And also, speed is really important if you play games.
One month ago, I ran Super Mario Bros. 3 at full speed on an NES emulator running on a Pentium 100 computer owned by a school. My current sub-GHz machine runs Game Boy Advance games at full speed. Games != bleeding-edge 3D games.
for example, builds are often CPU-bound, so [build speed will scale linearly with CPU clock]
Wrong. Building a large project is highly disk I/O bound. Normally, GCC on my PIII 866 MHz compiles the one source file I have changed within about two seconds. Linking takes the most time because it has to retrieve dozens of.o and.a files to produce a.exe file.
Only the builds on "Clobber" tinderboxes, where the system does a "make clean" before rebuilding the software, are CPU-bound. Builds on customers' machines are CPU-bound, but they can run while the client is reading some web comic (I/O and user bound).
All my NES, Super NES, Game Boy, and Game Boy Advance games run just fine on my old Dell with a PIII 866. Games != bleeding edge 3D games. Besides, nowadays, bleeding edge 3D performance depends more on the video card than on the x86 CPU.
Video compression/editing?
Without training, most budding directors will make crap even worse than that movie Dana Carvey just starred in. If I had enough money for a cinematography class, I would probably also have enough money for a box with the new processor in it.
this new cpu allow to people to convert their CD to MP3 even faster!
MP3 and Ogg encoding at archival quality already run faster than realtime on my old PIII 866 MHz. If the encoder can compress the audio faster than the CD drive can read it reliably[1], then what's the point of being able to encode faster?
[1] CD drives typically rip audio CDs slower than data CDs because Red Book audio carries less error correction coding than data or MP3 audio.
But do you even really need the encoding to run faster? I typically put a few CDs into CDex and then encode them in the background while I read/.
That's the problem. You appear to just be spouting out your own technique. You're not actually a font designer, are you?
I'll confess. I've only designed bitmap fonts for video games. I've written tools to help me do that, and the techniques I have described match what my tools do.
Fact: In the United States, you cannot copyright a bitmap font. However, you can copyright a program (in the Metafont, PostScript, or TrueType hinting language) that generates a font. To circumvent this, start by running an autotracer on a large L/G/C Unicode chart written in that font, but note that you do lose kernpairs in the process.
Letter forms are not generated algorithmically. A "b" is not an upside-down "p."
I knew that, and that's why I didn't suggest combining "b" and "p". However, in every roman (serif or sans) or oblique sans font I've seen, a "" is the top of a "b" and the bottom of a "p" stuck together.
These things have to be drawn by hand by people with talent-- more talent than I have, to be sure-- and refined at a level of detail that you wouldn't believe.
And that's why I explicitly stated that each generated glyph would be presented to the font designer for tweaking: "when creating a new glyph, copy parts from similar glyphs and present them to the font designer for further work" and "have the font designer stitch them together by editing the splines." (I should have said "editing the splines by hand.")
427 hours work of work to be done in 3 weeks, how many compentent programmers do you need
This question was asked on Slashdot's servers in the United States, and United States law states that a full-time week equals 40 hours. Thus, one pair of programmers can do 120 hours of work in the alloted time. Let n = the number of pairs of programmers. Then 427h = 120h * n, or 3.56 = n. Round up to four pairs of programmers, and you can give them nearly an hour of free time to unwind at the end of the day (Quack 3: The Duck Mod).
From what I know of these countries, it seems extremely unfair to lump Cuba in with Iraq et al as "Axis of Evil".
Don't blame me. Blame the people who write the export regulations. Blame USA Government. Blame PoizonBOx. The USA Government has considered Cuba a state sponsor of terrorism long before September 2001.
You'll see that the bar of the f is connected to the i, and that the dot above the i is absent.
In my technique, the program would copy parts from 'f' and 'Turkish dotless i' glyphs and have the font designer stitch them together by editing the splines.
Perhaps also a link to a program that fits your description of a "good font editor?"
There is none that I know of; pfaedit has the most potential (because it's free software, and people who know both C and font design can add features that scratch their itches), but it's definitely not 'there' yet.
People don't believe that the Internet is free and always will be. What they believe is that they should be able to pay a reasonable price to an ISP for access to a worldwide network.
So they have an access to a network. But what use is a network with nothing good on it? Writers have to eat somehow. That's the whole cable TV model: pay the cable company for access to the network; watch the ads for access to the content.
