Sorry but Intel have to take the blame also from many of the problem of the x86: when they switched from 16 to 32, they could have added more registers the way AMD did with x86-64. Also there is no reason why MMX, SSE and the like suck (sharing MMX registers with FPU registers for MMX was braindead) while PPC has Altivec.. And lets not forget about Netburst for the implementation part..
I can undestand why you like the features of OpenFirmware and while I hate x86 ISA, may I ask why do you care about the CPU used, are you doing assembly programming? If yes, I can understand, otherwise that doesn't make sense, who cares what happen between the compiler and the CPU?? Only price, performance, power consumption matter!
I think that having one fixed head per track is not very economical.. Now one thing that I've always wondered is why not several moving heads per disk? With two heads, this would lower the spinning latency from 1/2 to 1/4 if you make those two heads seek the same data or it would allow twice the bandwith..
Granted you can do the same thing with RAID, but while the added reliability is nice, it has also twice the power consumption, volume used, noice, heat..
And the drawback is that when the compiler is not able to predict correctly the flow of execution (for example, you have a bubble in your pipeline due a cachemiss), then the HW cannot reorder the instruction to 'work-around' the problem..
For me VLIW is a 'hit or miss' architecture: if you have very regular code, it can be very efficient, but if you have 'normal code' (OO with indirections, etc..) then it will suck. For the very regular code part, normal CPUs have also vector units which can help on some of these codes.. Anyway compared to the x86 juggernaut, even RISC failed, VLIWs don't stand a chance even supported by Intel: as shown by the x86-64 episode, Intel is not all powerful when it comes to dictating CPUs future..
Sorry but this is not contradicting to praise one and not the other. RISC are optimised for compilers: - compilers sucks with CPUs with few registers and work very well with big number of registers? Let's make CPU with a big number of register. - compilers suck with weird instruction? Let's use orthogonal instruction set, etc..
It worked very well: RISC with much less transistors were able to compete with CISC..
VLIW is different: OOO CPUs are too complex to make? Let's use just a bunch of ALUs, make them available to the compiler and hope that it will be able to use them efficiently. The only problem is that compilers are not able currently to use efficiently VLIWs except on very specific type of code..
Just a side note about your point 2: not everybody is living in the US, you know..:-)
I'm French and if memory serves, there is about 30% of French who are "non-believers", and among the believers only a little are going to church even once a year..
That said, even in my country, religious people would be hard pressed to find themselves "oppressed", it is just that people are loosing interest in religions (especially organised religions)..
>> Unfortunately, there is a history of people who are witnesses to spiritual events, even miracles, who openly dismiss them for one reason or another.
Uh? 1) spiritual events mean nothing, that people in a desesperate situation try to cling to religion as a way to keep their hopes is normal, but it doesn't mean that religions are true, just that it do work as an 'opium for the people'.
Anyway are you suggesting that their liberation was a miracle? What about those who died before being able to return home?
2) Currently I only know about three type of "miracles": -obvious scam -strange physical events: for the one who were real events, all have been explained without breaking physics law, the other only relies on reports and the events didn't reoccur.. -miraculous health recovery: we know so little about how human's body work that it isn't funny..
Now if someone is able to fly in front of me, or walk on the water, and that it is proven that it is not a trick, I would call this a miracle.
As for the 'going to church' part, in Europe people don't need to go to church to pray either, but when they go to church, they do it to pray, so it isn't simply a meeting as those atheist guys do..
Also about the religious freedom: 1) in France we have religious freedom protection, but for the private life, public schools are considered as a public institute, so no obnoxious religious sign are tolerated: a jewel will be accepted but not if it is too big, veils are not accepted. 2) Knowing that at the investiture, the president of the USA swears on the bible, somehow the religious freedom in the US seems to me a little weird, it would be very funny to hear the turmoil it would create, if one president would want to swear on the Coran for example:-)
> As an atheist, I hope you can appreciate that atheism is a religious viewpoint as well.
Nope, because my point of view 'there is no god' is falsifiable: if I'm witness to real miracles, I'd become religious (religious as in beleiving in supreme beings not as being part of a particular religion of course), wouldn't you?
