I was more reacting to the use of "FUD" whenever anybody does any comparison of themselves to others. Mostly I see if from Microsoft defenders. If somebody says "Linux runs 10 times faster than Windows" they will say "that's FUD" when in fact what they mean is "that's a lie". There is a difference, in my opinion, though I never see anybody trying to correct them.
1) High-contrast images such as credits can indeed be "projected at" the screen plane to prevent ghosting... but high-contrast images at any other plane will always have some ghosting.
True. However with this system at least the screen plane is guaranteed to not have ghosting. With two projectors that is not going to happen unless somebody carefully aligns them. And in the typical theatre that is not going to happen, or it will soon be bumped and knocked out of alignment.
mis-sync between the display of supposedly simultaneous left- and right-images
Yes I would agree that is certainly going to happen. Possibly a solution is to purposely render one of the images to be correct for 1/144 of a second later than the other one? Similar solutions are done for interlaced fields in TV, and I know this does not always work that well, so this may be an unsolvable problem.
"FUD" stands for "fear, uncertainty, and doubt". Apparently somehow it has had it's meaning twisted around here to mean "lies" about any subject. FUD can be entirely truthful, just worded to make people want to take the safer course.
I mostly see this here if anybody says anything negative about Microsoft, they are accused of "FUD". The term is wrong, except in a few cases such as when people give warnings like "DRM will destroy free speech" or that "Microsoft will bury/discontinue/etc that product, don't buy into it". These statements could be true or false, but they are FUD in that they are trying to scare people into disliking Microsoft or doing something Microsoft does not want, by predicting bad things to happen in the future.
Saying "Microsoft sucks" or "Linux is more than 10 times better than Windows" is NOT "FUD". It may be a false statement, but it is trying to make people choose Linux over Microsoft for positive reasons.
Conversely if Microsoft says their software is superior or has higher up time or lower tco, that is NOT "FUD". That may be truth or lies, but it is still not FUD. It is normal advertising, saying your product is better. They certainly do a lot more FUD than anybody else, if they say Linux might subject you to patent lawsuits, or copyright violations, or force you to give all your source code away, that is FUD. It can be true or false, it does not matter, what matters is that it is a vague threat of something that might happen in the future unless you do what they want.
I tried to make up something that Greenpeace would say that really is FUD but could not really come up with an example.
Does anybody agree with me, or is the use of "FUD" been completely twisted to mean almost any statement about a competitor?
There are already versions of this on Linux. One I use at work does it instead by making a new directory with the older versions of the files in it. To see yesterdays version of./foo.txt you look at./.snapshots/yesterday/foo.txt. This seems a lot nicer as you can more easily see when files are created and deleted as well as modified, though it is possible that technical limitations prevented this from using this naming scheme. I not sure what the system is, as it is a large file server, it is running Linux, though the filesystem is possibly proprietary.
I think the big deal is that this is a modification of the ext3 filesystem, rather than an all-new filesystem.
That must get real real hot with that bright light what is it, 7000-12000 watts going through a 2 inch space, must require lots of cold air being blown way fast for that to work.
The drawing in the article seems to show that the lcd shutter is more like the size of a large windowpane and is some distance in front of the projector. I would think this is on purpose to solve the heat problems. May also be necessary so it can be added without modifying the DLP.
THis cannot be retrofitted on current DLPs, no way.... your source media is in a fixed format, either 24fps film, or 24fps MPEG2/MPEG4 data. At most it would be 60/50fps DLP. To get 144/150
No, the source media is a computer that is just like the game machines that can do 150 or more frames per second. It may be reading two 24 fps streams for the left and right eye and displaying each image 3 times. However I agree with you that it seems doubtful that a typical DLP projector for a theatre, designed to show 24 fps source images, is going to handle a 144 fps signal.
The article is a little vague about how it works, trying to make it sound more magical.
What the system does is alternate projections of the left and right eye images using the same DLP projector. They said 144 frames per second, which I think means that each film frame (of which there are 24 per second) is projected 3 times for each eye, this means each eye sees the image flickering at 72 times a second, which is above the threshold for most people to see flickering. The real technology is a special lcd screen that is put in front of the lens of the projector that changes it's polarization 144 times per second so each image is polarized differently.
