>I still think that int should be the easiest size for the machine to deal with.
Correct, but any 64-bit CPU that I know about about instructions for manipulating 32bit variables easily too. These "64bit" CPU manipulate 32bit value as easily as 64bit values, which means K&R rules that the int should be the "natural" int of the CPU doesn't tell you anything about wether an int should be 32 or 64 bit..
Still in some case, the start-up times of some proprietary things are much better than the equivalent on OpenSource: BeOS booted to the graphical desktop in ~10s! On "old" machine even.. That is the equivalent of booting the kernel, starting X and KDE, which can easily takes minutes on slow PC (or with little memory)..
Well I suppose that this beta code got out of Microsoft, otherwise the DR-DOS guys wouldn't have had any case. Usually beta version are used by 'selected customers' (big enterprises) to evaluate, so they are important!
>Do you have a timeline + evidence of Microsoft doing this to another technology?
They added error messages on Windows (3.1?) so that people would get scared when they would run Windows on another DOS than MS-DOS. There was some litigation (with Caldera I think), but Microsoft settled for some money change (for them).
I don't know how this got to be modded interesting but, here I go: >will the DNA auto-repair any damage to the wire?
Uh, DNA in itself is just a big molecule which isn't capable to do anything at all! A cell is capable to repair its own DNA, but obviously the DNA they use is not inside a cell, where they put the wire in place..
>Couldn't a virus (biological, not computer) be used to re-write the DNA strand that is used to construct the devices, to make different components for sinister purposes?
If a virus modified the DNA, the most probable outcome (by very far) is that you would get a malfunctionning device ie the transistor won't work correctly..
You're right, helicopters are porr flying cars: - they are noisy as hell - you have to be very skilled to "drive" one, and even more skilled to avoid being killed if there is a failure. - half of their time is spent on maintenance. Anyway heavy maintenance is needed for *any* flying cars, otherwise there would be lots of death caused by malfunctions. I suspect that we won't be able to make products "perfect enough" for flying cars before the "nanotech world".
I wish like anybody else that OLED would be here now, but it has been reported that it still have problems getting uniform colour over a large surface for a long period of time..
Be careful if your in the US and you want to do the 3-channel alpha thingy, if I remember correctly a company (maybe Apple?) has a patent on having a per-color alpha channel..
Yes, it seems stupid, but stupid patents are quite common..
I think that all these proposal to increase the amount of energy avoid a potential problem: the corresponding increase of heat generated..
Eventually all this energy will turn into heat, so it is quite possible that this will eventually raise earth's temperature.. I think that it may be wiser to increase the efficiency usage of energy than to increase the amount of energy used, well unless of course we need to warm up the earth..
>can we even make useful robots at normal scale that self-replicate?
Well to product a robot at a normal scale, you'd have to produce parts with very high precision: motors, electronic,etc.. So in fact, producing true self-replicating "normal bots" is almost as difficult as producing nanobots.
We could probably produce, self-replicating bots which use lego-like elements, but what would be the point?
And there is BIG difference between going higher than the speed of light and creating nanobots. The first one is violating a physical principle, the other is "recreating" partly what nature has already achieved with life..
>In France, 90% of our electricity comes from nuclear plants and it's reasonnably cheap.
True, but AFAIK: 1) right now waste have not been truly disposed of, just treated "temporarily" in the recycling plant of LaHague. This is a political "hot potatoe", and I'd like that a permanent solution be found. 2) No nuclear plant have been disassembled, so the cost of the disassembling the nuclear plant is only theoretical, if they save not enough money for that, the cost of electricity will go up..
And indenpendently from this, there is a problem with the lack of transparency of the nuclear industry in France: when the Tchernobyl clouds of radioactivity went over France, officials lied about the level of radioactivity measured.
I'm not against the nuclear industry, but all the reasons give above makes me "cautious" about it.
On one end you have Ariane5 exploding because of a software error, on the other end, you have 10 clients within your enterprise which loose time because of software's bugs.
With such huge range of differing costs for finding the bug before or after the shipping of your product, the "average cost" of bugs is meaningless.
