Re:Good Idea for multistandard TV
on
Linux TV
·
· Score: 1
I already have a PAL compliant TV. Oh, wait, now I don't have to buy a NTSC compliant TV if I enter the states. Still no solution for the NTSC vhs tapes I get all the time..
Yeah, I do use that sometimes. You're right, it would be really useful if I made the keyboard-shortcut, and open it up in a new frame. But I'm a bit hesitant to configure emacs too much, it wouldn't be the first time I work all day on custumization 'to increase my productivity'... I do use info often in xemacs, that interface makes me hate info a lot less. Man pages work great in terminals, so that's what I use.
I agree, man pages rock, especially to just look up the signature of a function. And apropos of course (man -k). That make it really easy and fast to compensate for sloppy memory. Fast crucial here, any gui interface to man, like in nautilus, is too slow to be useful. For that kind of reading, html is more appropriate I would say.
But the gtk/gnome documentation is in sgml. I guess it's not to difficult to generate man-pages from it. Just the way the gtk html-reference is organized would cut it for me: huge man-pages with several functions per page. Just as long as there is a sym-link for every function to the page.
Ok, nobody will read this, but I just feel strongly about this. Please e-mail if you want to continue this discussion, Mr Skreet Nite.
I just really hate it if some questions are not allowed to be asked. Of course I may try to correlate race with intelligence! The outcome would possibly be: no valid answer can be given, since intellegence is ill-defined. Or race is ill-defined. Or maybe you can make some qualificationsm, probably something like "races are to uniform and intermixed for complex differences like brain-functioning to show. Only superficial traits like skin-color might differ". But you can ask every question, and try to answer it scientifically.
I just want to point out racism is a question of science
No, it isn't. And it never was. The racists had a theory and then set out to prove it. They never set out to disprove it. That is in itself unscientific. For in depth explanations try looking at the work of Stephen J Gould who nails their kind of pseudo science to wall.
Ok sorry, I should not have said racism. I meant something like 'studying genetic differences between races' or something. Still wrong, I use the word races. Maybe behavioural genomics, as another poster says. I did not mean the 'research' of the nazi's, or others of the like. Science can't have an agenda. I just hate political correctness, I like to use shocking words with an innocent meaning, sorry
On a related note, i won't be insulted if you compare me with Hitler. Only if the conclusion of your comparison is that we're similar, then I'll be mad!
BTW, have you tried to imagine a society where races quite obviously do differ, including in intelligence? This was probably the case with Neanderthalers and modern humans some 40.000 years ago. It must have been an awful society, and I'm glad we are all created equal. But that is probably only so because we killed everybody that was different:-(
henk@huitema.org
It doesn't surprise me because the very idea that there is such a thing as different races has never been actually proved, and is now definitively disproved.
Yes, ok, racism is luckily quite convincingly disproved (quite a while ago), but why doesn't that surprise you? It could easily have been otherwise!
And about the science of racism: it can be done well. There is nothing wrong with measuring skull sizes etc. and trying to correlate it with race and things like intelligence. It only gets wrong if you have prejudices about it, and it gets outragious if you base politics on dubious claims. But asking the question should be allright. BTW, it was not only Nazis that believed in racism; all 19th century books I've read were quite 'racistic'.
As an aside, by the law in my country, I am not even allowed to correlate race with anything. Racism is illegal here. Luckily the science backs it up, finding anything else is illegal!
To make it clear how ridiculous that can get: we had a new rule here that made it compulsary for some dog-races to be muled (pittbull etc.). There were some demonstrations which disproved against racism wrt dogs! Racism for dogs in is quite appropriate i would say.
(Yes of course I see the dangers of racism. I just want to point out racism is a question of science, you shouldn't be to dogmatic about it (pun intended). Probably I underestimate the dangers of racism in the current educuated age, and I should be more weary of it).
The story of man has often been about the search for a scapegoat before whom we can lay responsibility for our evil deeds. From a philosophical stand-point we've always known that this is wrong-headed; now science is confirming what we've long known in our hearts
We will always be responsible for our deeds, that's how we feel it, that is how humanity works. Science will never be able to change this. It might however be able one day to explain how it works. We shouldn't be afraid of understanding, it could never change the way we are. It might prove some religious feelings wrong though.
One of the 'surprises' of the results of the mapping was that there is no difference between so-called races. Well, IANAG but I could have told them that.
