite cite of Nazism) isn't just damaging to the world we live in, it can potentially destroy the whole world and all of human existence. Therefore, for pragmatic reasons, this behavior shouldn't be allowed by the forces that be, and if they practice this behavior, they should be thrown out and new powers installed.
That's a circular argument: Murder is illegal killing, and should therefore be considered "wrong".
Well,..., okaaaaaay.
The hard part is not arguing that murder is wrong: that's axiomatic. The hard part is determining when killing another is murder and when it isn't.
Now, even there, we've got a pretty good set of rules to guide us, but murder and killing, are extreme actions -- irreversable, given modern technology and medical skill (and, one could argue, by definition, since, if someone can be "brought back", then, perhaps, they weren't dead to begin with. Not really.).
The really hard questions are along the lines of "Should activity X be prohibited because it may lead to bad thing Y?".
May is not will, but increasing the mere liklihood of bad things is not acceptable, is it?
Think "Gun control" for a hard problem. (Philosophically, increased accountability for firearm use may render it a simple problem, but such rational approaches have not been socially accepted on a wide scale, so the problem remains "hard").
Re:Okaaaaay
on
Linus on DRM
·
· Score: 3, Insightful
OTOH, I don't think that Stallman should be trying to push his ideology on Linus any more than Linus should be trying to do so to Stallman. The difference is that Stallman tries to do exactly that with Linus, and Linus doesn't do so to Stallman.
I suppose that one shouldn't push the "ideology" that treating others worse than one would ever want to be treated one's self, either? Do we let others pursue their chosen "ideology" of murder?
If not, then why not, and what makes us "right" and others "wrong"? These are the (hard) questions to ask.
A very strong argument can be made that non-free software is so harmful that it ought to be outlawed. I don't particularly agree that that's the case, but neither do I see Stallman's "pushing" of his ideology as particularly forceful: he's not putting a gun to Linus' head to "comply", is he?
Stallman's "force" is nothing more than strength of conviction.
Torvalds' political apathy, on the other hand is irritating, in the same way that Canada's wishy-washy stance on the war in Iraq is irritating: trying to please all by doing nothing.
Guess what: driven people have strong opinions, and are willing to risk unpopularity (and often, much worse things), to stand by them. Such people make good friends and allies, though they do have enemies as well. When push comes to shove ("Linux is for terrorists"), will Linus be a "fair weather" friend to the Linux community, quietly disappearing into obscurity? I hope not, but, sadly, I'm not sure.
Re:Props to Linus
on
Linus on DRM
·
· Score: 5, Insightful
He simply shows that you don't have to be political (mp3 sharer, privacy guru, etc...) to enjoy Linux
You don't have to be political to enjoy a nice pair of new running shoes (made, possibly, with child labour), medical advances (made possible to some degree due to research done via unanesthesized vivisection of Jews by Nazis during WWII), or "free" health care (paid by tax dollars taken from those who now can't pay for their medical needs not covered by the "free" program).
You don't have to, certainly, but you should.
Politics, at its core, is the study of the philosophy of ethics. Ideally, the concern should be arriving at a means to determine whether actions are "right" or "wrong" without personal or group bias. Of course, "politics", as practiced, has nothing to do with ethics, and everything to do with special interests. And, no doubt, different people have different views of "right" and "wrong". It behooves them, to take part in the ethical debate.
With regard to Linux, particularly these days, that debate extends to whether it should be "permitted" to exist at all, supposedly being a "hackers'" and "terrorists'" tool. Surely, anyone who enjoys Linux should have an interest in the ethics surrounding it.
Now, should is not must, and people are free to live their lives in apolitical oblivion. However, the old mantra "evil prevails when good people do nothing", does nag at one's conscience, and such apathy in an individual is not a characteristic I particularly like to see.
In this regard, RMS is right to deride Linus Torvalds as merely an "engineer". Linus' pet operating system would not even exist, and have a strong ethical footing supporting the "goodness" of that existance, were it not for RMS' philosophical views. While this does not represent a "debt", per se., decent people generally respond to kindness (yes, the GPL is an act of kindness), by reciprocating.
