No, so people don't harm society with monopolies which do not provide incentive to producing new works. Copyright is not supposed to benefit the authors, it is supposed to benefit the public. Benefiting authors is only a side effect.
The value of the monopoly to the monopoly holder is a determining factor, as this is in important factor in how many resources they are likely to put into authoring the work. The long term value of the work to society is not, as this is impossible to predict.
Inefficient distribution means more of the copyright monopoly will be wasted. Less waste = less monopoly needed. Long term reputation is irrelevant to optimal monopoly length, because nobody produces work intending it to be only respected in the distant future.
The clock should start from date of first publication. A few thousand dollars isn't much spread over 10 years if your project is profitable. If your next project will take an unusually long time to complete and you can't afford to finance it yourself then you'll have to convince people to invest in it.
14 years is excessive, as improved distribution methods mean more people can access the work soon after publication than was possible when copyright was originally designed. Additionally, improved communications technology increases the pace of meme distribution, and as a great deal of value of a copyrighted work is in the novel social interactions it enables this shortens its time of highest value.
An automatic copyright of 5 years, with an extension of another 5 years available on paying a several thousand dollar fee sounds reasonable.
A popular belief, but not supported by scientific evidence. Humans read by recognizing individual letters in parallel. We read lowercase faster only because we have more practice at doing so.
Tomatoes are easy to grow, and by choosing a variety optimized for flavor rather than yield or disease resistance (not a major concern in low density farming) you can grow tomatoes better than anything you can buy. Once you've tasted a home grown tomato fresh from the vine you'll understand why tomato growing is popular.
A partial solution, albeit very CPU intensive. By speculatively running the emulation several frames ahead for a pruned tree of probable user inputs (eg. assume left+right aren't pressed at the same time, ignore the start/select buttons, and assume button state doesn't change every frame), classify sprites by future movement as moving or static relative to backgrounds. Track sprites by graphics data and position (assuming it's the same sprite if movement is under a certain threshold per frame) rather than number, to avoid being fooled by anti-flicker tricks and sprite duplication tricks. Track history of movement, so once a sprite is classified as mobile it remains classified as mobile. Analyze relative motion of mobile sprites to other mobile sprites and identify composite sprites, rendering each into its own buffer. Render static sprites directly onto the relevant background buffer. The further you look into the future the more accurate your classifications can be, but the combinatorial explosion greatly limits this in practice.
This does nothing to solve the false edge detection problem, and as with any graphics algorithm, the more cleverness you add the more jarring it becomes when it inevitably fails (see deinterlacing algorithms). It would be a lot of effort to implement for something that wouldn't necessarily be any better than nearest neighbor scaling.
Scaling each of those individually might cause artifacts at the seams.
Then the solution is to render all of the sprites to a single separate buffer, as well as each background layer to its own buffer, before filtering and compositing all the buffers. Optionally track dirty rectangles so you don't have to filter the whole thing.
Cesium is more reactive, but it does not produce a spectacular explosion:
Generally speaking, the hydrogen gas explosion contributes more to the overall visible size of the explosion than does the initial metal-water reaction. And this brings into play an important fact: When you go down the periodic table from lithium to cesium, the atomic weight goes up from 6.94 to 132.9. Higher atomic weight means fewer atoms per unit of weight, and the amount of hydrogen gas generated is directly proportional to the number of atoms. So 5 grams of cesium liberates only about one twentieth as much hydrogen as five grams of lithium, and a bit over one sixth as much as 5 grams of sodium.
Last I checked, ZSNES filtered its whole framebuffer instead of filtering the sprites independently then compositing them onto the background. This causes annoying artifacts as the sprites move. Do these filters even support alpha channel?
In addition you have the false edge detection artifacts, such as the Yoshi sign from the Super Mario World example, or the ropes on the hostage from the Metal Slug example.
To properly scale pixel art is "AI complete", as it involves figuring out what the original artist was trying to do and then redrawing the graphics at higher resolution but in the same style. I'd rather tolerate consistent blockiness than inconsistent scaler artifacts.
