that 65% figure, while true, is a bit extreme and only applies to the very highest income brackets.. most people pay 30-35% income tax in sweden
Swedish taxes are actually much higher than that, it's just that the employer pays most of it behind the scenes. To give someone a gross income of x (0.65x after taxes), the employer has to spend about 1.40x. The extra 0.4x is not called "income tax", but "lånekostnadspålägg" or something like that.
Still, I use [GIMP] because it's a free way to do some common picture operations (like fudging color)
For color fudging if you have a bunch of pictures from a digicam, nothing beats good old XV, IMO. It's shareware (free (beer) for personal use), it uses the antique Athena widget set, but I find it extremely effective for editing pictures from a digicam. Press "E" and you get a single color-editor pane with all I need: curves for brightness and R/G/B. Select a rectangle and press "C" for crop. Press Ctrl-S for save and click a few times (the new windows come with the OK button exactly under the mouse cursor) to overwrite the original. Press space to get the next picture from the command line. To do the same in Gimp takes ten times longer.
It's not deceptive as noise is also perceived on a logarithmic scale by humans.
Certainly the human ear works like that, but if you want to make a decision how to spend your money, it's hard to compare numbers in dB. In the original test, you might have read something like:
case fans: 5 dB PSU:1 dB GPU/CPU: 2.5 dB Materials/enclosure: 2 dB low-voltage resistor:7.5 dB.
This may lead you to believing that replacing the fans AND adding the resistor together will give you 12.5 dB noise reduction, while the rest gives you only 5.5 dB extra for . It is not meaningful to use dB in this situation, where you take out one noise source after the other. It would be meaningful to use dB if one were discussing an isolating enclosure for the whole computer.
The article shows how the noise level goes down by taking more and
more measures, with the biggest noise reduction at the last
step. However, this is deceptive since the noise is on a dB (i.e.,
logaritmic) scale. If his test computer produced 100 units of noise to
start with, then the reductions were:
73 units (low-noise case fans, 40 pounds)
5.5 units (low-noise power supply, 90 pounds)
9.6 units (CPU/GPU cooling, 75 pounds)
5.2 units (acoustic materials and HD enclosure, 128 pounds)
5.8 units (resistors on case fans, 0 pounds)
1.3 units (remaining)
So, by far the most bang for the bucks is in the case fans (with
resistors), accounting for 79% of the noise. The worst deal is the
acoustic materials and HD enclosure, which cost a whopping 128 pounds
for only 5% of the total noise.
I remember that applesoft basic had a way to renumber the whole thing in 10 increments
That was not included in Applesoft basic; you had to load a separate program first, that added some arcane command such as "REN 10,10000,500,10" to renumber the lines 10-10000 to 500 with 10 increments. Or something like that.
Let's see whether posting this will undo my stupid moderation...
anything printed offset uses halftones. the colors DON'T overlap.
Tell me what happens if you have a 70% covering level in cyan and 50% in magenta at the same time without any cyan overlapping the magenta. Does your paper surface magically expand by 20% ?
Do you know how many watts a 200HP car engine produces in theory?... So before you go apeshit over a video card that uses 100 piddly watts, get some perspective.
And how many hours in a week do you typically drive that car with the gas pedal all the way down?
Firstly as the parent poster correctly writes to produce a colour epaper display you don't actually need overlapping colours,
It seems that I'm the only one here who believes that you need to overlap colors. Can you explain me how C, M, Y, K pixels next to each other, each covering 1/4 of the surface, can give a black impression instead of a grey, when regarded from a distance?
Sorry for being unclear. The quote refers to my previous (long) post, where I tried to explain that CMY subpixels won't work. The Nature link shows a picture of 4 pixels, colored white/cyan, white/magenta, white/yellow, and white/black. It is obvious (to me at least) that the result can never be darker than some grey if all pixels are switched on fully.
What could work, is 8 subpixels, colored red, green, blue, cyan, magenta, yellow, white, and black, with varying sizes, such that a pixel can be fully covered by for example yellow, or half by yellow and half by black. Then you have the full gamut [disclaimer: I wrote part of that wikipedia article] of colors that you have with CMYK printing or RGB screens. But that requires a new technology.