Why doesn't that surprise you? I think it's a real wonderful surprise that there is no prove of any significant intellectual differences between 'races'. I could easily picture it otherwise. What if here were still Neanderthalers around, provably a little bit less (or more) stupid then we are? The social problems we have now between different groups would be a lot worse I would guess!
Luckily, the different races in humans seem to hardly vary in intellectual respect. Phew! The social problems of different environment alone are hard enough.
Recent discoveries show many genes code for up to 10 proteins. These proteins all have start codons on different parts of the same gene. Earlier estimates of gene count were often base on the number of start codons, that's one of the reasons for the appearant over estimation.
Humans are clearly more complex than roundworms. If we have hardly more genes then roundworms, then the complexity must stem from other reasons. Craig Venter suggested in his press release environmental factors are more important. That's too short a call for me. I agree, in the 70's everything was attributed to environmental factors. Since then, we've seen the acceptance of socio-biology, and maybe an overstressing of biological factors. But comparing with a roundworm, we clearly need some biological explanations.
If the number of genes is about the same, this complexity must lie in the complexity per gene, or in the complexity of how genes interact. Well, both show plenty of possibility for the missing complexity. Multiple proteins per genes is obvious; the intricacies of complex gene interactions is also a large potential source of ingenious heritage of information. Too be complete, junk maybe not junk. And for a real stretch: maybe some organization in egg-cells other than DNA is passed as information to offspring-cells.
Well said. Of course it's all about emergent phenomena
But I really hate the way Gould always makes this 'emergency' into some sort of mysticism. That's also the gist of this article: Gould says triomphantically: I told you so! It's all a lot more complicated! As if anybody ever claimed to understand the lot.
Gould's eternal enemy, Dawkins, has in my view a much saner approach: reductionism. Sure, the complexity is enormous, but if we don't try we'll never understand it. And of course the simple models we can make up using reductionism don't show the full spectrum, but at least it's a path to more knowledge. Instead of ridiculing them, as Gould does, you can improve them.
Gould's remarks about religion and science not biting eachother clearly shows were he stands: he has a deep religous feeling about the underpinnings of the world we live in, and he would be truely disappointed if we would unravel too much of the mystery. Not a healthy attitude in science, I would say.
Yes, this is prejudiced, based on the few things I read. I could be misguided...
It seems many human genes code for multiple proteins. In my naive view, this code easily solve the mystery; more complexity per gene instead of more genes....
No need to review nature-nurture debates as Craig Venter suggests (although those debates are always nice...)
Years ago a guy at our chemistry lab wrote a game where you could catch food on your head.
It used an SGI Indycam. It was quite fun to play,
but a bit inprecize, and background dependent.
See http://rugmd4.chem.rug.nl/hoesel/lu mbu s.html.
Re:Should the hard part be server or client-side?
on
XFree & Rendering
·
· Score: 1
Wrong. It starts to shift things to the server side.
O, wrong, i'm an idiot, i now see what you meant, and you were right, and a smart answer was already submitted... I'll stick to reading now:-(
Re:Should the hard part be server or client-side?
on
XFree & Rendering
·
· Score: 1
Why not move old API to extensions and new API to the X server core?
Well, yes, for the implementation side that is I guess the goal for Xfree5
But I meant the spec-side, API + wire protocol. There, you have to be backwards compatible and use extentions.
By the way, just read on the mailinglist that Keith already hacked in some font support, so all the key features seem to be there for a demo. Exciting to watch...
Re:Should the hard part be server or client-side?
on
XFree & Rendering
·
· Score: 1
This extension is starting to shift some of the load onto the client side
Wrong. It starts to shift things to the server side.
You seem to think that text should be rendered by the client using the new primitives. This would be a step back. But this is just the core, lower part of the new rendering engine, the introduction of new primitives (which can be hardware-accelareted) and an alpha-channel.
Once this infrastructure is inplace, higher level things, like text-api, can start to use it. It has all to be a new api though (X extensions), since IIRC the old api ties text to 2 color bitmaps.
I guess to grand overall back of the mind strategy is to make now X-extentions for all things that have to be improved, then when every toolkit and many applications use only the extensions, we can start thinking about X12 and doing away with the old cruft. At least, that's what i hope...