Nonetheless, I do still think we'd be better off if people said, "I use C because it's what everyone else uses," rather than the misleading "I use it because it gets me closer to the bare metal" (or other common lines).
Point noted, but I meant "... when compared to Java". C still has a lot of horsepower on uniprocessors, without esoteric memory and interconnect performance hierarchies. It isn't popular without good reason, those those reasons, as you note, are starting to get dated. Though, it is probably true that, when the need for performance exceeds what one can craft in C, its left as an exercise to the hardware folks to make a supporting ASIC.
o perhaps it's better to use a language where the default easy thing to do is to use the checked version (say, Java)?
In some instances, yes.
However, in some instances, those checks can be unnecessary time killers, when you are trying to squeeze the last bit of performance from a program, and have done careful static analysis to ensure they don't happen, you don't need to check at run-time.
IOW, you need to be able to chose between the "luxery car" with the "smooth ride", and the Formula-1 racer.
The trouble with using "the right language for the job" is that even fewer programmers are skilled in multiple languages, and this tends toward the desire for a "general purpose" language. C, and to a better degree, C++, fit these bills, but you'd be right in pointing out that they provide a million different ways to shoot yourself in the foot (conversely, Java is poor if you're into real-time video deinterlacing and rendering).
Also, in many applications, there is a small need for tight and high-performce "as close to the metal as possible" code, combined with less-demanding code as support. Again, the complexity of a multi-language approach goes against what might be a use of "the right tool for the right job".
So, we script in bash, or perl, or python (cleverly embedding an interpreter where appropriate), and use general purpose languages when we need to perform at the bleeding edge. It's hard to shoehorn Java into that environment, except, perhaps, where platform independence, or traditional web-based interfaces (which imply platform independence) are required (Java's "write once, debug everywhere" reputation discounted for the benefit of it's advocates).
C++ offers most of the language features to provide "safe" wrappers around unsafe, or otherwise tricky code. Unfortunately, it does not provide a decent mechansism for the programmer to explicitly disallow certain language features in certain portions of code: use of smart pointers, for example, is hard to enforce when one can "cheat". In all fairness, C++ was never intended to forcefully protect the programmer from him or herself.
I think the answer likes in mutable languages (of which neither C++ nor Java is one), but those are among the worst for programmer skill, because they basically embrace the philosophy of: "Consider your problem. Consider an elegant language for expressing its solution. Code in that language, using this mutable language base." Very few people are good at coding that way (though, it is a sign of the best programmers).
If most developers are still using these "trivial" funcs, I'm scared what other funcs are just as buggy!
Funny how one can forget all about these "harmless" bad practises. Time to add it to the internal coding standard.:)
eliciting parent's informative reference on proper coding practices.
I wish it were that simple. Recently starting at a new position, I was horified at the number of strcpy()s, strcat()s, and sprintf()s peppered throughout code -- mostly building up a string to be send to a logging routine. Variable function parameter lists would have helped a lot here (yeah, they are ugly and not type-safe, but considering the interface being emulated, i.e. printf(), used them, it seamed natural), as well, building up a buffer in one place, instead of peppered throughout the code.
When I suggested (a) the variable parameter list technique and (b) careful use of vsnprintf(), I was strongly rebuked for the following reasons:
1. (a) is "too hard" for programmers to understand. Of course the lack of type-safety was used to further bolster this reason why it should be shunned (/me mutters something about not hiring idiots in the first place.
2. It is a waste of time to make code robust and secure. This even though the original author got bitten on the ass by a buffer overflow segfaulting his program and spending more time debugging that than it would take to use the simple bounded versions of the functions in question).
As long as the mean programmer skill is so mediocre, and one can "legalize away" the need for security and robustness, trivial holes like this will continue to proliferate. It is perceived to be cheaper to leave them than fix them.
While I don't agree with this observation, for trivial code with minimal maintenance commitments, and an army of lawyers already on retainer, it may hold true. Unfortunately.
Of course, when the legal defense breaks down, the programmer who didn't fix the problem lest he or she lose his or her job for insubordination, will be used as the corporate scapegoat.
Slashdot Mirror
It's a free world. Oh, wait...
Seriously though: sure, whatever jiggles your bits.