The real advantage of this is that the same DLP projector used for non-3D films can be reused, just put the lcd in front of the lens when showing 3D. Any other system would require a second projector, which not only adds the cost of the projector, but the cost to mount it and add another aperture in the theatre wall. (actually another system would be shutter glasses with lcd lenses that turn on/off so each eye sees one side, but handing each customer an item that costs 10 or more dollars is probably out of the question) Also this system allows perfect alignment so that things that should appear at the screen plane really appear there, and high-contrast things like the credits can be projected at that distance with no ghosting.
It does appear fortunate that they can run at 144 frames per second, though if they were like consumer ones with a maximum of 90 or 100 it would still be an acceptable flicker rate of 45 or 50 (classic film projectors flickered 48 times a second due to having 1 extra vane on the shutter).
It is brilliant! And it certainly is different than Microsoft's implementation and not covered by any patent.
In fact it is a good deal simpler, it seems to reduce the color fringing naturally without having the calculation for one pixel effect nearby ones. It also only requires 2x horizontal resolution of the original rendering, rather than 3x.
From discussion here it sounds like what Microsoft patented is a filter, similar to their "font smoothing" filter, that turns a 1x resolution image into the colored cleartype image. That would mean it covers very little, and will be as obsolete as font smoothing in a year. However I kind of doubt this, I was under the impression that they were rendering at 3x horizontal resolution and then using this filter to put that in the lcd pixels without color artifacts. Anybody know?
In either case, rendering at 2x resolution and not having to do any filtering makes this be excluded from the Patent. Also looks like the results are better for very tiny fonts, too, mostly because there are only two color fringes, not 3.
That was completely misleading. The use of 256 levels for making antialiased fonts is really old and has nothing to do with this.
What sub pixel rendering does is make the pixel represent *more* than 256 different possible combinations of the fg and bg colors, where combinations are how an fully opaque edge falls into the square the pixel represents. Exactly how many is unclear, it is not 256^3, but I think it is 3*256 for the case of an antialiased vertical straight edge of an object significantly larger than a pixel (it goes down as the edge gets less vertical or if another edge gets within a few pixels of it). This is done by using hardware quirks so that various of the 256^3 possible colors are interpreted by the viewer as different coverages.
The Apple technique made a pixel represent more than the 2 possible coverages you mentioned. This is done by using hardware quirks so that the some of the 4 (16? i dunno) possible colors are interpreted by the viewer as different coverages.
They certainly are related, but I don't believe enough to invalidate Microsoft's patent. The difference is that the "hardware quirks" are vastly different.
Copyright is what prevents you from doing things. This new GPL is reducing the number of ways it allows you to violate the copyright. So it is granting freedom, just less than before.
Actually I suspect any Linux binaries from back then work, as the shared libraries they used are still there, and complex stuff was static-linked into the program.
The problems are more in the middle-age binaries, where GUI libraries started to show up (and this is still true now, so today's binaries may not work on the future machines). Windows is also sufferring from the same problem, though they do tend to be less likely to throw away the old api (Linux has the "excuse" that you can download and install the library if you need it, but in reality this is equivalent to "the api is missing and the software does not work").
I would also expect that this attempt to increase the gravity would result in a planet with enough gravity, but also a molten surface, and no water or atmosphere.
Re:I can't feel any responsiveness improvements.
on
Gnome 2.18 Released
·
· Score: 1
Actually you are off the mark. Threading has nothing to do with it, neither Windows or Gnome are using multiple threads to do the GUI (it actually would be a nightmare of complexity and probably far slower).
The problem is simply that Gnome, instead of using 15 guys passing instructions to each other, uses more like 200. That menu item probably has 20 or so ways of being changed to the correct foreign language and correctly formatted in swahili or whatever because there are many, many guys interested in how to change it. Meanwhile, the Windows system maybe has 15 guys (maybe each somewhat fatter and slower) and at most only one of them is interested in changing to swahili. In both cases it is enormously easier for any application that wants to display in swahili to just do it itself (with the equivalent of adding one really blindingly fast guy, comparatively speaking), so most of this overhead is a complete waste. It also does not help Gnome at all that those 200 workers are much more likely to do some completely unpredictable thing (something that makes perfect sense to that worker), on Windows a programmer can be reasonably certain that the text they say to draw on the screen will come out with at least the english letters correct.
Any real intelligent system would have one guy. The menu item should be a char* pointer to utf8 string, and it goes straight to the "draw this text here on the screen call". GUI's were usable on systems that ran at 5mhz, you know. Therefore even Windows must have an unnecessary overhead of at least 500 or so, and Gnome is even worse.