The nature article you point to has an image indicating that the display is of the variant "each pixel tunable between white and saturated color". That is not full-color.
The colors do not need to overlap. Why would they? As you noted, monitors use side-by-side colors rather than overlapping colors, and e-paper would be no different in this regard.
I can imagine technologies for colored pixels that result in either a scale from black to a saturated color (be it RGB or CMY), or from white to a saturated color (RGB or CMY). Let us take white-to-CMY as an example.
Colors are mixed ADDITIVELY. Why? Your eye is receiving light from all three pixels! But each pixel only represents 1/3 of the reflecting surface, so the
cyan pixels can represent RGB values between (0.0,0.33,0.33) (ON) and (.33,.33,.33) (OFF),
the magenta between (0.33,0,0.33) and (.33,.33,.33),
and the yellow between (0,0.33,0.33) and (.33,.33,.33).
(These RGB values are without gamma correction).
Examples:
All pixels switched off (white): white color (100% reflection of all wavelengths).
All pixels switched on: total RGB (.67,.67,.67). Light grey.
Only cyan switched on: 2/3 of the pixels, i.e. the magenta and yellow ones, are still white. 1/3 of the pixels is cyan. The result will be an unsaturated cyan, corresponding to RGB (0.67,1,1) (ignoring gamma correction), instead of pure cyan (0,1,1).
Cyan and magenta switched on. You may think that that results in blue (0,0,1), but it will actually be the ADDITIVE sum of cyan(0,0.33,0.33) + magenta(0.33,0,0.33) + white(0.33,0.33,0.33) = (0.67,0.67,1), an pale light blue.
So this display will be able to represent colors varying between white and light grey, and some pale colors lying inbetween.
Similar problems will occur with RGB pixels, or with a black background instead of a white.
In order to mix cyan and magenta into pure blue, you have to cover the whole surface with magenta, and then with cyan on top of that, so that the magenta layer will filter out all green light, and the cyan layer will filter out all the red light. That will be very hard, though not fundamentally impossible, to implement as e-paper.
in another couple years, i bet they'll have this with higher resolution, higher contrast, and full color
Don't wait too long for the full-color version. In order to do color reproduction without a backlight, you need overlapping colored pixels, and that is an order of magnitude harder than just putting colored pixels next to each other, as on a TFT or CRT screen. If you want to create white and put red, green, and blue reflecting pixels next to each other, the result will be reflecting roughly 1/3 of the light in the best case, which is grey. It's comparable to a colored mobile-phone display with the backlight switched off.
In addition, there is no special protocol needed to read them. A pin or sewing needle stuck in a paper cone will work!
True, when I was a kid, I used a home-built LEGO turntable with sewing needle and paper cone to play a couple of LPs that belonged to my father. I could listen to the music, but dad was not amused.:-)
I have found that writing very lightly with a soft, dark graphite pencil works well.
I wouldn't be so keen on having particles of electrically conducting graphite being spun off the disc inside the drive... But you're right that it probably won't damage the disc.
If you're very paranoid, you might consider not labelling the CD at all
Or write in the data-less area around the center of the disc.
The death star in Star Wars was able to shatter a planet to pieces. One can calculate that the energy needed to overcome the gravitational pull is about G*rho^2*r^5, where G=the gravitational constant, rho the planet's density, and r its radius. For an earth-sized planet, that amounts to 1e30 J, or 6e13 kg of matter to be converted into energy. If the Death Star were completely consisting of concentrated antimatter, then it would have been 3 km in diameter and be able to fire exactly one shot. Yes, that is more than 100 ft.:)
I remembered being told that this released some toxic gas in the process of being dissolved.
Styrofoam is the polymer polystyrene in foam form, that is, with a lot of small gas bubbles. Google tells me that commonly used gases are ethylene, CFCs (not commonly anymore), and HCFCs. These gases are not particularly toxic to humans, but can be an environmental issue (i.e., the hole in the ozone layer), especially CFCs and HCFCs.
