"Rail travel at high speeds is not possible because passengers, unable to breathe, would die of asphyxia."
Dionysius Lardner, Professor of Natural Philosophy and Astronomy at University College, London; 1830
He authored early works on the steam engine, amongst other things, and many people then would have though he knew his stuff. See the ususal http://en.wikipedia.org/wiki/Dionysius_Lardner for more details.
Bear in mind that in those days, "high speed" would have been taken to be "faster than a horse" (say 30mph) - as that was the fastest transport available then.
I can't find a reference to RMS correcting people _away_ from saying GNU/Linux for (non-gnu) Linux systems - can you provide one ?
I am, however, aware of examples of him assuming all Linux based systems are GNU/Linux, including the original change (to lignux) to configure files. I am a believer in looking at the code to see what a programmer means (not the docs), and when RMS made the change from outputting "linux" to "lignux" (and later "linux-gnu") that was literally the change that was made. All Linux based systems, as detected solely by "uname -s", became GNU/Linux.
He could have said (psuedocode):
if (kernel is linux)
then
if (traditional UNIX userspace is primarily composed of GNU software)
then
system is "{cpu name}-linux-gnu"
else
system is "{cpu name}-linux-unknown"
which is essentially what you state his position as. In fact, he didn't, he actually said:
if (kernel is linux)
then
system is "{cpu name}-linux-gnu"
Which is what I don't think is reasonable.
If you want some other quotes:
When SIGLINUX invited me to speak, it was a "Linux User Group"; that is, a group for users of the GNU/Linux system which calls the whole system "Linux".
No possibility of non-gnu linux based systems considered there (possibly linux user groups exclude such users / systems, but I don't think so).
Or from RMS's own FAQ on the issue:
Can't there be Linux systems without GNU? It is possible to make a system that uses Linux as the kernel but is not based on GNU. I'm told there are small systems, used for embedded development, that include Linux but not the GNU system. IBM was once rumored to be planning to put Linux as the kernel into AIX; whether or not they actually tried to do this, it is theoretically possible.
"It is possible". "It is theoretically possible". "I'm told there are".
Yes he was told, right at the time of "lignux", and even though it's now over a decade later, and this is a _F_AQ, he still doesn't appear to _know_ whether such systems actually exist. He won't even dignify the question with the small amount of research it would take to provide a definitive answer.
Does he think all Linux systems are GNU/Linux ? That I can't say, however, he gives that impression and has failed to make a clear statement of the opposite when he has the opportunity (see above).
As to planes, he possibly hasn't yet realised that "fly by wire" means software, not pulling on real wires. If he ever finds out that FBW (and similar) software developement can require independent implementations, developed by separaate teams who are not allowed to look at each other's code, then I guess he won't fly.
Rugged portables from a couple of decades ago (or maybe a bit more than that).
Built to be dropped in the mud by soldiers, then driven over (by a tank if possible), then to have the mud washed (yes, washed) off afterwards - through all of which your data capture would probably still be running. Dust proof, waterproof, shockproof and practically bomb proof.
Stolen password would be a vector into Apache servers on other platforms as well - but the reports so far seem to be saying it is just Apache + Linux servers.
If that _is_ the case, then it is more likely to be a weakness specific to that stack (possibly only certain version combinations) than a password compromise. The weakness may also be in something else installed on that stack, and thus applicable to only a small fraction of the 75M apache servers (eg. LAMP + particular PHP extension / application).
a) no one knows the initial attack vector (on the _servers_) b) the malware (on the _servers_) seems to be difficult to detect c) and no one seems to know how to remove it successfully either - some have suggested it is a rootkit as the apache part of the infection seems to reappear when removed. d) possibly as a result of (b), estimates are vague on number of infected servers, but I've seen estimates from "hundreds" to "tens of thousands" e) seems to be Linux + Apache stack that is targeted
> When was the last time anyone here interviewed the company they ended up working for
Every ****ing time.
And the ones I didn't end up working for.
And I've had to "sell" the company many times when interviewing (in fact, if I don't have to... it's probably not an interview that will end with an offer).
Sometimes you write to your target audience. In this case the submitter was writing about recruitment, for people involved in recruitment, and therefore assumed the reader would know what CV meant (in that context). As would I.
Oh, and note that I _am_ aware that "résumé" is used interchangeably for CV in some places. That is the point - I would expect anyone involved in technical or professional recruitment to know both terms.
[Aside: did "résumé" get substituted at all in the USA around the time of "freedom fries" or did the great leaders not know that the word is French ?]
> exactly, the GPL has never been about protecting the Developer or the Developer's freedoms, but protecting the User and insuring the code is always available.