Which make this *very different* from the faith of religious people.
As for your meetings that you say are equivalent as "going to church": usually when people go to church, they do it to pray, so..
For the rest, I'm French and we do have very strict limitations on religious right of people in the schools, and I agree with those. And we have no backlash, currently at least..
OK, so I won't tell you why it won't work but ask you questions: -how do you plan to make CBS OS and CBS based applications available to everyone? -does building CBS applications require a new language or can it be done with existing language such as C,C++?
Well agreed but by *every religion* include these type of elements, not only church as you said, but all the religions..
So if there was some kind of god without those attributes, it would be totally different from what is currently described as god, so why name it 'god'?
I don't rule out that "something unkwnow" exist, but as I know nothing about it, I apply Occam razor: nothing exist until you can proove me otherwise --> atheist.
>>> Part of the problem is that some people assume they know stuff when they really don't. They believe, but they don't actually know. For example, somebody was saying the other day that he knew all along that Iraq didn't have WMDs. Um, no, he didn't. He hadn't ever even been to Iraq. Heh.
You're right, but maybe he believed what the UN inspectors who have been to Irak told? It isn't first hand knowledge of course, but this is the same about many things: you know that relativity is correct, right? But have you checked it yourself?
There are several definition about atheism.. But I don't consider myself as an agnostic: I don't think that the two possibilites 'there is no god' and 'there are god(s)' have the same weight! I don't know what happened at the beginning of the universe or why it happened, but I consider that god(s) as described by religions (which have human look, that you can pray, which shield us from death through one way or another etc..) are about as interesting as 'father christmas' as a concept, so I consider myself atheist, not agnostic.
"Everyone you say who says that they have no religious beliefs is just so certain about their belief that they accept it as truth. If you just start asking probing questions, and they start getting mad, then you've found their religion." --Orson Scott Card.
Well, can you expect another view from someone who has a religion? Talk about a biased viewpoint! As an atheist, when someone tells me but how is-it possible than 'physic laws' has produced such complex universe where some little modification in the laws would have destroyed it, I don't get upset, I just answer 'I don't know'.
You know saying that there is a god, is just as bad: who created the god? So this "answer" is not an answer at all..
So we all don't know, but that's the religious people who tend to get upset when one say this: they know, they have their 'faith'.. Bah, talk about wearing blindhorses.
For something to be true or false, it must first be understandable with putting so much buzzwords in your description, you obscured to the point I don't care if it's true or false..
Depends on what are your criteria to decide good or bad but converting every integer into its ASCII representation is not a good over-the-wire protocol IMHO. Not efficient to say the least.
> Reiser, and for that matter VFS and the other myriad of database enhanced filesystems, are tools. Good ones, but tools...
While I agree that filesystems are very different from DB, I consider also that DB are tools: even a super-complex distributed DB is a tool, in my book.
Intelligence is a concept applied to animals, it is a weakly defined concept sure, but this is not a reason to apply this concept to a natural phenomenon (evolution) or to a mythical super-being (God).
Applying concept related to human/animals on natural phenomenom is rarely useful..
Well, I find those LED for brake lights / turn lights highly annoying.. When I'm stopped behind someone who lets his turn lights on, sometimes I used my hand to cover the area to avoid being annoyed by the blinding light.
Do you really believes that games couldn't use your CPU to improve the AI of the characters? the physics simulation? You know someone is trying to sell add-on cards to handle physic simulation because CPUs are too slow..
The reason why you see games not pushing your CPU is that you have a 'top' CPU and game manufacturers will not go through the hassle of trying to improve so subbtle things as the AI or the physics for the benefit of a few people: improved visual is much more likely to give good reviews on a game.
Slashdot Mirror
Sorry but Intel have to take the blame also from many of the problem of the x86: when they switched from 16 to 32, they could have added more registers the way AMD did with x86-64.
Also there is no reason why MMX, SSE and the like suck (sharing MMX registers with FPU registers for MMX was braindead) while PPC has Altivec..
And lets not forget about Netburst for the implementation part..