Before anyone here says that fonts are easy to make, you're probably forgetting the non-western character sets and the thousands of unicode characters.
Just as there are fonts that specialize in "CJK" (Chinese Japanese Korean) glyphs, there can also be "LGC" fonts for Latin, Greek, and Cyrillic. A good font editor will have the user draw a bunch of glyphs representing A-Z in Latin, the Greek, Cyrillic, and IPA glyphs that do not match the Latin glyphs, and then some diacritics. Then from that data, it'll "compose" glyphs for the first 1500 or so characters in Unicode.
Another optimization: when creating a new glyph, copy parts from similar glyphs and present them to the font designer for further work. For example, from b and p, you get (thorn). From D, you get Ð (edh). From l and n, you get h. From n, you can infer most of m. From f, you get long s, and from long s and normal s, you get ß.
If you post it on the net, then there's nothing stopping [Axis of Evil] countries from getting the code. Hence, doesn't that put one in murky legal waters?
Not under the definition of "knowingly" used by BIS. Did you read the regulations I linked to? 740.13(e)(6) clearly states (my emphasis):
Posting of source code or corresponding object code on the Internet (e.g., FTP or World Wide Web site) where it may be downloaded by anyone
would not establish "knowledge" of a prohibited export or reexport.
And if you really want to cover your rectum, you can make a "best effort" by looking up the IPv4 address ranges for the popular ISPs in the Axis of Evil, and just firewall those off.
Well, the U.S. government does place restrictions on one's right to give software away (in the case of strong cryptography). Hence OpenBSD is based in Canada.
But do these U.S. export restrictions apply to free software? The current crypto export regulations (section 740.13(e)) seem to grant an export License Exception for publicly available source code and object code compiled from publicly available source code provided that the original publisher of such code notifies crypt@bis.doc.gov (cc: enc@ncsc.mil) of the code's public availability. (Notification seems not to be required for mirrors.)
Hence Mozilla is based in the United States, where the only restriction on exporting OSI Certified(tm) open source encryption software is that it not implement a system primarily designed to restrict the fair use of a copyrighted work.
find DeCss, just like a google search or searching Disney's Go.
As soon as Disney found out that people were using its GO.com Infoseek search engine to look for DeCSS, which could decrypt Disney DVD titles, Disney killed Infoseek and went with Overture.
Hmmm, who the fsck would record a dvd to vhs anyways.
Parents of small children would, to avoid damaging their copy of "Adventures of Pinocchio" that the kids watch every night. Keep the purchased DVD copy as a backup and let the kids dest^H^H^H^Hwatch a copy on a $2 VHS tape. The Supreme Court has maintained that this is a fair use.
That is, until Congress enacted a bill that created 17 USC 1201, which gives publishers the right to outlaw fair use.
No, I did not miss the Fermat joke. Note that I accused dave_mcmillen of making a "bad joke". I've just noticed that Fermat jokes in Slashdot pure-math articles are quickly becoming as tired as Beowulf cluster jokes in hardware articles.
I have assuredly found an admirable proof of this, but the Post Comment box is too narrow to contain it.
I suspect that you're either bluffing or just making a bad joke analogous to the DMCA jokes, the Beowulf jokes, the "All Your Base" jokes, the "2. ???; 3. PROFIT!" jokes, the "Priceless" jokes, etc. that get moderated to -1 on Slashdot.
On the other hand, if you really do have a proof or disproof of the existence of Lychrel numbers, then fire up Emacs, make a web page outlining your proof, and post it on a public server. Go to GeoShitties if you have to. Leave the finest details to the reader if you have to. Pure mathematicians want to see it. Bad.
This is not pure mathematics, as it relies on the numbers being represented in a certain base.
Number theory per se is considered part of pure mathematics, the N in NIGGERS. When my number theory book defined some basic concepts in terms of first principles, it defined "base" soon after it defined "multiplication". It roughly went like this: The base b representation of the positive integer n is a finite sequence of non-negative integers less than b such that the last element x > 0 and the sum[i = 0..n-1](x[i]*b^i) = n.
This statement about 196 and base 10 is a statement about sequences of numbers.