You act as though there is no good reason to have Linux pre-installed. And then in the very next paragraph you say:
your average person isn't going to know what to do if their wireless card isn't working, or if they don't have support for MP3s, etc
Now are you deliberately being dense?
I thought I'd leave it at that, but just in case you cannot see what should be blatently obvious: the reason for pre-installed Linux is to solve the exact same problem you quoted.
What I was trying to say was "if you are using 16 bits for each pixel, it is a waste to store an integer in them, rather than using them for a floating point number".
A half float will have enormously higher resolution near 0 where the posterization shows. It can also represent numbers greater than 1, negative numbers, +/- infinity, and NaN. And it can efficiently represent the actual intensity of light, rather than a gamma curve that is required for integers to get the samples near zero closer together.
16 bit integers (or any number greater than 8) is a waste of time in image editing, and it is in some ways fortunate that Gimp avoids it, since it would take as much work to replace it as to add new data types.
If you are going to use more than 8 bits then using 16-bit "half" floats, or 32 bit floats, is a much better use of the space, and modern processors are easily able to keep up with this. It also helps a lot if the floating data represents linear light levels, rather than the log or gamma curve data that is in integer files.
Sounds like you are actually quite a genius. I can assure you from personal experience reading others and even my own code, it is far easier to write C# and C++ than to read them!
Actually you are supposed to use some rather obscure system call to get the "Documents and Settings" directory name, and another obscure call to get the "username". In addition the Desktop for the user is not under this directory. Certainly easier to do getenv("HOME").
In any case it is pretty obvious that Microsoft fixed this in Vista. It is now "/Users/username" and the desktop is this directory/Desktop. Oddly enough this is exactly the same as OS/X and the same as many Linux machines (some use "home" instead of "Users" but one symbolic link will fix that. Not sure if they put the directory name into an environment variable.
What I was shocked at was the amount of hoops they jumped through to avoid calling this directory "home" in the GUI. For instance, no icon and rather convoluted descriptions in the start menu, to avoid use of the word "home" or a house image. They are worried sick about being too compatable with Unix, no doubt.
They are using the 404 response correctly, this is what it was designed for! IE is trying to do something intellligent when it knows that the page is missing. What verisign did was fool every program (including IE) into thinking *all* pages exist, which breaks anything that wants to respond in a useful way to the page being mistyped.
I checked on a Windows machine, and they even let you change it! Didn't even bury it too deep in the configuration! You can go to google or bash-microsoft.net and thus the mistyped domains probably can hurt them!
Microsoft does plenty of evil and stupid things, but this is not amoung them.
Absolutely, removal of DRM would increase sales substantially, possibly to a phenominal level.
There are many, many people (like me) who are not buying anything because the DRM means it will not play on all my devices. I have poked around the Itunes site plenty and played the samples, but never bought anything, just said "I would buy that right now if they got rid of the DRM".
It does not apply to me, but I believe there will be substantial sales to people who intend to violate copyright. Somebody may be far more likely to buy a song if they know they can give it to all their friends. This is, I believe, an even larger market than people like me.
In both cases the users if they want the song, are forced to "steal" it, but would likely buy it, partly to be legal and not feel guilty, but mostly because it would be far easier to buy than steal.
How do locks on doors cause it to be easier to steal rather than buy (the #3 condition you said is caused by them)?
DRM literally means it is easier to steal than to buy. You steal, you get a copy that you know will play and can be put on any machine. If you buy, you are stuck with the DRM and have to remove it, a significant amount of work that was already done by the guy that uploaded the free "stealable" copy. The stolen copy is actually much more valuable as it includes this work.
Bzzt. You are wrong.
Re:Why are their two Mac Blenders? It's a drag.
on
New Blender Released
·
· Score: 1
For us OS/X performance is horrible if the/Applications directory is not on a local disk. Shared home directories are even worse.
You are running both ppc and intel machines from a shared/Applications directory? I would think that would make behavior so bad that accidentally running the ppc version on intel would be the least of your worries.
There is a switch to the linker that makes it look in the same directory as the executable for shared libraries, before looking in LD_LIBRARY_PATH and then in the built-in path.
Add this to gcc when linking the program: -Wl,--rpath,'$${ORIGIN}'
Slashdot Mirror
Yes your explanation makes a lot of sense.