For perfect graphics, you need no filtering and integer ratio nearest neighbor resizing. In sdlmame's mame.ini:
filter 0
unevenstretch 0
This won't necessarily fill the whole screen. If the game's pixel aspect isn't 1:1, which is very common, then you may have a problem. With a CRT you can adjust your display pixel aspect to match, but on an LCD you might have to suffer uneven pixel size from non-integer ratio scaling. This isn't so noticeable if the display resolution is much higher than the game resolution, but ideally there should be a single axis filtering scaling routine. Something like Mitchell-Netravali filtering along the scanlines of the original image and integer ratio nearest neighbor scaling on the other axis the other would look good. Alternatively, with a high enough display resolution or the right game pixel aspect the closest integer ratio scaling could be a good enough approximation.
You also have to consider temporal scaling. The display refresh rate needs to be an integer multiple of the game refresh rate, and the display vertical retrace must be used as the timing source. If that integer multiple is greater than 1 you'll probably have to modify the emulator for frame redrawing. This is usually trivial, eg. duplicate the graphics buffer swap call. Be aware that frame redrawing on CRTs sacrifices some of their superior motion quality, but I consider this a worthwhile tradeoff for avoiding flicker and improving control latency. This kind of simple frame redrawing increases CPU requirement. Avoid triple-buffering or any higher n-buffering, as it increases control latency.
If your LCD is locked to 60Hz refresh and you're playing a game that used a different refresh rate then you have no option but to run it at the wrong speed (I consider non 1:1 ratio pulldown completely unacceptable). This will break the audio, so you may want to investigate the "Cabmame" patchset which can resample the audio to match. It uses DirectX, so an sdlmame port will be difficult. Alternatively just tolerate the glitchy audio.
One other problem is CRT artifact simulation. I've never seen convincing fake scanlines or subpixels, so I ignore this altogether. If you really care then buy a real low resolution CRT.
They sometimes work, but when they fail the results are more distracting than nearest neighbor resampling. The "Yoshi" sign in Test Case 2 is a good example, where the algorithm has failed to identify the gradually sloping line and exaggerated the stepped appearance.
The graphics are crisp, intense duplicates of the original.
But from the screenshot we can see that both the sprites and the textures have been filtered. Filtering the textures is no problem, but the sprites are "pixel art" - they are designed around the pixel boundaries to pack more detail into a limited resolution. It's the same principle as manually hinting fonts. The only acceptable scaling method for pixel art is unfiltered "nearest neighbor" scaling, as used in the original game. This new version is not "crisp", it is an ugly blurred mess.
This is true, but music had much better mastering 20 years ago. Thanks to the loudness war, that "all-time favourite band you havn't heard yet" is probably ruined by clipping and compressor pumping. It doesn't matter how skilled the musicians are if the recording is unlistenable.
If somebody presents something as worthwhile when they do not believe it to be worthwhile they are lying. With the pens it is a reasonable assumption that 3 for the price of 2 is worthwhile. With the extended warranties it is not a reasonable assumption, as the only way it could be economically viable is if the user's time is extraordinarily valuable, in which case they are highly unlikely to be using products from Office Depot. Having unnecessary items on the shelves is not the same thing because nobody is actively trying to deceive people into buying them.
to present as worthy of confidence, acceptance, use, etc.; commend; mention favorably: to recommend an applicant for a job; to recommend a book.
to represent or urge as advisable or expedient: to recommend caution.
to advise, as an alternative; suggest (a choice, course of action, etc.) as appropriate, beneficial, or the like: He recommended the blue-plate special. The doctor recommended special exercises for her.
to make desirable or attractive: a plan that has very little to recommend it.
Common usage of "recommend" implies a benefit to the person being recommended to. In your example the salesperson can reasonably assume this to be true, as many people want more than one pen for cheap, but in the case of extended warranties the salesperson knows it is very unlikely to be beneficial to the customer. The intention is to trick the customer into buying something they don't need, so this is not using "recommend" as in common English usage.
Not at all. The true reason the salesperson recommends the extended warranty is because they get commission. The reason given in the script is an unrelated fact, so by following the script the salesperson is lying.
