>you'll see significantly more IBM machines > across the board then NEC
As widely known and publicly acknowledged also by one of the authors of that list (Prof. Meuer), the linpack benchmark used to build the top500 list is biased against vector supercomputers, like NEC's.
Supercomputer performance cannot be measured by a single number, really.
>especially when you consider it probably can't >even simulate the global weather fast enough to >predict it
The global weather can be - and is, on a daily basis - easily predicted by much slower computers. This one was built for *climate*, not weather simulations. While attending the NEC User Group conference I made some photo of the E.S. complex myself which I might or not put online at my site (see profile).
The early benchmarks on real applications are quite impressive.
Anyway these are old news. I submitted an early story about the Earth Simulator one year and a half ago, when it was being built, and it was rejected. Go figure.
Suffice it to say that even the biased top500 admits that the NEC SX/5, which is the perfect example of a shared memory architecture, can reach an effective performance equal to 96% of the theoretical peak performance, while for a typical SPP cluster it varies from 53% (CompaQ) to 69% (IBM):
http://www.top500.org/ORSC/2000/sx-5.html
Real life figures are more like 50-60% (NEC), 10-15% (IBM) and 5% (CompaQ) --
The top 500 list is currently extremely biased in favour of massively parallel SPP architectures, shared-memory architectures are very underrepresented.
Moreover the top 500 is based on a very unrealistic benchmarks which basically estimates peak performance - which is not relevant in practice - rather than sustained performance.
I've heard of a forthcoming revision of the top 500 criteria. --
I am a physicist with a strong background in chaos theory and I am also one of the team who ported (but not yet optimized) the modified Boulder model on the NEC supercomputer in Lugano/Manno (not Bern, as incorrectly stated).
So yes, the effort is multidisciplinary and - at least in Switzerland - climatologists are well aware of the implications (many of them are physicists).
Of course: that's because unlimited freedom is a contradictory concept (the old "my freedom ends where your freedom begins" saying).
Freedom for all cannot be achieved without some compromises, that's why it is important to keep fighting to conserve it.
Getting back to the case at hand: there is a conflict between the freedom of the ISP to voluntarily use MAPS BGP service to blackhole routes and the right of customers to get connectivity to blackholed sites.
This conflict can be resolved by debate - not like the present one, one between the ISP and its customers - by economic pressure or by a governmental (i.e. legislative) decision. --
The fact that a nation did not sign the international convention does not mean it has no copyright law at all. Many of the nations who did not sign the convention have their own laws about copyright.
>a browser that makes half the web look nothing like the designer intended,
This only goes to show that the above mentioned "designer" used the wrong tool or was otherwise clueless.
First of all HTML is a markup language: how pages appear is - or should be - defined by the browser's user, not by the author of the document. This is explicitly made obvious by the details of the HTML specification: HTML is designed that way, details of presentation are outside its scope.
CSS is the right tool to provide the browser with hints for the actual presentation of the document.
With respect to both HTML and CSS the gecko rendering engine which Mozilla uses is one of the most compliant around, alongside with the one used by Opera. You might not like its defaults, but - unlike what happens in some other browsers - you are allowed to change them. Properly authored, standard compliant documents look as they should. Mozilla is not and does not want to be a MSIE or Netscape emulator, and I certainly hope it would never reproduce their many pitfalls.
If you dislike CSS there are other alternatives for creating designed documents that look basically the same on all systems, such as PDF (which supports hyperlinks).
That said I'll cheerfully concede that Mozilla is not ready for prime time yet.
--
There is one essential caveat: grid-type projects are meant - at this stage - to be essentially private networks linking supercomputing centers for scientific data exchange only. Scientists - having now experienced bandwidth limitations - think that giving indiscriminate Internet access to the public at large has been a big mistake.
Grids are basically meant to be wide area networks allowing services that were previously available in local area networks only, such as load balancing between different supercomputing servers and real time analysis and visualization of several terabytes of simulation data.
I am aware of two such projects, one in the US, and the other one... NDA... (I think I'm allowed to say it isn't even in the planning stage yet, let's say I'm working very, very near one of the possible centers of operation:-).