You do bring up an interesting point, however: NetBSD is, of course, generally considered more than just the kernel, the same way that "Red Hat Linux" is more than just the Linux kernel.
Using my nomenclature, NetBSD/NetBSD would be awkward, and, if OS vs. kernel were clear from context, NetBSD alone would suffice for either one. NetBSD/Linux, of course, would refer to the "standard" parts of NetBSD on a Linux kernel.
But, I have to ask, could you really get away from GNU completely and still have a relatively complete O/S if you did that?
Why not, indeed.
First, why reinvent the wheel?
Second, there may be good reasons to. Gentoo, for example, likely fills an interesting need: auditable, tuned distribution generation, for one. But, unless your personal distro is likely to have a wider audience, or your need is strong, why bother?
I do think the <packager> <raw source package set>/<kernel> naming convention is a good one, though I admit that <packager> and <raw source package set> might be redundant, together. I do see <raw source package set> evolving into a place holder for a standard for same, thus, "Red Hat GNU/Linux" would adhere to the same common GNU standard that "Debian GNU/Linux" and "Foo GNU/Linux" would. That is not the state of affairs today.
Furthermore, reducing the GNU moniker to a standards certification might not make RMS all that happy, unless the standard is one of all software being free. Unfortunately, that is not the purpose of the standardization in this context: one of features as opposed to freedom. It would also require the moniker to be trademarked, and used under license, presumably from the FSF, to have any teeth. Perhaps "GSB" (GNU Standard Base) would be a better monker in that case, leaving GNU free to be used by anyone chosing to express software freedom ideals.
As for the mayhem that would result from a plethora of distros using such a naming convention, I think such a fear is overblown: only a few would be popular, existing trademark law could be used to stem confusion, at least for commercial distros. Besides, the Sun hack of using.a.domain.name.as.part.of.a.distro.name.compo nent.mydomain.org could always be used, as ugly as that would get (hmm, the logical equivalent of CNAMES in product branding? /me smells a Bezos-style patent).
I had addressed another aspect of the kind of structured thought that has to go into those efforts, and that is portability, which also takes time and skill.
However, then, it wouldn't be "Debian/NetBSD" but rather "Debian GNU/NetBSD" (Debian, at least, having already accepted the GNU/Linux moniker).
I really think there are practical advantages to distinguishing between 1) kernel, 2) "everything else", i.e. portable packages ported to run on that kernel, and with each other, and 3) who did the porting/distribution bundling.
Of course, when there is only one varient of one of those three components, i.e. the bits and pieces of what would be GNU, we tend to leave it out, as cumbersome, redundant, and unwieldly, RMS protests of the need to pay hommage (or at least use terminology that conveys GNU roots) notwithstanding. Thus, "Red Hat Linux": there is no other kind of "Linux" packaged by Red Hat (and since they did the bundling, they get to call it pretty much whatever they want (personally, I would have preferred "Up Yours, MS"/OS, but anyway...)), and no techical need for a "Red Hat GNU/Linux" designation.
RMS may want to see us embrace a phisosophical basis for free software, but, without the economic benefits that open source exemplifies, I doubt free software would have the contributions it does. Similarly, without a technical argument for "GNU/Linux", the moniker will likely not be popular. It would be nice, though, if the technical argument were there, so the philosophy and history could get some recognition and representation in common use.
I expect that might come in niche markets: where Linux is combined with severly scaled down portions of GNU, and non-GNU software, particularly in the embedded market: look at BusyBox -- a combination of utilities in a single executible for space reasons. We are seeing attempts to standardize "Linux" for the desktop (LSB), as well as for the embedded space. I, for one, wouldn't mind seeing LSB (Linux Standard Base) become GSB (GNU Standard Base), with LSB dealing solely with a kernel standard, in this vein.
O.K. ObGNU/Linux rant over.
Yes, it is (speaking from the relative omfort of familiarity with RH 5.2 to 7.3). Alternatives are good.
However I have never understood why they are so far behind other distributions?
Perhaps, because they take the time to ensure that we can see Debian running on everything known to man.
