I seem to have pissed off APK. It started, actually, with someone else’s post.
This is his modus operandi when somebody doesn’t quickly fold to his trolling, and I actually have a bit of history with him in past similar encounters so he’s a bit short-fused now.
The resolution at which artifacts are negligible is so much higher that you'd waste more processing power on rendering without antialiasing at that resolution than you'd need with antialiasing at a resolution which is just high enough that you can't resolve individual pixels anymore
If you can’t resolve individual pixels, your eyes are already doing anti-aliasing in that multiple pixels are triggering the same receptor cell in your eye, which averages them out.
Even if the display pixels are smaller than you can resolve, a non-antialiasing renderer will still produce visible aliasing. Aliasing is a frequency domain effect.
Photons are discrete elements, but they’re small enough that we don’t have the problems you describe with discrete sampling of a continuous function.
Ideally (mathematically) you want to be working with elementary points. No anti-aliasing is necessary at that level, and if your physical approximation is close enough to elementary points anti-aliasing also becomes unnecessary.
The reason nerds are so confused by it is that everyone knows that if something is 100% bigger than something else, it follows that the smaller thing is 50% smaller than the bigger thing.
That depends on whether you are thinking of pixels as rectangles of finite size (which should be partly one colour and partly another), or as individual dimensionless points of colour.
The “mathematically correct way” is to think of them as points.
A LCD uses two polarizing filters. One of them “flips” its polarization 90 degrees when you apply a current.
Depending on the current, the polarizing filters can either be lined up (0 degrees) or perpendicularly aligned (90 degrees), or anywhere in-between.
When the polarizing filters are lined up, the backlight shines through (or the ambient light from the room is reflected off of the back of the display. When the polarizing filters are perpendicular, the pixel is black.
The color itself is created by a normal filter; individual red, green, and blue sub-pixels are used to create any RGB value.
If A is 8 times larger than B, it follows that B is 8 times smaller than A. That’s the only interpretation that makes any sense at all. So of course 8 times smaller means 1/8th the size.
Matter is harder to copy, whereas if you’re trying to sell data a new copy is made every time someone sees it.
Just imagine how different car sales would be if everyone who looked at your cars instantly had one of their own. The whole sales model would change – you’d have to show them only a tiny part of the car, or a car with no engine, or a car with a hidden self-destruct feature...
I see where he's going -- the pirated version sucks so if you want the real deal you gotta come to me -- but then that's basically the old crippleware model, only letting the pirates do the crippling for you.
When you "previewed" Photoshop, you got the full experience.
But then it also has the 30-day free trial, presumably un-crippled, if you get it from him.
Windows (and its volume control utility) simply isn’t supposed to work with focus-follows-mouse-without-autoraise. Windows didn’t break its interface’s usability, your hack broke the default Win7 volume utility.
Suggestion: You could, of course, stop using the hack, or you could look around and see if you can find a volume utility that replaces the default Win7 one and doesn’t break when you turn on focus-follows-mouse. For instance, a quick search turned up this. Caveat: I haven’t tried it... it might have the same issue. However, it’s open source, so if it’s broken in the same way someone ought to be able to make it work. Simply closing after x milliseconds instead of immediately when it loses focus would probably fix the problem, if it exists.
Well, the phrase "...if I do say so myself" is supposed to serve as a qualifier for the preceding phrase. It points out that the preceding statement is probably based on some bias. That is the entire point of "IIDSSM".
That may or may not be how it is used, but its literal meaning is circular: “...and if I said it, it’s true” (or “the preceding statement is true because I said it” – or the favourite of parents: “because I said so!”). Which, ironically, is an example of begging the question.
It’s better to qualify biased opinions with “but that’s just my opinion”, IMHO... YMMV, etc.
He's stating that in order to perform a rotate you need to use inline asm (he only said assembly language though, but this is doable in C/C++, &/or Delphi for example via the inline asm directive) to do so, and he's correct that it would break C's "intelligent assembler/cross platform code independence" in doing so
No, you wouldn’t, retard. If you know the size of a data word is WORD_SIZE = 32 bits, a bitwise rotate right is simply:
a = (a >> 1) | ((a & 1) << (WORD_SIZE - 1));
And a bitwise rotate left is:
a = (a << 1) | (a >> (WORD_SIZE - 1));
This is perfectly extensible to any platform with only the adjustment of the definition of the WORD_SIZE constant – which, as I said, you would need to know anyway because the results are hardly meaningful if you rotate an element of an unknown size.
But hey... naturally you wouldn’t know this, because your idea of “programming” is building front-ends in Delphi or VB for system functions like “move” or “copy”. Hardly a step above batch programming, really.
Slashdot Mirror
They got Flash to work on the iPad?
And just in case they do delete it:
http://omploader.org/vNWRveg/ipawn.gif
Fake indeed! They took down the incriminating video, but forgot to delete it. Get it before it’s gone.
The last choice (“jeux vidéo”) clearly selects itself a moment before the token is actually set down onto it in the first video on their blog.
I seem to have pissed off APK. It started, actually, with someone else’s post.
This is his modus operandi when somebody doesn’t quickly fold to his trolling, and I actually have a bit of history with him in past similar encounters so he’s a bit short-fused now.
The solution to banding is more bits per pixel, not dithering.
But smaller, and they only need one polarizing filter because the silts already produce polarized light.