I think that the only thing to remember is: - bugs found late cost more to fix than bugs found earlier (any specific number is invalid) - finding bugs early is difficult and can be expensive.
Of which you can deduce that: - if late bugs can cost you very much (Ariane5 for exemple), you want to spend a lot of money on software testing|review at each level. - otherwise if tests can cost more than the fix (a small number of internal users with a non-critical software), then maybe you can use the clients as testers, but it must be managed well (tell the users, be in close contact with the users, don't let them wait the fixes too much).
Mod the parent up, if memory serves, Microsoft have had up to three years of being late for some OS, with major feature post-poned. So with Linux you get no roadmap, with Microsoft you get very unreliable roadmap, I'm not sure Microsoft has a very big advantage here..
Well I think that you treat a bit lightly the complexity of the "issue unit", if you want to have the maximal single thread performance, a 'single issue multiple units' makes sense, but every "brainiac" CPU usually has some limits on the way it can reorder the instructions, otherwise it would become too complex.
Look at the PPC970, it is a deep and wide CPU: it has so many instructions in flight that it packs the instructions into bundle, and sometimes some slot can be unused because of the limitations of the issue unit.. A multi-core CPU simplifies the logic of the issue unit: the Power5 is going to be SMT and SMP..
You are right that now many applications are programmed with the single-threading model, but in a few years, even entry-level CPU will be SMT (hopefully with less bottlenecks than the current P4) or SMP, so multi-threaded applications will gain a *big* advantage..
>So many pirates were private military contractors,
These were called "corsaires" *not* pirates!
Confusing both is like calling a soldier a hitman, technically soldiers are payed murderers like hitmen, but I'm sure that hitmen would feel insulted if you said that they were just soldiers;-)
>Anyone buying an Athlon-64 primarily to run the latest incarnation of Doom or Office is simply an idiot.
For Office I agree, but for games, why not? Because you don't care for games, doesn't mean that those who cares and have money to spend are idiot!
You really make yourself looks like an intolerant fool here: if you don't care about something, nobody should? Tss, let's talk about being open-minded!
>Pascal's successors (Oberon and others) are in active use at my university.
And universities are pretty much the only places where Pascal's successors are used: quite a failure when you realise that at a time Pascal was a big competitor of the C language.
The exception is Delphi, but it's not that big either compared to C,C++,Java(?)..
>Imagine needing the exact same qt/gtk version on every server and workstation.. it could become a mess.
Exactly the same version no. The interface would stay the same, even if the server change the way it displays.. There is already the same problem with X: what do you do if your client wants to send information with an X extention that the server doesn't support?
>removal of __P(), removal of the 'register' keyword and ANSIfication of old K&R code
*removal of __P()? Could you explain what this is? I suppose that this is FreeBSD specific..
*removal of the 'register' keyword If memory serves, the register keyword is only a hint to the compiler, so a global search and replace should be enough to remove the register keyword, no?
*ANSIfication of old K&R code K&R, ouch my memory is fading, but I think that some preprocessor can do it automatically, no?
Not that I'm not criticising: cleaning old code is good, but the two points I understood do not seem too difficult.
The thing which looked interesting in DragonFly is the execution of system call within the user's context if it is possible (I've read that the AmigaOS could do this before). I'm wondering what is the main problem to do this for an OS? There must be some difficulties otherwise every OS would do it..
Slashdot Mirror
>I still think that int should be the easiest size for the machine to deal with.
Correct, but any 64-bit CPU that I know about about instructions for manipulating 32bit variables easily too.
These "64bit" CPU manipulate 32bit value as easily as 64bit values, which means K&R rules that the int should be the "natural" int of the CPU doesn't tell you anything about wether an int should be 32 or 64 bit..
So I'm for the int=32bit and long=64bit rule..
Still in some case, the start-up times of some proprietary things are much better than the equivalent on OpenSource:
BeOS booted to the graphical desktop in ~10s! On "old" machine even..
That is the equivalent of booting the kernel, starting X and KDE, which can easily takes minutes on slow PC (or with little memory)..