Alternately, Xantech makes RS232 to IR converters which are designed to translate a character code sent over an RS232 line (i.e. your computer's serial port) to an IR signal which can be broadcast. However, the unit is quite large, and definitely requires an external power supply.
Since you apparently didn't even bother to do a rudimentary Google search for "RS232 to IR", or you would know this already, doing that is left as an excercize.
How this is perverted to be perceived as promulgation of a particular personal agenda is beyond me.
In this case, I encourage taking a position, one way or the other, and defending it. Nowehere do I argue that Linus is "wrong", bur rather wishy-washy.
No, of course not. I'm simply using extreme examples to illustrate that there are ethical issues worth considering. Because of the disruptive effects of open/free software, it should be examined from an ethical context.
For the record, if you have state funded healthcare, the rich, not the poor, pay for it disproportionately. So you're never going to be in a position where your taxes take enough money to leave you vulnerable to 'medical needs not covered by the "free" program'
Funny. That's exactly how my father died: he paid so much in taxes he could not afford essential surgery (about US$25,000 to US$50,000 with a 30% mortality rate -- his situation was grave), nor even the insurance that would cover it. Canada, of course, had the excuse that there were no qualified Canadian doctors to perform it, so, "Sorry, you die.".
I'm told they do send some people to the U.S. for care, now, but many still die that could be saved if they had bought real insurance and not the government bogus crap that was shoved down their throats as part of their taxes.
Avoid state healthcare like the plague!
Honestly, I think so too.
But, stranger things have happened, and if one is going to stick one's neck out, one may as well stick it out all the way. Though, that is one prediction I wouldn't be happy to "I told ya so!" about if it came to pass.
Watch that you don't confuse GNU with Linux as Stallman is ever quick to point out
I mean the community of users and developers around the Linux kernel.
However, GNU/Linux is an empowering operating system, that is disruptive to existing closed source interests, and can, no doubt (as can many things), be used to facilitiate hosting applications (particularly distributed, networked applications), that are favorable to "terrorism".
Linux is the heart of GNU/Linux, and Linus is the heart of Linux. Is it so far-fetched to think that the U.S. might wish to make an "example" out of him as "supporting terrorism", given the Constitution-shreading they've done since 9/11/2001?
If that happen to RMS, he'd not go down without kicking, screaming, generally making an annoying, attention-getting fuss, and making us think, "ya know, he's annoying, but doing that to him is wrong".
I can see Linus trying to "pull off a Ghandi" in the same predicament, in the best of circumstances, but, he doesn't come off even as confrontational enough for that, and would probably silently "let go" of Linux if the state guns were at his head (or, more likely, the heads of his wife and kids).
Yes, I am a cynic, and somewhat paranoid. Obviously, I hope that such a scenario does not unfold. But, I'd feel a lot more comfortable if Linus showed a bit more spine at times rather than constantly letting the water roll off his back, like a duck. Ducks get shot in hunting season.
How so?
First, ethics and law often diverge.
While I recognize the notions of copyright, trademark, and patent, I also note that the copyright and patent systems in the U.S. are highly broken. Thus, as far as the U.S. Constitution is concerned (and admitedly, it is but one legal framework), certain forms of *piracy* may very well be legal. I know I am a strong defender of fair use, and object to excessive copyright terms. Whether that makes me a law-breaker because I (a) back up my CDs and DVDs, and (b) use DeCSS to facilitate distribution of movies I have licensed, depends on how one interprets the law, and when -- perhaps the DMCA will, one day, be thrown out as unconsititutional. If that happens, present illegal violations of it would, legally, never happen (one of the rare cases where a law can be erased retroactively). But, this does not change the ethical notion of "not stealing", just because the question of what is property arises, and how, other than by contract, is property recognized (hint: state, i.e. mob, fiat).
For the record, I do not "pirate" movies or music, and have paid for every videocassette, DVD, and CD in my collection.
However, the ethics of ownership of thoughts and ideas are not clear: if "piracy" is so rampant, should it not be O.K.? If you want to keep something secret, well, ... keep it secret. In the absence of explicit contract that certain actions are a breech, ethically, those actions (to a contractarian, anyway), are neither moral, nor immoral, they are amoral.