Of course the technology will be there to be made available - at a later stage - to the general public, but don't hold your breath.
--
>What could realy make my day is publishing
>companies going after Public and University
>libraries and Xerox corporation.
Which they recently did in Italy, where a law forbidding Xeroxing of copyrighted material and withdrawing _almost_ entirely fair use rights has been approved on Sept 8, 2000, under pressure from editors and publishers of academic material (mainly located in the US).
I authored half a dozen scientific articles (Physical Review A, Physical Review Letters, Zeitschrift fur Physik...) and had to transfer my copyright to the publisher in order to get them published at all (academic press is even worse than the RIAA). According to the new law I - the author - can't legally make additional copies for my own private use anymore. Talk about screwed... anyway this is the typical way in which copyright works: it protects the rights of the publisher, certainly not those of the author.
On the other hand I relocated in Switzerland where fair use still holds. And stopped writing scientific articles.
>In essence, every note in an equal tempered scale is a wrong note! With the exception of the octaves, which are exact (that's the whole point).
On a C tempered scale, B is the worst approximated note, a fact that is quite easy to hear.
Western music is based on thirds and fifths: the more mathematically oriented among./ readers might want to build a two dimensional grid with powers of 3/2 on one axis and of 5/3 on the other one to see what happens and why the equal tempered tuning was introduced.
People who know about the continuous fraction expansion of real numbers may also find out why 12 semitones were used by expanding log(3/2) and log(5/3) [well, the historical reason has little to do with continuous fractions, which merely provide an explanation]. IIRC (I am probably slightly off) the next best approximation are at 42 (good third and bad fifth) and 54 subdivisions, both are quite unpractical.
A Faraday Cage around the plane would only shield it plane from EM radiation coming from outside the plane itself: it would be of no help in preventing interference from sources inside the plane.
A Faraday Cage around all EM sensitive instrumentation would be unpractical and possibily unfeasible (since it would also shield radiation meant to be picked up - at least in part).
Do you think you are joking? See below. http://www.nec.co.jp/engl ish/today/newsrel/0005/3001.html BTW, the NEC SX-5 - unlike massively parallel architectures - can effectively run near its theoretical peak performance for most applications, I'd say that the top 40 TFlops performance is a rather conservative estimate (NEC will have newer and faster technology by the time this beast starts being built).
I see no reason why this different licensing scheme could not and should not coexist. The GPL is not taking away any rights: just providing some further rights under a somewhat - often very - restrictive model.
I think that the plethora of available licensing schemes provides more freedom for programmers; anyway nobody is forced to release anything under the GPL unless he's basing his work on GPL'd code to start with.
I am slightly in favor of the BSD scheme vs. the GPL one, but I basically think each programmer should decide for herself.
Following my istinct I would release my code under the public domain, or with no restriction whatsoever: but I've seen code released like this being stolen and copyrighted by somebody else, leaving the original author with no rights to it; that's why licensing schemes such as GPL, BSD or Artistic exist in the first place.
>As the VC community realizes there is little to no money to be made in GPL software, the spigot will close.
I doubt it. It has been around quite a while, and it is expanding.
>as in freedom, as opposed to socialism
You might not realize it, but this is quite an inflamatory remark (to Western European ears, at least). Be more careful. [I, FWIW, happen to hold the exact opposite point of view]
Slashdot Mirror
He was ousted by the company several years ago and has been very critical of them ever since.
We still check in from time to time, though.
They are there as well. The classified computer project of the US government is named ASCI...
A test of a real application showed a performance of 28 Tflops sustained (70% efficiency). Not at all bad.
Super UX is an almost stock version of System V.
>you'll see significantly more IBM machines
> across the board then NEC
As widely known and publicly acknowledged also by one of the authors of that list (Prof. Meuer), the linpack benchmark used to build the top500 list is biased against vector supercomputers, like NEC's.
Supercomputer performance cannot be measured by a single number, really.
>especially when you consider it probably can't
>even simulate the global weather fast enough to
>predict it
The global weather can be - and is, on a daily basis - easily predicted by much slower computers. This one was built for *climate*, not weather simulations. While attending the NEC User Group conference I made some photo of the E.S. complex myself which I might or not put online at my site (see profile).