Yeah, that whole dithering thing never really caught on did it? Going from 16 colours to 16,777,216 hasn’t removed the need for it at all.
The resolution at which artifacts are negligible is so much higher that you'd waste more processing power on rendering without antialiasing at that resolution than you'd need with antialiasing at a resolution which is just high enough that you can't resolve individual pixels anymore
If you can’t resolve individual pixels, your eyes are already doing anti-aliasing in that multiple pixels are triggering the same receptor cell in your eye, which averages them out.
Even if the display pixels are smaller than you can resolve, a non-antialiasing renderer will still produce visible aliasing. Aliasing is a frequency domain effect.
Photons are discrete elements, but they’re small enough that we don’t have the problems you describe with discrete sampling of a continuous function.
Ideally (mathematically) you want to be working with elementary points. No anti-aliasing is necessary at that level, and if your physical approximation is close enough to elementary points anti-aliasing also becomes unnecessary.
I’m currently driving in my free Mozilla car with a number of also-free mod kits intalled. You were saying?
No, that’s exactly correct.
The reason nerds are so confused by it is that everyone knows that if something is 100% bigger than something else, it follows that the smaller thing is 50% smaller than the bigger thing.
That depends on whether you are thinking of pixels as rectangles of finite size (which should be partly one colour and partly another), or as individual dimensionless points of colour.
The “mathematically correct way” is to think of them as points.
The article actually said how they’d create color: red, green, and blue sub-pixels, just like LCDs currently use.
Red, green and blue pixel components could be made in one step by cutting arrays of slits in the stack, according to the researchers.
A LCD uses two polarizing filters. One of them “flips” its polarization 90 degrees when you apply a current.
Depending on the current, the polarizing filters can either be lined up (0 degrees) or perpendicularly aligned (90 degrees), or anywhere in-between.
When the polarizing filters are lined up, the backlight shines through (or the ambient light from the room is reflected off of the back of the display. When the polarizing filters are perpendicular, the pixel is black.
The color itself is created by a normal filter; individual red, green, and blue sub-pixels are used to create any RGB value.
iPhone pixels: ~78 microns
These (from the legend in the picture in TFA): ~10 microns
The point is it’s a reciprocal relation.
If A is 8 times larger than B, it follows that B is 8 times smaller than A. That’s the only interpretation that makes any sense at all. So of course 8 times smaller means 1/8th the size.
No, that’d be 8 times larger.
Um, does this answer your question:
It's a fucking SINGLE BOOLEAN SWITCH that validates the license, it doesn't require any Application Enhancer tricks or whatever.
Ok, ok, so he didn’t quite release a crack, but it’s practically a road-map.
Matter is harder to copy, whereas if you’re trying to sell data a new copy is made every time someone sees it.
Just imagine how different car sales would be if everyone who looked at your cars instantly had one of their own. The whole sales model would change – you’d have to show them only a tiny part of the car, or a car with no engine, or a car with a hidden self-destruct feature...
I see where he's going -- the pirated version sucks so if you want the real deal you gotta come to me -- but then that's basically the old crippleware model, only letting the pirates do the crippling for you.
When you "previewed" Photoshop, you got the full experience.
But then it also has the 30-day free trial, presumably un-crippled, if you get it from him.
Actually, I was used to right-click, “C”. Which they broke.
Windows (and its volume control utility) simply isn’t supposed to work with focus-follows-mouse-without-autoraise. Windows didn’t break its interface’s usability, your hack broke the default Win7 volume utility.
Suggestion: You could, of course, stop using the hack, or you could look around and see if you can find a volume utility that replaces the default Win7 one and doesn’t break when you turn on focus-follows-mouse. For instance, a quick search turned up this. Caveat: I haven’t tried it... it might have the same issue. However, it’s open source, so if it’s broken in the same way someone ought to be able to make it work. Simply closing after x milliseconds instead of immediately when it loses focus would probably fix the problem, if it exists.
Well, the phrase "...if I do say so myself" is supposed to serve as a qualifier for the preceding phrase. It points out that the preceding statement is probably based on some bias. That is the entire point of "IIDSSM".
That may or may not be how it is used, but its literal meaning is circular: “...and if I said it, it’s true” (or “the preceding statement is true because I said it” – or the favourite of parents: “because I said so!”). Which, ironically, is an example of begging the question.
It’s better to qualify biased opinions with “but that’s just my opinion”, IMHO... YMMV, etc.
He's stating that in order to perform a rotate you need to use inline asm (he only said assembly language though, but this is doable in C/C++, &/or Delphi for example via the inline asm directive) to do so, and he's correct that it would break C's "intelligent assembler/cross platform code independence" in doing so
No, you wouldn’t, retard. If you know the size of a data word is WORD_SIZE = 32 bits, a bitwise rotate right is simply:
a = (a >> 1) | ((a & 1) << (WORD_SIZE - 1));
And a bitwise rotate left is:
a = (a << 1) | (a >> (WORD_SIZE - 1));
This is perfectly extensible to any platform with only the adjustment of the definition of the WORD_SIZE constant – which, as I said, you would need to know anyway because the results are hardly meaningful if you rotate an element of an unknown size.
But hey... naturally you wouldn’t know this, because your idea of “programming” is building front-ends in Delphi or VB for system functions like “move” or “copy”. Hardly a step above batch programming, really.