GPLv2 ensures the code is available. The TIVO code is available.
The user doesn't get protected by the GPL (even v3) - only the recipient of the distribution (which may or may not be the user). In GPLv3 this is made explicit in conveyance vs. propagation - if the software is conveyed to a user, they get rights, if it is propagated, they do not.
For instance, it is made very clear by the FSF that, for example, the user (ie. the voter) of a voting machine with GPLv3 software would _not_ aquire modification rights.
Similarly, the user of a PVR running GPLv3 software would not get modification rights if that PVR was rented. Since the GPLv3 protects the user, and the users will want/need to modify their PVRs (that is the rationale for the anti-tivo clause), rented PVRs (eg. as part of a TV package) will clearly never take off and achieve significant market penetration...
> Linus did a huge mistake, and that doesn't relate to GPLv3: he locked the kernel into a license that CANNOT be changed.
It _can_ be changed, but only by consent of all the copyright holders. That (as I have previously read explained publically by Linus - wish I could find the reference now...) is a _feature_ not a bug.
Linus deliberately did it this way so that not even he could single handedly change the licence. No one person or entity can. This is a strength - there is no single point of failure, no single target to corrupt.
It is also a potential weakness if the GPLv2 turns out to be ineffective at doing what Linus wanted it to do - since it would then be difficult to fix.
Overall, only time will tell if Linus' choice was a mistake, but so far, after about a decade and a half, Linux and GPLv2 are still looking pretty successful.
> > The only thing you CAN'T do is run it on a specific piece of hardware.
> You cant do it on probably the only piece of hardware it runs on. Whats the point of a piece of software, if you cant run it on the hardware it is intended to run on?
It has many uses (that I can think of).
Suppose I ship a device containing software in ROM. You can't run modified software on that device, so what happens if that software is GPLed (2 or 3) ?
Is it:
a) I can't ship GPL software in ROM (at all) b) I don't have to ship you the source, because after all there would be "no point" in you having it or c) I still have to ship the source
If it is (c) then, by your argument, the GPL is pointless ?
> It started with a printer driver RMS wanted to customize but was denied the source code. [...] the code would be pretty much useless
Years ago I also had a problem with a printer driver - postcript interpreter, in ROM. Source code for that interpreter would have been very useful in diagnosing and fixing the problem. Even though I couldn't change the ROM.
If you can't see the use, then you aren't thinking hard enough.
> The freedom to "change and run" was one of the basic four GPL freedoms from the very beginnings, and is the same in relation to "run", as is "change and distribute" in relation to "distribute".
The publication history doesn't match this contention. The statement of freedoms postdates the first versions of the GPL, and the earliest statement of those freedoms contain only three, not four. Guess which one was added later ?
"...estimated ping time 1.25hrs, this could take a while.
It looks like you're frustrated by high ping times, would you like to be put into contact with someone who has had success with this? Confirmed. Now connecting...... it looks like you're getting frustrated by lack of oxygen, would you like to... "
> what's wrong with the pacific ocean? Surely if you've got a fire, ditching in water is perfect! Do I get a job in airline management now?
No, but you could maybe get a job in chemistry - just take a small piece of lithium, set fire to it and then throw it in a big bucket of water to put that fire out... should be "perfect", right ?
PS: if the penny hasn't dropped yet: "DON'T try this at home kids".
At breakfast, they're already powering up the combination library/videocam/audio recorder/music maker/drawing kits. At night, they're dozing off in front of them
My kids would (did once) do this too, on an ordinary PC - doesn't need any particular special kit. They don't get to do it now because they run into time limits etc. I've set those limits because I happen to think this behaviour is _not_ healthy or desirable (and IME leads to really tired, foul-tempered kids).
These OLPCs are most likely going into homes where the parents are not computer literate and will probably never learn as much about them as their kids. They won't be as able to restrict their kids and prevent them spending all their time on the computer instead of, say, going outside and kicking a ball.
Maybe this social experiment will turn out to be good for the kids - but from that quote above and my experience with kids, I'm skeptical.
Nowhere did GP say they had taught this behaviour.
My kids have picked up use of a PC with little or no instruction - interested kids have a way of just figuring stuff out by trying things (something we lose the tendency to do as we head into adulthood).
I have had to put controls in place to _stop_ them signing up for chatrooms, ebay, gambling sites etc. (because I found them doing it - way before I thought I would have to worry about it). The oldest is _6_. At no time have they been _taught_ how to do these things - they just click on stuff (including ads) and figure it out.