I can undestand why you like the features of OpenFirmware and while I hate x86 ISA, may I ask why do you care about the CPU used, are you doing assembly programming?
If yes, I can understand, otherwise that doesn't make sense, who cares what happen between the compiler and the CPU??
Only price, performance, power consumption matter!
I don't how a better 'download dependency tool' is any help to the problem discussed before which is: how to avoid package management hell.
While one head per track is bit extreme, I wonder what on one arm they couldn't fit two heads?
Grand this will increase the weight of the arm, but it would also double the bandwith!
I think that having one fixed head per track is not very economical..
Now one thing that I've always wondered is why not several moving heads per disk?
With two heads, this would lower the spinning latency from 1/2 to 1/4 if you make those two heads seek the same data or it would allow twice the bandwith..
Granted you can do the same thing with RAID, but while the added reliability is nice, it has also twice the power consumption, volume used, noice, heat..
And the drawback is that when the compiler is not able to predict correctly the flow of execution (for example, you have a bubble in your pipeline due a cachemiss), then the HW cannot reorder the instruction to 'work-around' the problem..
For me VLIW is a 'hit or miss' architecture: if you have very regular code, it can be very efficient, but if you have 'normal code' (OO with indirections, etc..) then it will suck.
For the very regular code part, normal CPUs have also vector units which can help on some of these codes..
Anyway compared to the x86 juggernaut, even RISC failed, VLIWs don't stand a chance even supported by Intel: as shown by the x86-64 episode, Intel is not all powerful when it comes to dictating CPUs future..
Sorry but this is not contradicting to praise one and not the other.
RISC are optimised for compilers:
- compilers sucks with CPUs with few registers and work very well with big number of registers?
Let's make CPU with a big number of register.
- compilers suck with weird instruction?
Let's use orthogonal instruction set, etc..
It worked very well: RISC with much less transistors were able to compete with CISC..
VLIW is different: OOO CPUs are too complex to make?
Let's use just a bunch of ALUs, make them available to the compiler and hope that it will be able to use them efficiently.
The only problem is that compilers are not able currently to use efficiently VLIWs except on very specific type of code..
Just a side note about your point 2: not everybody is living in the US, you know.. :-)
I'm French and if memory serves, there is about 30% of French who are "non-believers", and among the believers only a little are going to church even once a year..
That said, even in my country, religious people would be hard pressed to find themselves "oppressed", it is just that people are loosing interest in religions (especially organised religions)..
>>
:-)
Unfortunately, there is a history of people who are witnesses to spiritual events, even miracles, who openly dismiss them for one reason or another.
Uh?
1) spiritual events mean nothing, that people in a desesperate situation try to cling to religion as a way to keep their hopes is normal, but it doesn't mean that religions are true, just that it do work as an 'opium for the people'.
Anyway are you suggesting that their liberation was a miracle? What about those who died before being able to return home?
2) Currently I only know about three type of "miracles":
-obvious scam
-strange physical events: for the one who were real events, all have been explained without breaking physics law, the other only relies on reports and the events didn't reoccur..
-miraculous health recovery: we know so little about how human's body work that it isn't funny..
Now if someone is able to fly in front of me, or walk on the water, and that it is proven that it is not a trick, I would call this a miracle.
As for the 'going to church' part, in Europe people don't need to go to church to pray either, but when they go to church, they do it to pray, so it isn't simply a meeting as those atheist guys do..
Also about the religious freedom:
1) in France we have religious freedom protection, but for the private life, public schools are considered as a public institute, so no obnoxious religious sign are tolerated: a jewel will be accepted but not if it is too big, veils are not accepted.
2) Knowing that at the investiture, the president of the USA swears on the bible, somehow the religious freedom in the US seems to me a little weird, it would be very funny to hear the turmoil it would create, if one president would want to swear on the Coran for example
Moderate parent up.
Grand parent 'usability guy' said a stupid thing: every WindowsXP users have zip!
> As an atheist, I hope you can appreciate that atheism is a religious viewpoint as well.
Nope, because my point of view 'there is no god' is falsifiable: if I'm witness to real miracles, I'd become religious (religious as in beleiving in supreme beings not as being part of a particular religion of course), wouldn't you?