Do you really need much compute power in a walk-about machine to do email, web browsing, word smithing ? In a trade off give me battery time over machine horsepower every time.
Common sense would say so, but unfortunately, newer browsers, widget libraries, and window managers use a lot of resources. I used to use Redhat 7.1 with FVWM and Opera 6. Blazingly fast on my P3/450. Then, because of frustration with incompatible libraries for newer RPMs, I upgraded to Fedora/Opera 7. I still run Fvwm, but this new Opera version (with a newer Qt library, I presume) needs about 2 seconds of CPU time just for getting in and out of focus. If I look carefully, I can see that the borders of the windows inside the Opera window change a little bit depending on the focus. Emacs and xterm still run fine, but everything that has Gtk or Qt is slow as hell.
When starting, only use a Bic Roundstick pen. These are the classic cheap ball point pen,
I have read recommendations against ball point pens. The argument is that they tend to glide too smoothly over the paper to give you adequate feedback. Pencils and fountain pens are better in that respect: you are really in direct contact with the paper without a ball bearing in between. A good fountain pen will not blot ink when you use something better than newspaper to write on.
Before drawing letters, first practice on smooth movements in the basic elements that occur in letters: arcades (write something like mmmmmmm fluently), guirlandes (uuuuuuuuuuuu), upward loops (like lllll or eeeee), downward loops (mirror image of upward loops), and then the hard one: interleaved up/down loops, which you need in order to write an elegant 'f'.
Then study the humananistic cursive (see some other posts). The idea is that you mainly use the motion patterns you just practice, and lift the tip of the pen at the right moment. For example (assuming that you know how cursive handwriting is supposed to look like in theory):
When writing the combination "nd", lift the pen when you've finished "n" but continue moving in the northeast direction until you reach the starting point for the letter "d". If you have perfected your technique, the belly of the "d" will touch the tail of the "n", such that the result is connected.
This technique is in contrast to conventional cursive, where you keep contact, but you have to draw a curved line towards the starting point of the "d" and then you have to reverse direction and follow the line you just drew in backward direction. That is hard to do neatly and is unnessecary.
Don't draw loops on the letters bfghjkl. Do the same motion, but lift the pen during the first half of the loop. That will not take any longer, but the resulting writing will be much clearer without the redundant forest of loops.
Write the "r" like the printed r, not mirror-imaged.
Of course, learn the difference between UPPERCASE and lowercase. I don't know about USA habits, but I've seen many people who mix lowercase and uppercase letter shapes.
I find that a small amount of dishwashing detergent rubbed along the cable works great. It doesn't hurt the pet to be exposed to this small amount, and it leaves a nasty enough taste that they won't go back for a second try.
Actually, dishwashing agent is made to taste horrible to discourage children from eating it. You'd be better of with just that magical flavour without the dishwashing stuff around it.
If 99% of the world uses "hacker" in a negative context, I think the real hackers had better find a new term
But which term? An earlier discussion showed that alternatives such as "programming enthousiast" and "codesmith" do not really carry the same associations.
12 points = 1 pica = 1/6 of an inch height, which I have often seen ignored.
The point size of a font refers to the height of the ancient lead type, where every letter is sitting on a rectangular piece of lead. Some fonts are designed to never reach the borders of that rectangle, i.e. a parenthesis "(" in a 12-point font can actually be a bit less than 12 pt high. Only if a "(" is more than the specified 12pt, then something is really wrong.
Furthermore, 12 points are NOT the same as 1/6 inch. There are actually 72.27 (American printer's) points in an inch, but someday, Adobe decided that for digital typesetting, a round number such as 72 points per inch was easier. (The number 72.27 pt/in is easy to remember, but that is pure coincidence. See point units.)
Re:It's like Netscape v. Microsoft in that...
on
Google v. Microsoft
·
· Score: 1
>Would Microsoft be so low as to "embellish" the robots.txt file hosted on IIS sites so as to include a line forbidding the GoogleBot?