I have an 18month old who has found the main switch (at the power socket) that controls the home computer. He takes great delight in turning it off. Was he _taught_ to do that ? Hell no. He's figured it out by randomly playing with switches at some point and ealising that that one turns the screen off and gets a big reaction from the person using the computer.
I would expect a young child to pick up using any type of pc (XO or otherwise) pretty quickly - probably quicker than me these days. There is nothing special about OLPC in this respect, or indeed about any PC. Decades ago, thousands of children had similar experiences with the early home computers like the ZX81 etc. Most often, those computers went into homes where the parents had zero computing knowledge (we're talking 1981 remember), so there was no one to teach the children. The children still learnt, and fast, and with only a command line interface.
The biggest argument, for me, against this sort of disclosure is that the public in general is unfortunately (proven) too stupid to be entrusted with the information.
If soldiers, they would be POWs and under Geneva conventions.
If not, they are allegedly civilian criminals and should be prosecuted in the civilian judicial system.
Problem with Gitmo is the US has decided these people are neither soldiers nor civilians but fall in some black hole category in between, where they have no access to civilian justice and no POW rights either.
That's all fine (even agreed), but it isn't related to the question of whether or not "The original GPL code is still there" (as it had been "stolen").
You claimed it isn't, I disagreed. Looking for the file in your copy of the kernel source is left as an excercise for the reader...
Yes it is - you missed the word "original". Only the modified source code is allegedly not available. The _original_ GPL code will is still available where it was before, therefore it hasn't been "stolen".
All of which misunderstanding demonstrates why "theft" and similar words are not appropriate in the context of copyright violation.
The design has inherent susceptibility to going into VRS (vortex ring state) asymetrically - ie. one rotor stops lifting. This is bad enough news in an normal single rotor helicopter but in a side-by-side rotor design it's seriously bad news (uncommanded, uncontrollable roll). Also, it seems that the V22 can go into asymmetric VRS very quickly with little or no warning to the pilot (much less warning than a normal helicopter).
This is a fundamental design issue that isn't going to get worked out in the field.
Slashdot Mirror
The quote is rather (in)famous.
"Rail travel at high speeds is not possible because passengers, unable to breathe, would die of asphyxia."
Dionysius Lardner, Professor of Natural Philosophy and Astronomy at University College, London; 1830
He authored early works on the steam engine, amongst other things, and many people then would have though he knew his stuff. See the ususal http://en.wikipedia.org/wiki/Dionysius_Lardner for more details.
Bear in mind that in those days, "high speed" would have been taken to be "faster than a horse" (say 30mph) - as that was the fastest transport available then.
I am, however, aware of examples of him assuming all Linux based systems are GNU/Linux, including the original change (to lignux) to configure files. I am a believer in looking at the code to see what a programmer means (not the docs), and when RMS made the change from outputting "linux" to "lignux" (and later "linux-gnu") that was literally the change that was made. All Linux based systems, as detected solely by "uname -s", became GNU/Linux.
He could have said (psuedocode):
if (kernel is linux)
then
if (traditional UNIX userspace is primarily composed of GNU software)
then
system is "{cpu name}-linux-gnu"
else
system is "{cpu name}-linux-unknown"
which is essentially what you state his position as. In fact, he didn't, he actually said:
if (kernel is linux)
then
system is "{cpu name}-linux-gnu"
Which is what I don't think is reasonable.
If you want some other quotes:
No possibility of non-gnu linux based systems considered there (possibly linux user groups exclude such users / systems, but I don't think so).
Or from RMS's own FAQ on the issue:
"It is possible". "It is theoretically possible". "I'm told there are".
Yes he was told, right at the time of "lignux", and even though it's now over a decade later, and this is a _F_AQ, he still doesn't appear to _know_ whether such systems actually exist. He won't even dignify the question with the small amount of research it would take to provide a definitive answer.
Does he think all Linux systems are GNU/Linux ? That I can't say, however, he gives that impression and has failed to make a clear statement of the opposite when he has the opportunity (see above).
I don't think RMS drives.
As to planes, he possibly hasn't yet realised that "fly by wire" means software, not pulling on real wires. If he ever finds out that FBW (and similar) software developement can require independent implementations, developed by separaate teams who are not allowed to look at each other's code, then I guess he won't fly.
But it runs on Linux - Uh, sorry - GNU/Linux.
Essentially, all linux systems are GNU/Linux to RMS (check out the source to configure).
Rugged portables from a couple of decades ago (or maybe a bit more than that).
Built to be dropped in the mud by soldiers, then driven over (by a tank if possible), then to have the mud washed (yes, washed) off afterwards - through all of which your data capture would probably still be running. Dust proof, waterproof, shockproof and practically bomb proof.