Which make this *very different* from the faith of religious people.
As for your meetings that you say are equivalent as "going to church": usually when people go to church, they do it to pray, so..
For the rest, I'm French and we do have very strict limitations on religious right of people in the schools, and I agree with those. And we have no backlash, currently at least..
OK, so I won't tell you why it won't work but ask you questions:
-how do you plan to make CBS OS and CBS based applications available to everyone?
-does building CBS applications require a new language or can it be done with existing language such as C,C++?
Well agreed but by *every religion* include these type of elements, not only church as you said, but all the religions..
So if there was some kind of god without those attributes, it would be totally different from what is currently described as god, so why name it 'god'?
I don't rule out that "something unkwnow" exist, but as I know nothing about it, I apply Occam razor: nothing exist until you can proove me otherwise --> atheist.
>>>
Part of the problem is that some people assume they know stuff when they really don't. They believe, but they don't actually know. For example, somebody was saying the other day that he knew all along that Iraq didn't have WMDs. Um, no, he didn't. He hadn't ever even been to Iraq. Heh.
You're right, but maybe he believed what the UN inspectors who have been to Irak told?
It isn't first hand knowledge of course, but this is the same about many things: you know that relativity is correct, right? But have you checked it yourself?
There are several definition about atheism..
But I don't consider myself as an agnostic: I don't think that the two possibilites 'there is no god' and 'there are god(s)' have the same weight!
I don't know what happened at the beginning of the universe or why it happened, but I consider that god(s) as described by religions (which have human look, that you can pray, which shield us from death through one way or another etc..) are about as interesting as 'father christmas' as a concept, so I consider myself atheist, not agnostic.
Well, can you expect another view from someone who has a religion?
Talk about a biased viewpoint!
As an atheist, when someone tells me but how is-it possible than 'physic laws' has produced such complex universe where some little modification in the laws would have destroyed it, I don't get upset, I just answer 'I don't know'.
You know saying that there is a god, is just as bad: who created the god? So this "answer" is not an answer at all..
So we all don't know, but that's the religious people who tend to get upset when one say this: they know, they have their 'faith'..
Bah, talk about wearing blindhorses.
But does it have an "anonymous" publish/search feature?
For me that is what distinguish a p2p use mainly to share p0rn from BT..
For me, now that Java is here, Pascal has no more advantage.
I'm not a fan of Java, but I think it took the "niche" of 'clean' language to learn at school.
For something to be true or false, it must first be understandable with putting so much buzzwords in your description, you obscured to the point I don't care if it's true or false..
In other words: KISS!!
>But it isn't bad as an over-the-wire protocol.
Depends on what are your criteria to decide good or bad but converting every integer into its ASCII representation is not a good over-the-wire protocol IMHO.
Not efficient to say the least.
> Reiser, and for that matter VFS and the other myriad of database enhanced filesystems, are tools. Good ones, but tools...
While I agree that filesystems are very different from DB, I consider also that DB are tools: even a super-complex distributed DB is a tool, in my book.
> because evolution is a form of intelligence.
Gasp, what a bad analogy..
Intelligence is a concept applied to animals, it is a weakly defined concept sure, but this is not a reason to apply this concept to a natural phenomenon (evolution) or to a mythical super-being (God).
Applying concept related to human/animals on natural phenomenom is rarely useful..
I'm impressed by how little the server use bandwith as well: it returns 'error 500' ;-)
How efficient!
Well, I find those LED for brake lights / turn lights highly annoying..
When I'm stopped behind someone who lets his turn lights on, sometimes I used my hand to cover the area to avoid being annoyed by the blinding light.
Do you really believes that games couldn't use your CPU to improve the AI of the characters? the physics simulation?
You know someone is trying to sell add-on cards to handle physic simulation because CPUs are too slow..
The reason why you see games not pushing your CPU is that you have a 'top' CPU and game manufacturers will not go through the hassle of trying to improve so subbtle things as the AI or the physics for the benefit of a few people: improved visual is much more likely to give good reviews on a game.