Copyright and patents, are legal fabrications that make non-tangible ideas ownable, presumably to stimulate their creation and reward. A tricky balance between innovation and exploitation is necessary, and the present state of this balance drives, among other things, "pirating" behaviour.
So, no, I do not encourage, nor do I practice, what would be considered (by most -- the law might disagree on technical details, vis. use of DeCSS while in the U.S.A.) "piracy". However, I note that "piracy" is a highly disruptive social and economic force, and therefore software that can support it is, by extention, powerful, as a social mover. Powerful things should be examined in their ethical context.
On the other side of the coin, the whole internet can be viewed as leveraging very powerful software to effect societal change. While generally considered a constructive application of software, it too has a "dark side" that deserves ethical review (spam, porn, particularly exploitive (i.e. kiddie) porn).
My main objection to Linus is that he comes off as a little to blase with regard to potentially ethical aspects of Linux. I am not suggesting that he take a binary yes/no stance with regard to things like "DRM support in the kernel" (whatever that might mean), but rather express his views about DRM with a bit more than a virtual shrug.
However, like a firearm, software can be used for constructive as well as disruptive purposes, and to a surprising degree (though, the disruptive ones tend to be the ones making the news).
Thus, software raises significant ethical issues.
I think Linus shrugs them off a little too lightly. It isn't so much a DRM vs. "DRM verbotten" issue. It's more along the lines of "If DRM matters, what kinds of DRM would we think are O.K., and what aren't? And, are there things that we can do to encourage or facilitate the 'O.K.' kinds?". Linus appears to take the "not my problem" stance.
Would we consider support for, say, signed or encrypted modules in the kernel as a good thing? Perhaps, espescially if it could be excluded, so individuals could chose. Now, that's so defeatable as to not be strong DRM, but could still be leveraged in embedded applications in a crippling fashion -- remember most users don't build custom kernels, and what's easy for us isn't for the masses.
No, it's not. Get over yourself.
1. DeCSS
2. Napster
3. P2P networks
I'd hardly consider those non-disruptive (to some very established interests) and take exception to your view: software can be very powerful indeed.
Actually, it's the other way around: many citizens here are strong supporters of U.S. action, Canadians have served on the front lines (two have died, actually), yet the government is "not involved".
Fortunately, you don't, ... and what was that about proportion?
This is software we're talking about, for chrissakes! Not sweatshop labor, not Nazi death camps, not... whatever that third thing you said was supposed to mean.
Ethical issues surround us. Some of them are nasty, hence the extreme examples I provide. Ergo, one should have an interest in ethics. Lacking an interest even in the ethics surrounding one's own creation suggests an apathy of a degree that even extreme issues don't matter.
Furthermore, history, and recent events in the U.S., suggest that the difference between extreme measures, and apparently innocent things like software is not as large as one might like, making taking an ethical stance even nore compelling (one way or the other).
(You seem to be saying that doctors shouldn't be paid for their services.
No, I am saying that forcefully taking money from one to pay doctors that serve another, to the point where the individuals so impoverished can no longer purchase the medical care they need, is wrong.
When you talk about software in the same breath as Mengele, you're demonstrating your lack of a sense of proportion.
If you check, I didn't, but simply used the example of Mengele as substantiation of an earlier claim.
Software is becoming an extremely powerful force on this planet: either empowering, or controlling. I do not think it is extreme to compare the ramifications of how popular software evolves to very bad situations, given that powerful things can be used for good or evil.
Communists (the practical as opposed to the philosophical kind) practice pursuasion from the end of a gun. RMS does no such thing, and thus does give others the choice to agree or disagree with him.
My complaint is not that Linus doesn't kowtow to RMS, in fact I disagree with many of RMS' views. My complaint is that Linus ducks the hard questions.
Sounds like a good hack of the legal system to me (though your example would constitute trespassing), and fitting for a hacker.
I do suppose, though, that when (not if) Linus and RMS get thrown in jail for "manufacturing tools that support terrorism", Linus, though far less political, will have far greater support for his release. [... they always jail the intellectuals ...]