All mp3 compression is digital. Whether the audio was originally analog or not is unrelated to hard it will be to compress. The 128kbps mp3s were preferred in the case of music with a lot of high frequency content, which is hard to compress. Because of this it is more distorted, and this "sizzle" distortion if what some people prefer.
Slashdot Mirror
No, so people don't harm society with monopolies which do not provide incentive to producing new works. Copyright is not supposed to benefit the authors, it is supposed to benefit the public. Benefiting authors is only a side effect.
The value of the monopoly to the monopoly holder is a determining factor, as this is in important factor in how many resources they are likely to put into authoring the work. The long term value of the work to society is not, as this is impossible to predict.
Inefficient distribution means more of the copyright monopoly will be wasted. Less waste = less monopoly needed. Long term reputation is irrelevant to optimal monopoly length, because nobody produces work intending it to be only respected in the distant future.
The clock should start from date of first publication. A few thousand dollars isn't much spread over 10 years if your project is profitable. If your next project will take an unusually long time to complete and you can't afford to finance it yourself then you'll have to convince people to invest in it.
14 years is excessive, as improved distribution methods mean more people can access the work soon after publication than was possible when copyright was originally designed. Additionally, improved communications technology increases the pace of meme distribution, and as a great deal of value of a copyrighted work is in the novel social interactions it enables this shortens its time of highest value.
An automatic copyright of 5 years, with an extension of another 5 years available on paying a several thousand dollar fee sounds reasonable.
A popular belief, but not supported by scientific evidence. Humans read by recognizing individual letters in parallel. We read lowercase faster only because we have more practice at doing so.
Detailed explanation:
http://www.microsoft.com/typography/ctfonts/WordRecognition.aspx
It doesn't matter how old that usage is -- it's still propaganda.
Tomatoes are easy to grow, and by choosing a variety optimized for flavor rather than yield or disease resistance (not a major concern in low density farming) you can grow tomatoes better than anything you can buy. Once you've tasted a home grown tomato fresh from the vine you'll understand why tomato growing is popular.
Weather is a chaotic system. There's no guarantee things would return to normal.
You may be interested in this:
http://solarcooking.wikia.com/
Solar cooking isn't practical in every climate, but when it works it can be very effective.
A partial solution, albeit very CPU intensive. By speculatively running the emulation several frames ahead for a pruned tree of probable user inputs (eg. assume left+right aren't pressed at the same time, ignore the start/select buttons, and assume button state doesn't change every frame), classify sprites by future movement as moving or static relative to backgrounds. Track sprites by graphics data and position (assuming it's the same sprite if movement is under a certain threshold per frame) rather than number, to avoid being fooled by anti-flicker tricks and sprite duplication tricks. Track history of movement, so once a sprite is classified as mobile it remains classified as mobile. Analyze relative motion of mobile sprites to other mobile sprites and identify composite sprites, rendering each into its own buffer. Render static sprites directly onto the relevant background buffer. The further you look into the future the more accurate your classifications can be, but the combinatorial explosion greatly limits this in practice.
This does nothing to solve the false edge detection problem, and as with any graphics algorithm, the more cleverness you add the more jarring it becomes when it inevitably fails (see deinterlacing algorithms). It would be a lot of effort to implement for something that wouldn't necessarily be any better than nearest neighbor scaling.
Then the solution is to render all of the sprites to a single separate buffer, as well as each background layer to its own buffer, before filtering and compositing all the buffers. Optionally track dirty rectangles so you don't have to filter the whole thing.
http://www.theodoregray.com/PeriodicTable/AlkaliBangs/
Cesium is more reactive, but it does not produce a spectacular explosion:
Last I checked, ZSNES filtered its whole framebuffer instead of filtering the sprites independently then compositing them onto the background. This causes annoying artifacts as the sprites move. Do these filters even support alpha channel?
In addition you have the false edge detection artifacts, such as the Yoshi sign from the Super Mario World example, or the ropes on the hostage from the Metal Slug example.
To properly scale pixel art is "AI complete", as it involves figuring out what the original artist was trying to do and then redrawing the graphics at higher resolution but in the same style. I'd rather tolerate consistent blockiness than inconsistent scaler artifacts.