The early benchmarks on real applications are quite impressive.
Anyway these are old news. I submitted an early story about the Earth Simulator one year and a half ago, when it was being built, and it was rejected. Go figure.
Suffice it to say that even the biased top500 admits that the NEC SX/5, which is the perfect example of a shared memory architecture, can reach an effective performance equal to 96% of the theoretical peak performance, while for a typical SPP cluster it varies from 53% (CompaQ) to 69% (IBM):
http://www.top500.org/ORSC/2000/sx-5.html
Real life figures are more like 50-60% (NEC), 10-15% (IBM) and 5% (CompaQ)
--
They are referring to the Japanese earth simulator:
<http://www.nec.co.jp/english/today/newsrel/000 5/3001.html>
--
The top 500 list is currently extremely biased in favour of massively parallel SPP architectures, shared-memory architectures are very underrepresented.
Moreover the top 500 is based on a very unrealistic benchmarks which basically estimates peak performance - which is not relevant in practice - rather than sustained performance.
I've heard of a forthcoming revision of the top 500 criteria.
--
I am a physicist with a strong background in chaos theory and I am also one of the team who ported (but not yet optimized) the modified Boulder model on the NEC supercomputer in Lugano/Manno (not Bern, as incorrectly stated).
So yes, the effort is multidisciplinary and - at least in Switzerland - climatologists are well aware of the implications (many of them are physicists).
See <http://www.climate.unibe.ch> (I hope that NCCR/Climate gets a real website up and running soon) and <http://www.cscs.ch>
--
>No. Now *you* are asking me to limit my rights.
Of course: that's because unlimited freedom is a contradictory concept (the old "my freedom ends where your freedom begins" saying).
Freedom for all cannot be achieved without some compromises, that's why it is important to keep fighting to conserve it.
Getting back to the case at hand: there is a conflict between the freedom of the ISP to voluntarily use MAPS BGP service to blackhole routes and the right of customers to get connectivity to blackholed sites.
This conflict can be resolved by debate - not like the present one, one between the ISP and its customers - by economic pressure or by a governmental (i.e. legislative) decision.
--
>If you want to defend freedom, you have to defend *all* freedom.
No. Now *you* are asking me to limit my rights.
I reserve my right to fight for what I like and fight against what I dislike. This is pretty much *my* own definition of freedom!
You have the right to disagree, but do not try to define freedom for me.
--
The fact that a nation did not sign the international convention does not mean it has no copyright law at all. Many of the nations who did not sign the convention have their own laws about copyright.
Memo to self: better read the message I'm following up to next time.
Oh well, take the above as a generic rant.
This only goes to show that the above mentioned "designer" used the wrong tool or was otherwise clueless.
First of all HTML is a markup language: how pages appear is - or should be - defined by the browser's user, not by the author of the document. This is explicitly made obvious by the details of the HTML specification: HTML is designed that way, details of presentation are outside its scope.
CSS is the right tool to provide the browser with hints for the actual presentation of the document.
With respect to both HTML and CSS the gecko rendering engine which Mozilla uses is one of the most compliant around, alongside with the one used by Opera. You might not like its defaults, but - unlike what happens in some other browsers - you are allowed to change them. Properly authored, standard compliant documents look as they should. Mozilla is not and does not want to be a MSIE or Netscape emulator, and I certainly hope it would never reproduce their many pitfalls.
If you dislike CSS there are other alternatives for creating designed documents that look basically the same on all systems, such as PDF (which supports hyperlinks).
That said I'll cheerfully concede that Mozilla is not ready for prime time yet.
--
There is one essential caveat: grid-type projects are meant - at this stage - to be essentially private networks linking supercomputing centers for scientific data exchange only. Scientists - having now experienced bandwidth limitations - think that giving indiscriminate Internet access to the public at large has been a big mistake.
... NDA ... (I think I'm allowed to say it isn't even in the planning stage yet, let's say I'm working very, very near one of the possible centers of operation :-).
Grids are basically meant to be wide area networks allowing services that were previously available in local area networks only, such as load balancing between different supercomputing servers and real time analysis and visualization of several terabytes of simulation data.