Most of the engineers I know would probably come up with a "remote detonator" or "timer" solution at least.
The terrorists do too. IEDs are a major problem in IRAQ and they are getting better, now including, allegedly, shaped charges know-how from Iran.
However, they also use the Mk1-dumb-homo-sapiens remote detonator / timer. Why ? Because it works.
[You don't think the master bomb-makers, ie. the engineers, actually blow themselves up rather than use mules, do you ?]
Stolen password would be a vector into Apache servers on other platforms as well - but the reports so far seem to be saying it is just Apache + Linux servers.
If that _is_ the case, then it is more likely to be a weakness specific to that stack (possibly only certain version combinations) than a password compromise. The weakness may also be in something else installed on that stack, and thus applicable to only a small fraction of the 75M apache servers (eg. LAMP + particular PHP extension / application).
That's the secondary infection.
The tough question is what is the malware that is infecting the servers themselves. There have been reports of this for weeks now, and apparently it may go back months (see eg. http://www.channelregister.co.uk/2008/01/16/mysterious_web_infection_continues/), and AFAICS:
a) no one knows the initial attack vector (on the _servers_)
b) the malware (on the _servers_) seems to be difficult to detect
c) and no one seems to know how to remove it successfully either - some have suggested it is a rootkit as the apache part of the infection seems to reappear when removed.
d) possibly as a result of (b), estimates are vague on number of infected servers, but I've seen estimates from "hundreds" to "tens of thousands"
e) seems to be Linux + Apache stack that is targeted
> When was the last time anyone here interviewed the company they ended up working for
Every ****ing time.
And the ones I didn't end up working for.
And I've had to "sell" the company many times when interviewing (in fact, if I don't have to... it's probably not an interview that will end with an offer).
Sometimes you write to your target audience. In this case the submitter was writing about recruitment, for people involved in recruitment, and therefore assumed the reader would know what CV meant (in that context). As would I.
Oh, and note that I _am_ aware that "résumé" is used interchangeably for CV in some places. That is the point - I would expect anyone involved in technical or professional recruitment to know both terms.
[Aside: did "résumé" get substituted at all in the USA around the time of "freedom fries" or did the great leaders not know that the word is French ?]
Mr Coward, looking at your posting history you have _way_ more than a couple of UIDs...
> exactly, the GPL has never been about protecting the Developer or the Developer's freedoms, but protecting the User and insuring the code is always available.
GPLv2 ensures the code is available. The TIVO code is available.
The user doesn't get protected by the GPL (even v3) - only the recipient of the distribution (which may or may not be the user). In GPLv3 this is made explicit in conveyance vs. propagation - if the software is conveyed to a user, they get rights, if it is propagated, they do not.
For instance, it is made very clear by the FSF that, for example, the user (ie. the voter) of a voting machine with GPLv3 software would _not_ aquire modification rights.
Similarly, the user of a PVR running GPLv3 software would not get modification rights if that PVR was rented. Since the GPLv3 protects the user, and the users will want/need to modify their PVRs (that is the rationale for the anti-tivo clause), rented PVRs (eg. as part of a TV package) will clearly never take off and achieve significant market penetration...
> Linus did a huge mistake, and that doesn't relate to GPLv3: he locked the kernel into a license that CANNOT be changed.
It _can_ be changed, but only by consent of all the copyright holders. That (as I have previously read explained publically by Linus - wish I could find the reference now...) is a _feature_ not a bug.
Linus deliberately did it this way so that not even he could single handedly change the licence. No one person or entity can. This is a strength - there is no single point of failure, no single target to corrupt.
It is also a potential weakness if the GPLv2 turns out to be ineffective at doing what Linus wanted it to do - since it would then be difficult to fix.
Overall, only time will tell if Linus' choice was a mistake, but so far, after about a decade and a half, Linux and GPLv2 are still looking pretty successful.
> > The only thing you CAN'T do is run it on a specific piece of hardware.
> You cant do it on probably the only piece of hardware it runs on. Whats the point of a piece of software, if you cant run it on the hardware it is intended to run on?
It has many uses (that I can think of).
Suppose I ship a device containing software in ROM. You can't run modified software on that device, so what happens if that software is GPLed (2 or 3) ?
Is it:
a) I can't ship GPL software in ROM (at all)
b) I don't have to ship you the source, because after all there would be "no point" in you having it
or
c) I still have to ship the source
If it is (c) then, by your argument, the GPL is pointless ?