Focus on a stable architecturable basis, and implement only basic features. In this way you can release early, and start building a userbase since your software is actually useful for some. Then many hackers will get itches, and hopefully it will be easy for them to start fixing bux/adding features. Once you get it going and approach 1.0 you can start thinking big for 2.0
Just my 2 cents. Look at mozilla for another approach:)
Well, here in Holland we have one protocol (GSM), but 5 providers who all have a nation wide (well ok that's not much) network, so it is not uncommon to see 3 antennas on buildings here...
At least that's what a new report of the EU says...
"The European Commission reported on ? that European and foreign-owned firms seeking access to the US market for
communication services still face considerable barriers, particularly in the satellite services and the mobile services
sectors. According to the Commission "this situation is not in line with the market access policy advocated by the US, and
provides a competitive advantage to the significant number of US companies that have already access to the European
market in these fields." Furthermore, the Commission has also identified a number of US laws and policies on Internet and
electronic commerce which could impact negatively on the business of European companies, particularly in the fields of
Internet domain names and cybersquatting, encryption and patentability of software and business methods."
I have to wonder what the point is in massive beowulf clusters like these. Sure they are fast
and give you more Mips than flanders next door, but they surely lack the memory bandwidth
that makes traditional mainframes and supercomputers so powerful.
If you want massively parallel systems then I would honestly think that something like
processtree would be a good solution since you can rent a phenomenal block of cpu time.
Well, obviously these machines are something inbetween the extremes you mention, and there are
applications for which this is sort of a sweet-spot.
I have used an application for which this type of machines are excellent: molecular dynamics
simulations.
The usual strategy for this type of software is to partion your system by giving every proc a share of
the atoms. Then you start calculating forces and motions etc for each part for a short time period, and
then compare them. Many forces extend to neighbouring parts, and atoms can move to other parts,
so quite a lot of communications between the nodes is necessary. After exchanging this info, each
node can compute the next timestep. This works quite well if most interactions between atoms are
relatively short.
This type of app is excellently suited for a large cluster. It is naturally suited for message-passing, so
programming it using MPI is easy. If you partion the system well, the memory use of one node is quite
small, and fits for a large part in cache.
IO between nodes has to happen quite often, so latency is a problem. So processtree is obviously no
option.
These simulations scale quite well to larger molecular systems. Unfortunately, many researchers don't
want more atoms in their systems, they want the simulation of their small system done faster.
Unfortunately, this scaling is bad; if you end up with only a few atoms per node, the communication
overhead boggs it down.
FYI, here are some old benchmarks of the software i used (gromacs). Although this software is
considered to scale excellent, a 64 node machine is only 32 times as fast as a single-node machine...
Sorry if all this is incrompehensible, i guess i want to say too much too fast...
Not if you compare it with this machine (which is ranked 5th on the top 500 supercomputers list).
Ok, that one is faster... (>770 MB/s internode using MPI, no mentioning of latency). But it doesn't qualify as a beowolf-style machine; it is all specialized Hitachi stuff.
This compaq is in my opinion 'beowolf-style': it uses standard 8-way SMP machines using PCI network cards and fast switches for interconnection. For this, the QSW products still look impressive to me.
At least, they seem to be custom-built by a company the sepcializes on such things.
Which has an excellent product page
here. 2.35 usec latency for a short message. 340 MB/s peak, 210 MB/s sustained throughput. Fault tolerant redundant links. Tru64, Solaris and Linux support. I know nothing about this, but it sounds impressive to me.
Slashdot Mirror
I already have a PAL compliant TV. Oh, wait, now I don't have to buy a NTSC compliant TV if I enter the states. Still no solution for the NTSC vhs tapes I get all the time..
Yeah, I do use that sometimes. You're right, it would be really useful if I made the keyboard-shortcut, and open it up in a new frame. But I'm a bit hesitant to configure emacs too much, it wouldn't be the first time I work all day on custumization 'to increase my productivity'... I do use info often in xemacs, that interface makes me hate info a lot less. Man pages work great in terminals, so that's what I use.
But the gtk/gnome documentation is in sgml. I guess it's not to difficult to generate man-pages from it. Just the way the gtk html-reference is organized would cut it for me: huge man-pages with several functions per page. Just as long as there is a sym-link for every function to the page.
(Yeah, should code it myself I guess..)
Microsoft could also still compete with win32 of course, no easy risk of a monopoly there.
I just really hate it if some questions are not allowed to be asked. Of course I may try to correlate race with intelligence! The outcome would possibly be: no valid answer can be given, since intellegence is ill-defined. Or race is ill-defined. Or maybe you can make some qualificationsm, probably something like "races are to uniform and intermixed for complex differences like brain-functioning to show. Only superficial traits like skin-color might differ". But you can ask every question, and try to answer it scientifically.