I'm not sure if OpenBios is interesting for normal (non-embeded) usage..
I didn't find the list of hardware supported, I guess it is tool early..
>only on beta code
Well I suppose that this beta code got out of Microsoft, otherwise the DR-DOS guys wouldn't have had any case.
Usually beta version are used by 'selected customers' (big enterprises) to evaluate, so they are important!
>Do you have a timeline + evidence of Microsoft doing this to another technology?
They added error messages on Windows (3.1?) so that people would get scared when they would run Windows on another DOS than MS-DOS.
There was some litigation (with Caldera I think), but Microsoft settled for some money change (for them).
I don't know how this got to be modded interesting but, here I go:
>will the DNA auto-repair any damage to the wire?
Uh, DNA in itself is just a big molecule which isn't capable to do anything at all!
A cell is capable to repair its own DNA, but obviously the DNA they use is not inside a cell, where they put the wire in place..
>Couldn't a virus (biological, not computer) be used to re-write the DNA strand that is used to construct the devices, to make different components for sinister purposes?
If a virus modified the DNA, the most probable outcome (by very far) is that you would get a malfunctionning device ie the transistor won't work correctly..
You're right, helicopters are porr flying cars:
- they are noisy as hell
- you have to be very skilled to "drive" one, and even more skilled to avoid being killed if there is a failure.
- half of their time is spent on maintenance.
Anyway heavy maintenance is needed for *any* flying cars, otherwise there would be lots of death caused by malfunctions.
I suspect that we won't be able to make products "perfect enough" for flying cars before the "nanotech world".
>Most signs of intelligence can be mimiced by sufficent brute force.
:-)
It is very true, and I think that our intelligence is only the result of our brain "brute force".
Well, it is not the same way of processing than computers, but parallel processing is quite good too..
Mode the parent up.
I wish like anybody else that OLED would be here now, but it has been reported that it still have problems getting uniform colour over a large surface for a long period of time..
Be careful if your in the US and you want to do the 3-channel alpha thingy, if I remember correctly a company (maybe Apple?) has a patent on having a per-color alpha channel..
Yes, it seems stupid, but stupid patents are quite common..
I think that all these proposal to increase the amount of energy avoid a potential problem: the corresponding increase of heat generated..
Eventually all this energy will turn into heat, so it is quite possible that this will eventually raise earth's temperature..
I think that it may be wiser to increase the efficiency usage of energy than to increase the amount of energy used, well unless of course we need to warm up the earth..
>can we even make useful robots at normal scale that self-replicate?
Well to product a robot at a normal scale, you'd have to produce parts with very high precision: motors, electronic,etc..
So in fact, producing true self-replicating "normal bots" is almost as difficult as producing nanobots.
We could probably produce, self-replicating bots which use lego-like elements, but what would be the point?
And there is BIG difference between going higher than the speed of light and creating nanobots.
The first one is violating a physical principle, the other is "recreating" partly what nature has already achieved with life..
>The article says that they could do it from any mammalian blood, but use human blood for "ethical reasons". PETA members, perhaps?
I don't know. The question is: have they really stripped out 100% of the "nasty things"?
If it is not 100%, but 99.99% using cow's blood could create new illness: animals illness transmited to men..
>In France, 90% of our electricity comes from nuclear plants and it's reasonnably cheap.
True, but AFAIK:
1) right now waste have not been truly disposed of, just treated "temporarily" in the recycling plant of LaHague. This is a political "hot potatoe", and I'd like that a permanent solution be found.
2) No nuclear plant have been disassembled, so the cost of the disassembling the nuclear plant is only theoretical, if they save not enough money for that, the cost of electricity will go up..
And indenpendently from this, there is a problem with the lack of transparency of the nuclear industry in France: when the Tchernobyl clouds of radioactivity went over France, officials lied about the level of radioactivity measured.
I'm not against the nuclear industry, but all the reasons give above makes me "cautious" about it.
On one end you have Ariane5 exploding because of a software error, on the other end, you have 10 clients within your enterprise which loose time because of software's bugs.
With such huge range of differing costs for finding the bug before or after the shipping of your product, the "average cost" of bugs is meaningless.