I am aware of two such projects, one in the US, and the other one
Of course the technology will be there to be made available - at a later stage - to the general public, but don't hold your breath.
--
>What could realy make my day is publishing
>companies going after Public and University
>libraries and Xerox corporation.
Which they recently did in Italy, where a law forbidding Xeroxing of copyrighted material and withdrawing _almost_ entirely fair use rights has been approved on Sept 8, 2000, under pressure from editors and publishers of academic material (mainly located in the US).
I authored half a dozen scientific articles (Physical Review A, Physical Review Letters, Zeitschrift fur Physik...) and had to transfer my copyright to the publisher in order to get them published at all (academic press is even worse than the RIAA). According to the new law I - the author - can't legally make additional copies for my own private use anymore. Talk about screwed... anyway this is the typical way in which copyright works: it protects the rights of the publisher, certainly not those of the author.
On the other hand I relocated in Switzerland where fair use still holds. And stopped writing scientific articles.
Well said, but...
./ readers might want to build a two dimensional grid with
>In essence, every note in an equal tempered scale is a wrong note!
With the exception of the octaves, which are exact (that's the whole point).
On a C tempered scale, B is the worst approximated note, a fact that is quite easy to hear.
Western music is based on thirds and fifths: the more mathematically oriented among
powers of 3/2 on one axis and of 5/3 on the other one to see what happens and why the equal tempered tuning was introduced.
People who know about the continuous fraction expansion of real numbers may also find out why 12 semitones were used by expanding log(3/2) and log(5/3) [well, the historical reason has little to do with continuous fractions, which merely provide an explanation]. IIRC (I am probably slightly off) the next best approximation are at 42 (good third and bad fifth) and 54 subdivisions, both are quite unpractical.
>You really didn't answer my one question, which is where is WordPerfect's file format located on Corel's website?
.corel.com/partners_developers/ds/CO32SDK/docs/ff/ A_FRNTFF.htm
Here:
http://www
--
A Faraday Cage around the plane would only shield it plane from EM radiation coming from outside the plane itself: it would be of no help in preventing interference from sources inside the plane.
A Faraday Cage around all EM sensitive instrumentation would be unpractical and possibily unfeasible (since it would also shield radiation meant to be picked up - at least in part).
Do you think you are joking? See below.
http://www.nec.co.jp/engl ish/today/newsrel/0005/3001.html
BTW, the NEC SX-5 - unlike massively parallel architectures - can effectively run near its theoretical peak performance for most applications, I'd say that the top 40 TFlops performance is a rather conservative estimate (NEC will have newer and faster technology by the time this beast starts being built).
>So then how does Microsoft not fit into that?
Insofar as it's preventing others from achieving the same goal no matter what.
I'm quite sure you are not, but deep inside I still dimly hope you are joking.
Unless I'm gravely mistaken notepad.exe is the only test editor shipped by default with Windows.
Wordpad and Word are word processors and using them as text editors is at best an unnecessary complication - if it is at all possible.
IE is a completely different beast. I thought that confusing text editors and word processors was bad enough... but web browsers?
I see no reason why this different licensing scheme could not and should not coexist. The GPL is not taking away any rights: just providing some further rights under a somewhat - often very - restrictive model.
I think that the plethora of available licensing schemes provides more freedom for programmers; anyway nobody is forced to release anything under the GPL unless he's basing his work on GPL'd code to start with.
I am slightly in favor of the BSD scheme vs. the GPL one, but I basically think each programmer should decide for herself.
Following my istinct I would release my code under the public domain, or with no restriction whatsoever: but I've seen code released like this being stolen and copyrighted by somebody else, leaving the original author with no rights to it; that's why licensing schemes such as GPL, BSD or Artistic exist in the first place.
>As the VC community realizes there is little to no money to be made in GPL software, the spigot will close.
I doubt it. It has been around quite a while, and it is expanding.
>as in freedom, as opposed to socialism
You might not realize it, but this is quite an inflamatory remark (to Western European ears, at least). Be more careful. [I, FWIW, happen to hold the exact opposite point of view]
--