X-Box: 120 millon polygons/sec. GCN: 6-12 million polygons/sec. Guess which has better graphics! You guessed right: GCN
That's because Nintendo measures performance with lit textured triangles of a decent size, rather than one-pixel flat shaded triangles. An XBox developer attested to this: one of the tech demos included in the XDK uses one-pixel triangles to render a particle system and does manage to break 1.7 million triangles per frame (100 million triangle per second), but this situation won't occur very often in a real game.
The GameCube has a comb filter, which reduces flicker and covers up non-antialiased edges on NTSC 480i, PAL/M 480i, and PAL 576i displays, which cover the vast majority of displays used for game consoles in Japan, the Americas, and Europe.
But the real reason GameCube games look better may not be because of technical ability but because Nintendo has Mr. Shigeru Miyamoto producing exclusive titles.
MGhz for Mghz a RISC chip kicked the shit out of CISC and stole their lunch money. If I'm not mistaken, they still do.
Not especially. Modern CISC CPUs such as the Athlon, the P4, and the Crusoe recompile CISC bytecode into RISC micro-operations internally. The problem with the P4 is that the decoder isn't fast enough (one micro-op per clock for non-cached instructions; three micro-ops per clock for cached instructions) to feed the P4's nine functional units.
At work, I can rip a CD with CDex in about 16-18 minutes per disk using the SSE enabled Lame encoder. On my laptop at home, it takes less than 5 minutes to rip a CD with iTunes. What gives?
How much of that is the physical speed of the CD-ROM drive? My PlexWriter 12/10/32A burner reads data at 10x to 32x (CAV) but reads audio at 10x across the whole CD, limiting me to an 8-minute rip.
But still, I've never understood how people can just rip and encode to MP3 simultaneously. Without an intermediate step where the recording exists as a wav file, there's no chance to fix up pops in the audio, silence explicit language for a play-in-front-of-your-parents edit (I'd rather not pay twice for the clean and dirty versions), or remove leading or trailing silence.
Slashdot Mirror
How does Combat look on these babies?
Mortal or non-Mortal?
And also, speed is really important if you play games.
One month ago, I ran Super Mario Bros. 3 at full speed on an NES emulator running on a Pentium 100 computer owned by a school. My current sub-GHz machine runs Game Boy Advance games at full speed. Games != bleeding-edge 3D games.
for example, builds are often CPU-bound, so [build speed will scale linearly with CPU clock]
Wrong. Building a large project is highly disk I/O bound. Normally, GCC on my PIII 866 MHz compiles the one source file I have changed within about two seconds. Linking takes the most time because it has to retrieve dozens of .o and .a files to produce a .exe file.
Only the builds on "Clobber" tinderboxes, where the system does a "make clean" before rebuilding the software, are CPU-bound. Builds on customers' machines are CPU-bound, but they can run while the client is reading some web comic (I/O and user bound).
Games?
All my NES, Super NES, Game Boy, and Game Boy Advance games run just fine on my old Dell with a PIII 866. Games != bleeding edge 3D games. Besides, nowadays, bleeding edge 3D performance depends more on the video card than on the x86 CPU.
Video compression/editing?
Without training, most budding directors will make crap even worse than that movie Dana Carvey just starred in. If I had enough money for a cinematography class, I would probably also have enough money for a box with the new processor in it.
this new cpu allow to people to convert their CD to MP3 even faster!
MP3 and Ogg encoding at archival quality already run faster than realtime on my old PIII 866 MHz. If the encoder can compress the audio faster than the CD drive can read it reliably[1], then what's the point of being able to encode faster?
[1] CD drives typically rip audio CDs slower than data CDs because Red Book audio carries less error correction coding than data or MP3 audio.
But do you even really need the encoding to run faster? I typically put a few CDs into CDex and then encode them in the background while I read /.
Oh, and writing the name of my company with a dollar sign instead of an 's' is childish.
Didn't you write a BASIC interpreter for the Altair and the Apple II? If so, isn't this correct BASIC?
In old versions of BASIC (i.e. before QuickBASIC), names of string variables ended in a $ character. M$ is shorter to type than "Microsoft".
That's the problem. You appear to just be spouting out your own technique. You're not actually a font designer, are you?
I'll confess. I've only designed bitmap fonts for video games. I've written tools to help me do that, and the techniques I have described match what my tools do.