I just want to point out racism is a question of science
No, it isn't. And it never was. The racists had a theory and then set out to prove it. They never set out to disprove it. That is in itself unscientific. For in depth explanations try looking at the work of Stephen J Gould who nails their kind of pseudo science to wall.
Ok sorry, I should not have said racism. I meant something like 'studying genetic differences between races' or something. Still wrong, I use the word races. Maybe behavioural genomics, as another poster says. I did not mean the 'research' of the nazi's, or others of the like. Science can't have an agenda. I just hate political correctness, I like to use shocking words with an innocent meaning, sorry
On a related note, i won't be insulted if you compare me with Hitler. Only if the conclusion of your comparison is that we're similar, then I'll be mad!
BTW, have you tried to imagine a society where races quite obviously do differ, including in intelligence? This was probably the case with Neanderthalers and modern humans some 40.000 years ago. It must have been an awful society, and I'm glad we are all created equal. But that is probably only so because we killed everybody that was different :-(
henk@huitema.org
Yes, ok, racism is luckily quite convincingly disproved (quite a while ago), but why doesn't that surprise you? It could easily have been otherwise!
And about the science of racism: it can be done well. There is nothing wrong with measuring skull sizes etc. and trying to correlate it with race and things like intelligence. It only gets wrong if you have prejudices about it, and it gets outragious if you base politics on dubious claims. But asking the question should be allright. BTW, it was not only Nazis that believed in racism; all 19th century books I've read were quite 'racistic'.
As an aside, by the law in my country, I am not even allowed to correlate race with anything. Racism is illegal here. Luckily the science backs it up, finding anything else is illegal!
To make it clear how ridiculous that can get: we had a new rule here that made it compulsary for some dog-races to be muled (pittbull etc.). There were some demonstrations which disproved against racism wrt dogs! Racism for dogs in is quite appropriate i would say.
(Yes of course I see the dangers of racism. I just want to point out racism is a question of science, you shouldn't be to dogmatic about it (pun intended). Probably I underestimate the dangers of racism in the current educuated age, and I should be more weary of it).
We will always be responsible for our deeds, that's how we feel it, that is how humanity works. Science will never be able to change this. It might however be able one day to explain how it works. We shouldn't be afraid of understanding, it could never change the way we are. It might prove some religious feelings wrong though.
Why doesn't that surprise you? I think it's a real wonderful surprise that there is no prove of any significant intellectual differences between 'races'. I could easily picture it otherwise. What if here were still Neanderthalers around, provably a little bit less (or more) stupid then we are? The social problems we have now between different groups would be a lot worse I would guess!
Luckily, the different races in humans seem to hardly vary in intellectual respect. Phew! The social problems of different environment alone are hard enough.
You imply that if we would ever find out how it all works, we wouldn't be responsible for our deeds anymore. I fail to see that consequence.
Recent discoveries show many genes code for up to 10 proteins. These proteins all have start codons on different parts of the same gene. Earlier estimates of gene count were often base on the number of start codons, that's one of the reasons for the appearant over estimation.
Humans are clearly more complex than roundworms. If we have hardly more genes then roundworms, then the complexity must stem from other reasons. Craig Venter suggested in his press release environmental factors are more important. That's too short a call for me. I agree, in the 70's everything was attributed to environmental factors. Since then, we've seen the acceptance of socio-biology, and maybe an overstressing of biological factors. But comparing with a roundworm, we clearly need some biological explanations.
If the number of genes is about the same, this complexity must lie in the complexity per gene, or in the complexity of how genes interact. Well, both show plenty of possibility for the missing complexity. Multiple proteins per genes is obvious; the intricacies of complex gene interactions is also a large potential source of ingenious heritage of information. Too be complete, junk maybe not junk. And for a real stretch: maybe some organization in egg-cells other than DNA is passed as information to offspring-cells.
But I really hate the way Gould always makes this 'emergency' into some sort of mysticism. That's also the gist of this article: Gould says triomphantically: I told you so! It's all a lot more complicated! As if anybody ever claimed to understand the lot.
Gould's eternal enemy, Dawkins, has in my view a much saner approach: reductionism. Sure, the complexity is enormous, but if we don't try we'll never understand it. And of course the simple models we can make up using reductionism don't show the full spectrum, but at least it's a path to more knowledge. Instead of ridiculing them, as Gould does, you can improve them.