I was more reacting to the use of "FUD" whenever anybody does any comparison of themselves to others. Mostly I see if from Microsoft defenders. If somebody says "Linux runs 10 times faster than Windows" they will say "that's FUD" when in fact what they mean is "that's a lie". There is a difference, in my opinion, though I never see anybody trying to correct them.
I think it may be Network Appliances.
1) High-contrast images such as credits can indeed be "projected at" the screen plane to prevent ghosting... but high-contrast images at any other plane will always have some ghosting.
True. However with this system at least the screen plane is guaranteed to not have ghosting. With two projectors that is not going to happen unless somebody carefully aligns them. And in the typical theatre that is not going to happen, or it will soon be bumped and knocked out of alignment.
mis-sync between the display of supposedly simultaneous left- and right-images
Yes I would agree that is certainly going to happen. Possibly a solution is to purposely render one of the images to be correct for 1/144 of a second later than the other one? Similar solutions are done for interlaced fields in TV, and I know this does not always work that well, so this may be an unsolvable problem.
"FUD" stands for "fear, uncertainty, and doubt". Apparently somehow it has had it's meaning twisted around here to mean "lies" about any subject. FUD can be entirely truthful, just worded to make people want to take the safer course.
I mostly see this here if anybody says anything negative about Microsoft, they are accused of "FUD". The term is wrong, except in a few cases such as when people give warnings like "DRM will destroy free speech" or that "Microsoft will bury/discontinue/etc that product, don't buy into it". These statements could be true or false, but they are FUD in that they are trying to scare people into disliking Microsoft or doing something Microsoft does not want, by predicting bad things to happen in the future.
Saying "Microsoft sucks" or "Linux is more than 10 times better than Windows" is NOT "FUD". It may be a false statement, but it is trying to make people choose Linux over Microsoft for positive reasons.
Conversely if Microsoft says their software is superior or has higher up time or lower tco, that is NOT "FUD". That may be truth or lies, but it is still not FUD. It is normal advertising, saying your product is better. They certainly do a lot more FUD than anybody else, if they say Linux might subject you to patent lawsuits, or copyright violations, or force you to give all your source code away, that is FUD. It can be true or false, it does not matter, what matters is that it is a vague threat of something that might happen in the future unless you do what they want.
I tried to make up something that Greenpeace would say that really is FUD but could not really come up with an example.
Does anybody agree with me, or is the use of "FUD" been completely twisted to mean almost any statement about a competitor?
There are already versions of this on Linux. One I use at work does it instead by making a new directory with the older versions of the files in it. To see yesterdays version of ./foo.txt you look at ./.snapshots/yesterday/foo.txt. This seems a lot nicer as you can more easily see when files are created and deleted as well as modified, though it is possible that technical limitations prevented this from using this naming scheme. I not sure what the system is, as it is a large file server, it is running Linux, though the filesystem is possibly proprietary.
I think the big deal is that this is a modification of the ext3 filesystem, rather than an all-new filesystem.
That must get real real hot with that bright light what is it, 7000-12000 watts going through a 2 inch space, must require lots of cold air being blown way fast
for that to work.
The drawing in the article seems to show that the lcd shutter is more like the size of a large windowpane and is some distance in front of the projector. I would think this is on purpose to solve the heat problems. May also be necessary so it can be added without modifying the DLP.
THis cannot be retrofitted on current DLPs, no way.... your source media is in a fixed format, either 24fps film, or 24fps MPEG2/MPEG4 data. At most it would be 60/50fps DLP. To get 144/150
No, the source media is a computer that is just like the game machines that can do 150 or more frames per second. It may be reading two 24 fps streams for the left and right eye and displaying each image 3 times. However I agree with you that it seems doubtful that a typical DLP projector for a theatre, designed to show 24 fps source images, is going to handle a 144 fps signal.
The article is a little vague about how it works, trying to make it sound more magical.
What the system does is alternate projections of the left and right eye images using the same DLP projector. They said 144 frames per second, which I think means that each film frame (of which there are 24 per second) is projected 3 times for each eye, this means each eye sees the image flickering at 72 times a second, which is above the threshold for most people to see flickering. The real technology is a special lcd screen that is put in front of the lens of the projector that changes it's polarization 144 times per second so each image is polarized differently.