Though Joe Average might accept the default settings for the search button, webmasters will certainly get very pissed if their webserver prevents them from showing up in the #1 or (shiver) #2 search engine Google... Even Microsoft wouldn't risk loosing the IIS market.
Most keyboards slope towards you; you have to pull your fingers and hands back just to keep from pushing a key.
Most keyboards have small foldable feet to decrease the sloping angle. For some reason, most people believe that it is more comfortable with the feet extended, but fold the damn things down if you have wrist problems! Get a wrist pad to decrease the angle even further, especially if you have the bad habit of supporting your wrists on the table. (I can't type without either resting my wrists on the pad/table or sitting in a chair with properly adjusted armrests)
This is particularly true with ergonomic keyboards.
Some years ago I tried the MS natural keyboard when it was new on the market. Though I didn't like it because of how the "click" felt, it was actually sloping the other way around: lifted on the front and lower on the back (where the function keys are). I'm a bit surprised to hear that modern split keyboards do have those evil feet on the back side.
Slashdot Mirror
Swedish taxes are actually much higher than that, it's just that the employer pays most of it behind the scenes. To give someone a gross income of x (0.65x after taxes), the employer has to spend about 1.40x. The extra 0.4x is not called "income tax", but "lånekostnadspålägg" or something like that.
For color fudging if you have a bunch of pictures from a digicam, nothing beats good old XV, IMO. It's shareware (free (beer) for personal use), it uses the antique Athena widget set, but I find it extremely effective for editing pictures from a digicam. Press "E" and you get a single color-editor pane with all I need: curves for brightness and R/G/B. Select a rectangle and press "C" for crop. Press Ctrl-S for save and click a few times (the new windows come with the OK button exactly under the mouse cursor) to overwrite the original. Press space to get the next picture from the command line. To do the same in Gimp takes ten times longer.
Try one of these:
or
This will temporarily degrade your CDROM to a quiet 20-speed model, if you run the correct OS, that is.Certainly the human ear works like that, but if you want to make a decision how to spend your money, it's hard to compare numbers in dB. In the original test, you might have read something like:
case fans: 5 dB
PSU:1 dB
GPU/CPU: 2.5 dB
Materials/enclosure: 2 dB
low-voltage resistor:7.5 dB.
This may lead you to believing that replacing the fans AND adding the resistor together will give you 12.5 dB noise reduction, while the rest gives you only 5.5 dB extra for . It is not meaningful to use dB in this situation, where you take out one noise source after the other. It would be meaningful to use dB if one were discussing an isolating enclosure for the whole computer.
73 units (low-noise case fans, 40 pounds)
5.5 units (low-noise power supply, 90 pounds)
9.6 units (CPU/GPU cooling, 75 pounds)
5.2 units (acoustic materials and HD enclosure, 128 pounds)
5.8 units (resistors on case fans, 0 pounds)
1.3 units (remaining)
So, by far the most bang for the bucks is in the case fans (with resistors), accounting for 79% of the noise. The worst deal is the acoustic materials and HD enclosure, which cost a whopping 128 pounds for only 5% of the total noise.
That was not included in Applesoft basic; you had to load a separate program first, that added some arcane command such as "REN 10,10000,500,10" to renumber the lines 10-10000 to 500 with 10 increments. Or something like that.
Let's see whether posting this will undo my stupid moderation...
Tell me what happens if you have a 70% covering level in cyan and 50% in magenta at the same time without any cyan overlapping the magenta. Does your paper surface magically expand by 20% ?
And how many hours in a week do you typically drive that car with the gas pedal all the way down?
It seems that I'm the only one here who believes that you need to overlap colors. Can you explain me how C, M, Y, K pixels next to each other, each covering 1/4 of the surface, can give a black impression instead of a grey, when regarded from a distance?
Sorry for being unclear. The quote refers to my previous (long) post, where I tried to explain that CMY subpixels won't work. The Nature link shows a picture of 4 pixels, colored white/cyan, white/magenta, white/yellow, and white/black. It is obvious (to me at least) that the result can never be darker than some grey if all pixels are switched on fully.