Gould's remarks about religion and science not biting eachother clearly shows were he stands: he has a deep religous feeling about the underpinnings of the world we live in, and he would be truely disappointed if we would unravel too much of the mystery. Not a healthy attitude in science, I would say.
Yes, this is prejudiced, based on the few things I read. I could be misguided...
No need to review nature-nurture debates as Craig Venter suggests (although those debates are always nice...)
Years ago a guy at our chemistry lab wrote a game where you could catch food on your head. It used an SGI Indycam. It was quite fun to play, but a bit inprecize, and background dependent. See http://rugmd4.chem.rug.nl/hoesel/lu mbu s.html.
And i'm assuming you are seeing the sum of all threads... Top on Linux currently can't distinguish threads and processes.
I read something about an extra field in the process-table being implemented for clone, so this might be fixed in 2.4.
(Finally, a discussion where i can gripe about miserable support for pthreads on Linux :)
... from autobi
O, wrong, i'm an idiot, i now see what you meant, and you were right, and a smart answer was already submitted... I'll stick to reading now :-(
Well, yes, for the implementation side that is I guess the goal for Xfree5
But I meant the spec-side, API + wire protocol. There, you have to be backwards compatible and use extentions.
By the way, just read on the mailinglist that Keith already hacked in some font support, so all the key features seem to be there for a demo. Exciting to watch...
You seem to think that text should be rendered by the client using the new primitives. This would be a step back. But this is just the core, lower part of the new rendering engine, the introduction of new primitives (which can be hardware-accelareted) and an alpha-channel.
Once this infrastructure is inplace, higher level things, like text-api, can start to use it. It has all to be a new api though (X extensions), since IIRC the old api ties text to 2 color bitmaps.
I guess to grand overall back of the mind strategy is to make now X-extentions for all things that have to be improved, then when every toolkit and many applications use only the extensions, we can start thinking about X12 and doing away with the old cruft. At least, that's what i hope...
Just my 2 cents. Look at mozilla for another approach :)
Well, here in Holland we have one protocol (GSM), but 5 providers who all have a nation wide (well ok that's not much) network, so it is not uncommon to see 3 antennas on buildings here...
If you want massively parallel systems then I would honestly think that something like processtree would be a good solution since you can rent a phenomenal block of cpu time.
Well, obviously these machines are something inbetween the extremes you mention, and there are applications for which this is sort of a sweet-spot.
I have used an application for which this type of machines are excellent: molecular dynamics simulations.
The usual strategy for this type of software is to partion your system by giving every proc a share of the atoms. Then you start calculating forces and motions etc for each part for a short time period, and then compare them. Many forces extend to neighbouring parts, and atoms can move to other parts, so quite a lot of communications between the nodes is necessary. After exchanging this info, each node can compute the next timestep. This works quite well if most interactions between atoms are relatively short.
This type of app is excellently suited for a large cluster. It is naturally suited for message-passing, so programming it using MPI is easy. If you partion the system well, the memory use of one node is quite small, and fits for a large part in cache. IO between nodes has to happen quite often, so latency is a problem. So processtree is obviously no option.
These simulations scale quite well to larger molecular systems. Unfortunately, many researchers don't want more atoms in their systems, they want the simulation of their small system done faster. Unfortunately, this scaling is bad; if you end up with only a few atoms per node, the communication overhead boggs it down.
FYI, here are some old benchmarks of the software i used (gromacs). Although this software is considered to scale excellent, a 64 node machine is only 32 times as fast as a single-node machine...
Sorry if all this is incrompehensible, i guess i want to say too much too fast...
Ok, that one is faster... (>770 MB/s internode using MPI, no mentioning of latency). But it doesn't qualify as a beowolf-style machine; it is all specialized Hitachi stuff.
This compaq is in my opinion 'beowolf-style': it uses standard 8-way SMP machines using PCI network cards and fast switches for interconnection. For this, the QSW products still look impressive to me.
Which has an excellent product page here. 2.35 usec latency for a short message. 340 MB/s peak, 210 MB/s sustained throughput. Fault tolerant redundant links. Tru64, Solaris and Linux support. I know nothing about this, but it sounds impressive to me.
I thought the intention is to save the government some money; "Give the Samba guys the specs they need so we don't need server licenses"