The real advantage of this is that the same DLP projector used for non-3D films can be reused, just put the lcd in front of the lens when showing 3D. Any other system would require a second projector, which not only adds the cost of the projector, but the cost to mount it and add another aperture in the theatre wall. (actually another system would be shutter glasses with lcd lenses that turn on/off so each eye sees one side, but handing each customer an item that costs 10 or more dollars is probably out of the question) Also this system allows perfect alignment so that things that should appear at the screen plane really appear there, and high-contrast things like the credits can be projected at that distance with no ghosting.
It does appear fortunate that they can run at 144 frames per second, though if they were like consumer ones with a maximum of 90 or 100 it would still be an acceptable flicker rate of 45 or 50 (classic film projectors flickered 48 times a second due to having 1 extra vane on the shutter).
It is brilliant! And it certainly is different than Microsoft's implementation and not covered by any patent.
In fact it is a good deal simpler, it seems to reduce the color fringing naturally without having the calculation for one pixel effect nearby ones. It also only requires 2x horizontal resolution of the original rendering, rather than 3x.
From discussion here it sounds like what Microsoft patented is a filter, similar to their "font smoothing" filter, that turns a 1x resolution image into the colored cleartype image. That would mean it covers very little, and will be as obsolete as font smoothing in a year. However I kind of doubt this, I was under the impression that they were rendering at 3x horizontal resolution and then using this filter to put that in the lcd pixels without color artifacts. Anybody know?
In either case, rendering at 2x resolution and not having to do any filtering makes this be excluded from the Patent. Also looks like the results are better for very tiny fonts, too, mostly because there are only two color fringes, not 3.
Holy confusion, batman!
That was completely misleading. The use of 256 levels for making antialiased fonts is really old and has nothing to do with this.
What sub pixel rendering does is make the pixel represent *more* than 256 different possible combinations of the fg and bg colors, where combinations are how an fully opaque edge falls into the square the pixel represents. Exactly how many is unclear, it is not 256^3, but I think it is 3*256 for the case of an antialiased vertical straight edge of an object significantly larger than a pixel (it goes down as the edge gets less vertical or if another edge gets within a few pixels of it). This is done by using hardware quirks so that various of the 256^3 possible colors are interpreted by the viewer as different coverages.
The Apple technique made a pixel represent more than the 2 possible coverages you mentioned. This is done by using hardware quirks so that the some of the 4 (16? i dunno) possible colors are interpreted by the viewer as different coverages.
They certainly are related, but I don't believe enough to invalidate Microsoft's patent. The difference is that the "hardware quirks" are vastly different.
Copyright is what prevents you from doing things. This new GPL is reducing the number of ways it allows you to violate the copyright. So it is granting freedom, just less than before.
Yea, goddammit! If god didn't want us to drive cars, he would not have put roads and highways all over the earth!
Actually I suspect any Linux binaries from back then work, as the shared libraries they used are still there, and complex stuff was static-linked into the program.
The problems are more in the middle-age binaries, where GUI libraries started to show up (and this is still true now, so today's binaries may not work on the future machines). Windows is also sufferring from the same problem, though they do tend to be less likely to throw away the old api (Linux has the "excuse" that you can download and install the library if you need it, but in reality this is equivalent to "the api is missing and the software does not work").
I would also expect that this attempt to increase the gravity would result in a planet with enough gravity, but also a molten surface, and no water or atmosphere.
Actually you are off the mark. Threading has nothing to do with it, neither Windows or Gnome are using multiple threads to do the GUI (it actually would be a nightmare of complexity and probably far slower).
The problem is simply that Gnome, instead of using 15 guys passing instructions to each other, uses more like 200. That menu item probably has 20 or so ways of being changed to the correct foreign language and correctly formatted in swahili or whatever because there are many, many guys interested in how to change it. Meanwhile, the Windows system maybe has 15 guys (maybe each somewhat fatter and slower) and at most only one of them is interested in changing to swahili. In both cases it is enormously easier for any application that wants to display in swahili to just do it itself (with the equivalent of adding one really blindingly fast guy, comparatively speaking), so most of this overhead is a complete waste. It also does not help Gnome at all that those 200 workers are much more likely to do some completely unpredictable thing (something that makes perfect sense to that worker), on Windows a programmer can be reasonably certain that the text they say to draw on the screen will come out with at least the english letters correct.
Any real intelligent system would have one guy. The menu item should be a char* pointer to utf8 string, and it goes straight to the "draw this text here on the screen call". GUI's were usable on systems that ran at 5mhz, you know. Therefore even Windows must have an unnecessary overhead of at least 500 or so, and Gnome is even worse.