For perfect graphics, you need no filtering and integer ratio nearest neighbor resizing. In sdlmame's mame.ini:
filter 0
unevenstretch 0
This won't necessarily fill the whole screen. If the game's pixel aspect isn't 1:1, which is very common, then you may have a problem. With a CRT you can adjust your display pixel aspect to match, but on an LCD you might have to suffer uneven pixel size from non-integer ratio scaling. This isn't so noticeable if the display resolution is much higher than the game resolution, but ideally there should be a single axis filtering scaling routine. Something like Mitchell-Netravali filtering along the scanlines of the original image and integer ratio nearest neighbor scaling on the other axis the other would look good. Alternatively, with a high enough display resolution or the right game pixel aspect the closest integer ratio scaling could be a good enough approximation. You also have to consider temporal scaling. The display refresh rate needs to be an integer multiple of the game refresh rate, and the display vertical retrace must be used as the timing source. If that integer multiple is greater than 1 you'll probably have to modify the emulator for frame redrawing. This is usually trivial, eg. duplicate the graphics buffer swap call. Be aware that frame redrawing on CRTs sacrifices some of their superior motion quality, but I consider this a worthwhile tradeoff for avoiding flicker and improving control latency. This kind of simple frame redrawing increases CPU requirement. Avoid triple-buffering or any higher n-buffering, as it increases control latency. If your LCD is locked to 60Hz refresh and you're playing a game that used a different refresh rate then you have no option but to run it at the wrong speed (I consider non 1:1 ratio pulldown completely unacceptable). This will break the audio, so you may want to investigate the "Cabmame" patchset which can resample the audio to match. It uses DirectX, so an sdlmame port will be difficult. Alternatively just tolerate the glitchy audio. One other problem is CRT artifact simulation. I've never seen convincing fake scanlines or subpixels, so I ignore this altogether. If you really care then buy a real low resolution CRT.
They sometimes work, but when they fail the results are more distracting than nearest neighbor resampling. The "Yoshi" sign in Test Case 2 is a good example, where the algorithm has failed to identify the gradually sloping line and exaggerated the stepped appearance.
But from the screenshot we can see that both the sprites and the textures have been filtered. Filtering the textures is no problem, but the sprites are "pixel art" - they are designed around the pixel boundaries to pack more detail into a limited resolution. It's the same principle as manually hinting fonts. The only acceptable scaling method for pixel art is unfiltered "nearest neighbor" scaling, as used in the original game. This new version is not "crisp", it is an ugly blurred mess.
You can disable the labels from the "Interface" tab of the "Appearance Preferences".
This is true, but music had much better mastering 20 years ago. Thanks to the loudness war, that "all-time favourite band you havn't heard yet" is probably ruined by clipping and compressor pumping. It doesn't matter how skilled the musicians are if the recording is unlistenable.
If somebody presents something as worthwhile when they do not believe it to be worthwhile they are lying. With the pens it is a reasonable assumption that 3 for the price of 2 is worthwhile. With the extended warranties it is not a reasonable assumption, as the only way it could be economically viable is if the user's time is extraordinarily valuable, in which case they are highly unlikely to be using products from Office Depot. Having unnecessary items on the shelves is not the same thing because nobody is actively trying to deceive people into buying them.
Where do you get that 86dB figure? I've always seen 16 bit dynamic range quoted as 96dB.
Common usage of "recommend" implies a benefit to the person being recommended to. In your example the salesperson can reasonably assume this to be true, as many people want more than one pen for cheap, but in the case of extended warranties the salesperson knows it is very unlikely to be beneficial to the customer. The intention is to trick the customer into buying something they don't need, so this is not using "recommend" as in common English usage.
Not at all. The true reason the salesperson recommends the extended warranty is because they get commission. The reason given in the script is an unrelated fact, so by following the script the salesperson is lying.
CD quality is already overkill: http://www.aes.org/e-lib/browse.cfm?elib=14195
All mp3 compression is digital. Whether the audio was originally analog or not is unrelated to hard it will be to compress. The 128kbps mp3s were preferred in the case of music with a lot of high frequency content, which is hard to compress. Because of this it is more distorted, and this "sizzle" distortion if what some people prefer.