> It started with a printer driver RMS wanted to customize but was denied the source code. [...] the code would be pretty much useless
Years ago I also had a problem with a printer driver - postcript interpreter, in ROM. Source code for that interpreter would have been very useful in diagnosing and fixing the problem. Even though I couldn't change the ROM.
If you can't see the use, then you aren't thinking hard enough.
> The freedom to "change and run" was one of the basic four GPL freedoms from the very beginnings, and is the same in relation to "run", as is "change and distribute" in relation to "distribute".
The publication history doesn't match this contention. The statement of freedoms postdates the first versions of the GPL, and the earliest statement of those freedoms contain only three, not four. Guess which one was added later ?
> Confirmed. Now connecting...
... it looks like you're getting frustrated by lack of oxygen, would you like to... "
"...estimated ping time 1.25hrs, this could take a while.
It looks like you're frustrated by high ping times, would you like to be put into contact with someone who has had success with this? Confirmed. Now connecting...
wait... you mean this _isn't_ work ?!
> what's wrong with the pacific ocean? Surely if you've got a fire, ditching in water is perfect! Do I get a job in airline management now?
No, but you could maybe get a job in chemistry - just take a small piece of lithium, set fire to it and then throw it in a big bucket of water to put that fire out... should be "perfect", right ?
PS: if the penny hasn't dropped yet: "DON'T try this at home kids".
Umm... you want to put out a _lithium_ fire using _water_ ?
My kids would (did once) do this too, on an ordinary PC - doesn't need any particular special kit. They don't get to do it now because they run into time limits etc. I've set those limits because I happen to think this behaviour is _not_ healthy or desirable (and IME leads to really tired, foul-tempered kids).
These OLPCs are most likely going into homes where the parents are not computer literate and will probably never learn as much about them as their kids. They won't be as able to restrict their kids and prevent them spending all their time on the computer instead of, say, going outside and kicking a ball.
Maybe this social experiment will turn out to be good for the kids - but from that quote above and my experience with kids, I'm skeptical.
Nowhere did GP say they had taught this behaviour.
My kids have picked up use of a PC with little or no instruction - interested kids have a way of just figuring stuff out by trying things (something we lose the tendency to do as we head into adulthood).
I have had to put controls in place to _stop_ them signing up for chatrooms, ebay, gambling sites etc. (because I found them doing it - way before I thought I would have to worry about it). The oldest is _6_. At no time have they been _taught_ how to do these things - they just click on stuff (including ads) and figure it out.
I have an 18month old who has found the main switch (at the power socket) that controls the home computer. He takes great delight in turning it off. Was he _taught_ to do that ? Hell no. He's figured it out by randomly playing with switches at some point and ealising that that one turns the screen off and gets a big reaction from the person using the computer.
I would expect a young child to pick up using any type of pc (XO or otherwise) pretty quickly - probably quicker than me these days. There is nothing special about OLPC in this respect, or indeed about any PC. Decades ago, thousands of children had similar experiences with the early home computers like the ZX81 etc. Most often, those computers went into homes where the parents had zero computing knowledge (we're talking 1981 remember), so there was no one to teach the children. The children still learnt, and fast, and with only a command line interface.
Oh, it's worse than that. We live in a world where people take vigilante action because they believe "child doctor" equals "baby rapist".
http://news.independent.co.uk/uk/this_britain/article271248.ece
The biggest argument, for me, against this sort of disclosure is that the public in general is unfortunately (proven) too stupid to be entrusted with the information.
If soldiers, they would be POWs and under Geneva conventions.
If not, they are allegedly civilian criminals and should be prosecuted in the civilian judicial system.
Problem with Gitmo is the US has decided these people are neither soldiers nor civilians but fall in some black hole category in between, where they have no access to civilian justice and no POW rights either.
That's all fine (even agreed), but it isn't related to the question of whether or not "The original GPL code is still there" (as it had been "stolen").
You claimed it isn't, I disagreed. Looking for the file in your copy of the kernel source is left as an excercise for the reader...
> No its not.
Yes it is - you missed the word "original". Only the modified source code is allegedly not available. The _original_ GPL code will is still available where it was before, therefore it hasn't been "stolen".
All of which misunderstanding demonstrates why "theft" and similar words are not appropriate in the context of copyright violation.
The design has inherent susceptibility to going into VRS (vortex ring state) asymetrically - ie. one rotor stops lifting. This is bad enough news in an normal single rotor helicopter but in a side-by-side rotor design it's seriously bad news (uncommanded, uncontrollable roll). Also, it seems that the V22 can go into asymmetric VRS very quickly with little or no warning to the pilot (much less warning than a normal helicopter).
This is a fundamental design issue that isn't going to get worked out in the field.