I think that the only thing to remember is:
- bugs found late cost more to fix than bugs found earlier (any specific number is invalid)
- finding bugs early is difficult and can be expensive.
Of which you can deduce that:
- if late bugs can cost you very much (Ariane5 for exemple), you want to spend a lot of money on software testing|review at each level.
- otherwise if tests can cost more than the fix (a small number of internal users with a non-critical software), then maybe you can use the clients as testers, but it must be managed well (tell the users, be in close contact with the users, don't let them wait the fixes too much).
>Does anybody care about Mercedes's share of the entire automotive industry? Of course not.
Your analogy is flawed: you don't use a car to run software, whereas a computer is useless if you cannot run on it the software you want.
Mod the parent up, if memory serves, Microsoft have had up to three years of being late for some OS, with major feature post-poned.
So with Linux you get no roadmap, with Microsoft you get very unreliable roadmap, I'm not sure Microsoft has a very big advantage here..
Well I think that you treat a bit lightly the complexity of the "issue unit", if you want to have the maximal single thread performance, a 'single issue multiple units' makes sense, but every "brainiac" CPU usually has some limits on the way it can reorder the instructions, otherwise it would become too complex.
Look at the PPC970, it is a deep and wide CPU: it has so many instructions in flight that it packs the instructions into bundle, and sometimes some slot can be unused because of the limitations of the issue unit..
A multi-core CPU simplifies the logic of the issue unit: the Power5 is going to be SMT and SMP..
You are right that now many applications are programmed with the single-threading model, but in a few years, even entry-level CPU will be SMT (hopefully with less bottlenecks than the current P4) or SMP, so multi-threaded applications will gain a *big* advantage..
>So many pirates were private military contractors,
;-)
These were called "corsaires" *not* pirates!
Confusing both is like calling a soldier a hitman, technically soldiers are payed murderers like hitmen, but I'm sure that hitmen would feel insulted if you said that they were just soldiers
>Anyone buying an Athlon-64 primarily to run the latest incarnation of Doom or Office is simply an idiot.
For Office I agree, but for games, why not?
Because you don't care for games, doesn't mean that those who cares and have money to spend are idiot!
You really make yourself looks like an intolerant fool here: if you don't care about something, nobody should?
Tss, let's talk about being open-minded!
>Pascal's successors (Oberon and others) are in active use at my university.
And universities are pretty much the only places where Pascal's successors are used: quite a failure when you realise that at a time Pascal was a big competitor of the C language.
The exception is Delphi, but it's not that big either compared to C,C++,Java(?)..
Except that the clipboard selected by Ctrl+C/X is different from the one used by the mouse select.
So you can use two clipboard easily, I use the "mouse select" for simple, short term copy and the other way for more "long term" memory.
And I really miss the Unix way to do the copy/paste, when I'm using Windows!
>Imagine needing the exact same qt/gtk version on every server and workstation.. it could become a mess.
Exactly the same version no.
The interface would stay the same, even if the server change the way it displays..
There is already the same problem with X: what do you do if your client wants to send information with an X extention that the server doesn't support?
I did a google search before posting, but without success, thanks for the link.
;-)
>The problem probably isn't doing the work, but getting the changes accepted by the powers-that-don't-like-change.
Well now that he has his own codebase, this shouldn't be too much a problem
>removal of __P(), removal of the 'register' keyword and ANSIfication of old K&R code
*removal of __P()?
Could you explain what this is? I suppose that this is FreeBSD specific..
*removal of the 'register' keyword
If memory serves, the register keyword is only a hint to the compiler, so a global search and replace should be enough to remove the register keyword, no?
*ANSIfication of old K&R code
K&R, ouch my memory is fading, but I think that some preprocessor can do it automatically, no?
Not that I'm not criticising: cleaning old code is good, but the two points I understood do not seem too difficult.
The thing which looked interesting in DragonFly is the execution of system call within the user's context if it is possible (I've read that the AmigaOS could do this before).
I'm wondering what is the main problem to do this for an OS?
There must be some difficulties otherwise every OS would do it..