A prominent Nazi scientist such as von Verschuer certainly had first-hand knowledge of the Final Solution policy that had recently been formalized in Berlin by the top members of the Nazi hierarchy. He would have also correspondingly been aware of Nazi plans to construct enormous concentration camps across Europe, and that such camps held untold opportunities for in vivo experiments, living genetic research to be conducted on human subjects. Within a year after being posted to Berlin, Dr. Josef Mengele received a new assignment. In May of 1943, Mengele departed from Berlin for his next assignment: the Nazi concentration camp at Auschwitz, Poland.
--http://www.crimelibrary.com/serial_killers/histo ry/mengele/nazi_3.html?sect=6
That just hints at what went on, of course, but you can do the rest of the research for yourself.
You, of course, are free to ignore him, and dislike him for it.
Linus is in a position to lend support to RMS' views and thus it is to be expected that RMS will apply pressure to try to bring such support about. Linus, of course, can resist such pressures as he chooses, and it is somewhat to his credit that he does.
However, it is not clear that he does this because of ideological differences with RMS, or because he has no political views about anything.
Bluntly put, is he a Finn, or a Swiss national in disguise? (or worse, a Canadian wanna-be).
That's a circular argument: Murder is illegal killing, and should therefore be considered "wrong".
Well, ..., okaaaaaay.
The hard part is not arguing that murder is wrong: that's axiomatic. The hard part is determining when killing another is murder and when it isn't.
Now, even there, we've got a pretty good set of rules to guide us, but murder and killing, are extreme actions -- irreversable, given modern technology and medical skill (and, one could argue, by definition, since, if someone can be "brought back", then, perhaps, they weren't dead to begin with. Not really.).
The really hard questions are along the lines of "Should activity X be prohibited because it may lead to bad thing Y?".
May is not will, but increasing the mere liklihood of bad things is not acceptable, is it?
Think "Gun control" for a hard problem. (Philosophically, increased accountability for firearm use may render it a simple problem, but such rational approaches have not been socially accepted on a wide scale, so the problem remains "hard").
I suppose that one shouldn't push the "ideology" that treating others worse than one would ever want to be treated one's self, either? Do we let others pursue their chosen "ideology" of murder? If not, then why not, and what makes us "right" and others "wrong"? These are the (hard) questions to ask.
A very strong argument can be made that non-free software is so harmful that it ought to be outlawed. I don't particularly agree that that's the case, but neither do I see Stallman's "pushing" of his ideology as particularly forceful: he's not putting a gun to Linus' head to "comply", is he?
Stallman's "force" is nothing more than strength of conviction.
Torvalds' political apathy, on the other hand is irritating, in the same way that Canada's wishy-washy stance on the war in Iraq is irritating: trying to please all by doing nothing.
Guess what: driven people have strong opinions, and are willing to risk unpopularity (and often, much worse things), to stand by them. Such people make good friends and allies, though they do have enemies as well. When push comes to shove ("Linux is for terrorists"), will Linus be a "fair weather" friend to the Linux community, quietly disappearing into obscurity? I hope not, but, sadly, I'm not sure.
You don't have to be political to enjoy a nice pair of new running shoes (made, possibly, with child labour), medical advances (made possible to some degree due to research done via unanesthesized vivisection of Jews by Nazis during WWII), or "free" health care (paid by tax dollars taken from those who now can't pay for their medical needs not covered by the "free" program).
You don't have to, certainly, but you should.
Politics, at its core, is the study of the philosophy of ethics. Ideally, the concern should be arriving at a means to determine whether actions are "right" or "wrong" without personal or group bias. Of course, "politics", as practiced, has nothing to do with ethics, and everything to do with special interests. And, no doubt, different people have different views of "right" and "wrong". It behooves them, to take part in the ethical debate.
With regard to Linux, particularly these days, that debate extends to whether it should be "permitted" to exist at all, supposedly being a "hackers'" and "terrorists'" tool. Surely, anyone who enjoys Linux should have an interest in the ethics surrounding it.
Now, should is not must, and people are free to live their lives in apolitical oblivion. However, the old mantra "evil prevails when good people do nothing", does nag at one's conscience, and such apathy in an individual is not a characteristic I particularly like to see.