Fact: In the United States, you cannot copyright a bitmap font. However, you can copyright a program (in the Metafont, PostScript, or TrueType hinting language) that generates a font. To circumvent this, start by running an autotracer on a large L/G/C Unicode chart written in that font, but note that you do lose kernpairs in the process.
Letter forms are not generated algorithmically. A "b" is not an upside-down "p."
I knew that, and that's why I didn't suggest combining "b" and "p". However, in every roman (serif or sans) or oblique sans font I've seen, a "" is the top of a "b" and the bottom of a "p" stuck together.
These things have to be drawn by hand by people with talent-- more talent than I have, to be sure-- and refined at a level of detail that you wouldn't believe.
And that's why I explicitly stated that each generated glyph would be presented to the font designer for tweaking: "when creating a new glyph, copy parts from similar glyphs and present them to the font designer for further work" and "have the font designer stitch them together by editing the splines." (I should have said "editing the splines by hand.")
427 hours work of work to be done in 3 weeks, how many compentent programmers do you need
This question was asked on Slashdot's servers in the United States, and United States law states that a full-time week equals 40 hours. Thus, one pair of programmers can do 120 hours of work in the alloted time. Let n = the number of pairs of programmers. Then 427h = 120h * n, or 3.56 = n. Round up to four pairs of programmers, and you can give them nearly an hour of free time to unwind at the end of the day (Quack 3: The Duck Mod).
From what I know of these countries, it seems extremely unfair to lump Cuba in with Iraq et al as "Axis of Evil".
Don't blame me. Blame the people who write the export regulations. Blame USA Government. Blame PoizonBOx. The USA Government has considered Cuba a state sponsor of terrorism long before September 2001.
Imagine a system where the government is so intrusive that it even controls the amount of food which one is entitled to live off.
You'll see that the bar of the f is connected to the i, and that the dot above the i is absent.
In my technique, the program would copy parts from 'f' and 'Turkish dotless i' glyphs and have the font designer stitch them together by editing the splines.
Perhaps also a link to a program that fits your description of a "good font editor?"
There is none that I know of; pfaedit has the most potential (because it's free software, and people who know both C and font design can add features that scratch their itches), but it's definitely not 'there' yet.
People don't believe that the Internet is free and always will be. What they believe is that they should be able to pay a reasonable price to an ISP for access to a worldwide network.
So they have an access to a network. But what use is a network with nothing good on it? Writers have to eat somehow. That's the whole cable TV model: pay the cable company for access to the network; watch the ads for access to the content.
Before anyone here says that fonts are easy to make, you're probably forgetting the non-western character sets and the thousands of unicode characters.
Just as there are fonts that specialize in "CJK" (Chinese Japanese Korean) glyphs, there can also be "LGC" fonts for Latin, Greek, and Cyrillic. A good font editor will have the user draw a bunch of glyphs representing A-Z in Latin, the Greek, Cyrillic, and IPA glyphs that do not match the Latin glyphs, and then some diacritics. Then from that data, it'll "compose" glyphs for the first 1500 or so characters in Unicode.
Another optimization: when creating a new glyph, copy parts from similar glyphs and present them to the font designer for further work. For example, from b and p, you get (thorn). From D, you get Ð (edh). From l and n, you get h. From n, you can infer most of m. From f, you get long s, and from long s and normal s, you get ß.
If you post it on the net, then there's nothing stopping [Axis of Evil] countries from getting the code. Hence, doesn't that put one in murky legal waters?
Not under the definition of "knowingly" used by BIS. Did you read the regulations I linked to? 740.13(e)(6) clearly states (my emphasis):
And if you really want to cover your rectum, you can make a "best effort" by looking up the IPv4 address ranges for the popular ISPs in the Axis of Evil, and just firewall those off.
Well, the U.S. government does place restrictions on one's right to give software away (in the case of strong cryptography). Hence OpenBSD is based in Canada.
But do these U.S. export restrictions apply to free software? The current crypto export regulations (section 740.13(e)) seem to grant an export License Exception for publicly available source code and object code compiled from publicly available source code provided that the original publisher of such code notifies crypt@bis.doc.gov (cc: enc@ncsc.mil) of the code's public availability. (Notification seems not to be required for mirrors.)
Hence Mozilla is based in the United States, where the only restriction on exporting OSI Certified(tm) open source encryption software is that it not implement a system primarily designed to restrict the fair use of a copyrighted work.
find DeCss, just like a google search or searching Disney's Go.