Linux has to try harder. Apparently Google can make Microsoft throw a chair, but Linux cannot yet.
You act as though there is no good reason to have Linux pre-installed. And then in the very next paragraph you say:
your average person isn't going to know what to do if their wireless card isn't working, or if they don't have support for MP3s, etc
Now are you deliberately being dense?
I thought I'd leave it at that, but just in case you cannot see what should be blatently obvious: the reason for pre-installed Linux is to solve the exact same problem you quoted.
What I was trying to say was "if you are using 16 bits for each pixel, it is a waste to store an integer in them, rather than using them for a floating point number".
A half float will have enormously higher resolution near 0 where the posterization shows. It can also represent numbers greater than 1, negative numbers, +/- infinity, and NaN. And it can efficiently represent the actual intensity of light, rather than a gamma curve that is required for integers to get the samples near zero closer together.
16 bit integers (or any number greater than 8) is a waste of time in image editing, and it is in some ways fortunate that Gimp avoids it, since it would take as much work to replace it as to add new data types.
If you are going to use more than 8 bits then using 16-bit "half" floats, or 32 bit floats, is a much better use of the space, and modern processors are easily able to keep up with this. It also helps a lot if the floating data represents linear light levels, rather than the log or gamma curve data that is in integer files.
I can read c# but i don't really program with it.
Sounds like you are actually quite a genius. I can assure you from personal experience reading others and even my own code, it is far easier to write C# and C++ than to read them!
Actually you are supposed to use some rather obscure system call to get the "Documents and Settings" directory name, and another obscure call to get the "username". In addition the Desktop for the user is not under this directory. Certainly easier to do getenv("HOME").
/Desktop. Oddly enough this is exactly the same as OS/X and the same as many Linux machines (some use "home" instead of "Users" but one symbolic link will fix that. Not sure if they put the directory name into an environment variable.
In any case it is pretty obvious that Microsoft fixed this in Vista. It is now "/Users/username" and the desktop is this directory
What I was shocked at was the amount of hoops they jumped through to avoid calling this directory "home" in the GUI. For instance, no icon and rather convoluted descriptions in the start menu, to avoid use of the word "home" or a house image. They are worried sick about being too compatable with Unix, no doubt.
They are using the 404 response correctly, this is what it was designed for! IE is trying to do something intellligent when it knows that the page is missing. What verisign did was fool every program (including IE) into thinking *all* pages exist, which breaks anything that wants to respond in a useful way to the page being mistyped.
I checked on a Windows machine, and they even let you change it! Didn't even bury it too deep in the configuration! You can go to google or bash-microsoft.net and thus the mistyped domains probably can hurt them!
Microsoft does plenty of evil and stupid things, but this is not amoung them.
Absolutely, removal of DRM would increase sales substantially, possibly to a phenominal level.
There are many, many people (like me) who are not buying anything because the DRM means it will not play on all my devices. I have poked around the Itunes site plenty and played the samples, but never bought anything, just said "I would buy that right now if they got rid of the DRM".
It does not apply to me, but I believe there will be substantial sales to people who intend to violate copyright. Somebody may be far more likely to buy a song if they know they can give it to all their friends. This is, I believe, an even larger market than people like me.
In both cases the users if they want the song, are forced to "steal" it, but would likely buy it, partly to be legal and not feel guilty, but mostly because it would be far easier to buy than steal.
How do locks on doors cause it to be easier to steal rather than buy (the #3 condition you said is caused by them)?
DRM literally means it is easier to steal than to buy. You steal, you get a copy that you know will play and can be put on any machine. If you buy, you are stuck with the DRM and have to remove it, a significant amount of work that was already done by the guy that uploaded the free "stealable" copy. The stolen copy is actually much more valuable as it includes this work.
Bzzt. You are wrong.
For us OS/X performance is horrible if the /Applications directory is not on a local disk. Shared home directories are even worse.
/Applications directory? I would think that would make behavior so bad that accidentally running the ppc version on intel would be the least of your worries.
You are running both ppc and intel machines from a shared
There is a switch to the linker that makes it look in the same directory as the executable for shared libraries, before looking in LD_LIBRARY_PATH and then in the built-in path.
Add this to gcc when linking the program: -Wl,--rpath,'$${ORIGIN}'