In this regard, RMS is right to deride Linus Torvalds as merely an "engineer". Linus' pet operating system would not even exist, and have a strong ethical footing supporting the "goodness" of that existance, were it not for RMS' philosophical views. While this does not represent a "debt", per se., decent people generally respond to kindness (yes, the GPL is an act of kindness), by reciprocating.
Point noted, but I meant "... when compared to Java". C still has a lot of horsepower on uniprocessors, without esoteric memory and interconnect performance hierarchies. It isn't popular without good reason, those those reasons, as you note, are starting to get dated. Though, it is probably true that, when the need for performance exceeds what one can craft in C, its left as an exercise to the hardware folks to make a supporting ASIC.
o perhaps it's better to use a language where the default easy thing to do is to use the checked version (say, Java)?
In some instances, yes.
However, in some instances, those checks can be unnecessary time killers, when you are trying to squeeze the last bit of performance from a program, and have done careful static analysis to ensure they don't happen, you don't need to check at run-time.
IOW, you need to be able to chose between the "luxery car" with the "smooth ride", and the Formula-1 racer.
The trouble with using "the right language for the job" is that even fewer programmers are skilled in multiple languages, and this tends toward the desire for a "general purpose" language. C, and to a better degree, C++, fit these bills, but you'd be right in pointing out that they provide a million different ways to shoot yourself in the foot (conversely, Java is poor if you're into real-time video deinterlacing and rendering).
Also, in many applications, there is a small need for tight and high-performce "as close to the metal as possible" code, combined with less-demanding code as support. Again, the complexity of a multi-language approach goes against what might be a use of "the right tool for the right job".
So, we script in bash, or perl, or python (cleverly embedding an interpreter where appropriate), and use general purpose languages when we need to perform at the bleeding edge. It's hard to shoehorn Java into that environment, except, perhaps, where platform independence, or traditional web-based interfaces (which imply platform independence) are required (Java's "write once, debug everywhere" reputation discounted for the benefit of it's advocates).
C++ offers most of the language features to provide "safe" wrappers around unsafe, or otherwise tricky code. Unfortunately, it does not provide a decent mechansism for the programmer to explicitly disallow certain language features in certain portions of code: use of smart pointers, for example, is hard to enforce when one can "cheat". In all fairness, C++ was never intended to forcefully protect the programmer from him or herself.
I think the answer likes in mutable languages (of which neither C++ nor Java is one), but those are among the worst for programmer skill, because they basically embrace the philosophy of: "Consider your problem. Consider an elegant language for expressing its solution. Code in that language, using this mutable language base." Very few people are good at coding that way (though, it is a sign of the best programmers).
If most developers are still using these "trivial" funcs, I'm scared what other funcs are just as buggy!
Funny how one can forget all about these "harmless" bad practises. Time to add it to the internal coding standard. :)
eliciting parent's informative reference on proper coding practices.
I wish it were that simple. Recently starting at a new position, I was horified at the number of strcpy()s, strcat()s, and sprintf()s peppered throughout code -- mostly building up a string to be send to a logging routine. Variable function parameter lists would have helped a lot here (yeah, they are ugly and not type-safe, but considering the interface being emulated, i.e. printf(), used them, it seamed natural), as well, building up a buffer in one place, instead of peppered throughout the code.
When I suggested (a) the variable parameter list technique and (b) careful use of vsnprintf(), I was strongly rebuked for the following reasons:
1. (a) is "too hard" for programmers to understand. Of course the lack of type-safety was used to further bolster this reason why it should be shunned (/me mutters something about not hiring idiots in the first place.
2. It is a waste of time to make code robust and secure. This even though the original author got bitten on the ass by a buffer overflow segfaulting his program and spending more time debugging that than it would take to use the simple bounded versions of the functions in question).
As long as the mean programmer skill is so mediocre, and one can "legalize away" the need for security and robustness, trivial holes like this will continue to proliferate. It is perceived to be cheaper to leave them than fix them.
While I don't agree with this observation, for trivial code with minimal maintenance commitments, and an army of lawyers already on retainer, it may hold true. Unfortunately.
Of course, when the legal defense breaks down, the programmer who didn't fix the problem lest he or she lose his or her job for insubordination, will be used as the corporate scapegoat.