As soon as Disney found out that people were using its GO.com Infoseek search engine to look for DeCSS, which could decrypt Disney DVD titles, Disney killed Infoseek and went with Overture.
It didn't help. Overture takes you straight to DeCSS as well, in both flavors (DVD and HTML).
Yes, but what region is space ? I would think it would be region 0 - aka region-free.
Special international venues such as air and space are region 8.
Hmmm, who the fsck would record a dvd to vhs anyways.
Parents of small children would, to avoid damaging their copy of "Adventures of Pinocchio" that the kids watch every night. Keep the purchased DVD copy as a backup and let the kids dest^H^H^H^Hwatch a copy on a $2 VHS tape. The Supreme Court has maintained that this is a fair use.
That is, until Congress enacted a bill that created 17 USC 1201, which gives publishers the right to outlaw fair use.
the joke you so obviously missed
No, I did not miss the Fermat joke. Note that I accused dave_mcmillen of making a "bad joke". I've just noticed that Fermat jokes in Slashdot pure-math articles are quickly becoming as tired as Beowulf cluster jokes in hardware articles.
I have assuredly found an admirable proof of this, but the Post Comment box is too narrow to contain it.
I suspect that you're either bluffing or just making a bad joke analogous to the DMCA jokes, the Beowulf jokes, the "All Your Base" jokes, the "2. ???; 3. PROFIT!" jokes, the "Priceless" jokes, etc. that get moderated to -1 on Slashdot.
On the other hand, if you really do have a proof or disproof of the existence of Lychrel numbers, then fire up Emacs, make a web page outlining your proof, and post it on a public server. Go to GeoShitties if you have to. Leave the finest details to the reader if you have to. Pure mathematicians want to see it. Bad.
This is not pure mathematics, as it relies on the numbers being represented in a certain base.
Number theory per se is considered part of pure mathematics, the N in NIGGERS. When my number theory book defined some basic concepts in terms of first principles, it defined "base" soon after it defined "multiplication". It roughly went like this: The base b representation of the positive integer n is a finite sequence of non-negative integers less than b such that the last element x > 0 and the sum[i = 0..n-1](x[i]*b^i) = n.
This statement about 196 and base 10 is a statement about sequences of numbers.
X-Box: 120 millon polygons/sec. GCN: 6-12 million polygons/sec. Guess which has better graphics! You guessed right: GCN
That's because Nintendo measures performance with lit textured triangles of a decent size, rather than one-pixel flat shaded triangles. An XBox developer attested to this: one of the tech demos included in the XDK uses one-pixel triangles to render a particle system and does manage to break 1.7 million triangles per frame (100 million triangle per second), but this situation won't occur very often in a real game.
The GameCube has a comb filter, which reduces flicker and covers up non-antialiased edges on NTSC 480i, PAL/M 480i, and PAL 576i displays, which cover the vast majority of displays used for game consoles in Japan, the Americas, and Europe.
But the real reason GameCube games look better may not be because of technical ability but because Nintendo has Mr. Shigeru Miyamoto producing exclusive titles.
MGhz for Mghz a RISC chip kicked the shit out of CISC and stole their lunch money. If I'm not mistaken, they still do.
Not especially. Modern CISC CPUs such as the Athlon, the P4, and the Crusoe recompile CISC bytecode into RISC micro-operations internally. The problem with the P4 is that the decoder isn't fast enough (one micro-op per clock for non-cached instructions; three micro-ops per clock for cached instructions) to feed the P4's nine functional units.
At work, I can rip a CD with CDex in about 16-18 minutes per disk using the SSE enabled Lame encoder. On my laptop at home, it takes less than 5 minutes to rip a CD with iTunes. What gives?
How much of that is the physical speed of the CD-ROM drive? My PlexWriter 12/10/32A burner reads data at 10x to 32x (CAV) but reads audio at 10x across the whole CD, limiting me to an 8-minute rip.
But still, I've never understood how people can just rip and encode to MP3 simultaneously. Without an intermediate step where the recording exists as a wav file, there's no chance to fix up pops in the audio, silence explicit language for a play-in-front-of-your-parents edit (I'd rather not pay twice for the clean and dirty versions), or remove leading or trailing silence.