What could work, is 8 subpixels, colored red, green, blue, cyan, magenta, yellow, white, and black, with varying sizes, such that a pixel can be fully covered by for example yellow, or half by yellow and half by black. Then you have the full gamut [disclaimer: I wrote part of that wikipedia article] of colors that you have with CMYK printing or RGB screens. But that requires a new technology.
The nature article you point to has an image indicating that the display is of the variant "each pixel tunable between white and saturated color". That is not full-color.
The colors do not need to overlap. Why would they? As you noted, monitors use side-by-side colors rather than overlapping colors, and e-paper would be no different in this regard.
I can imagine technologies for colored pixels that result in either a scale from black to a saturated color (be it RGB or CMY), or from white to a saturated color (RGB or CMY). Let us take white-to-CMY as an example.
Colors are mixed ADDITIVELY. Why? Your eye is receiving light from all three pixels! But each pixel only represents 1/3 of the reflecting surface, so the
- cyan pixels can represent RGB values between (0.0,0.33,0.33) (ON) and (.33,.33,.33) (OFF),
- the magenta between (0.33,0,0.33) and (.33,.33,.33),
- and the yellow between (0,0.33,0.33) and (.33,.33,.33).
(These RGB values are without gamma correction).Examples:
- All pixels switched off (white): white color (100% reflection of all wavelengths).
- All pixels switched on: total RGB (.67,.67,.67). Light grey.
- Only cyan switched on: 2/3 of the pixels, i.e. the magenta and yellow ones, are still white. 1/3 of the pixels is cyan. The result will be an unsaturated cyan, corresponding to RGB (0.67,1,1) (ignoring gamma correction), instead of pure cyan (0,1,1).
- Cyan and magenta switched on. You may think that that results in blue (0,0,1), but it will actually be the ADDITIVE sum of cyan(0,0.33,0.33) + magenta(0.33,0,0.33) + white(0.33,0.33,0.33) = (0.67,0.67,1), an pale light blue.
So this display will be able to represent colors varying between white and light grey, and some pale colors lying inbetween.Similar problems will occur with RGB pixels, or with a black background instead of a white.
In order to mix cyan and magenta into pure blue, you have to cover the whole surface with magenta, and then with cyan on top of that, so that the magenta layer will filter out all green light, and the cyan layer will filter out all the red light. That will be very hard, though not fundamentally impossible, to implement as e-paper.
Don't wait too long for the full-color version. In order to do color reproduction without a backlight, you need overlapping colored pixels, and that is an order of magnitude harder than just putting colored pixels next to each other, as on a TFT or CRT screen. If you want to create white and put red, green, and blue reflecting pixels next to each other, the result will be reflecting roughly 1/3 of the light in the best case, which is grey. It's comparable to a colored mobile-phone display with the backlight switched off.
True, when I was a kid, I used a home-built LEGO turntable with sewing needle and paper cone to play a couple of LPs that belonged to my father. I could listen to the music, but dad was not amused. :-)
I wouldn't be so keen on having particles of electrically conducting graphite being spun off the disc inside the drive... But you're right that it probably won't damage the disc.
If you're very paranoid, you might consider not labelling the CD at all
Or write in the data-less area around the center of the disc.
The death star in Star Wars was able to shatter a planet to pieces. One can calculate that the energy needed to overcome the gravitational pull is about G*rho^2*r^5, where G=the gravitational constant, rho the planet's density, and r its radius. For an earth-sized planet, that amounts to 1e30 J, or 6e13 kg of matter to be converted into energy. If the Death Star were completely consisting of concentrated antimatter, then it would have been 3 km in diameter and be able to fire exactly one shot. Yes, that is more than 100 ft. :)
Styrofoam is the polymer polystyrene in foam form, that is, with a lot of small gas bubbles. Google tells me that commonly used gases are ethylene, CFCs (not commonly anymore), and HCFCs. These gases are not particularly toxic to humans, but can be an environmental issue (i.e., the hole in the ozone layer), especially CFCs and HCFCs.
Common sense would say so, but unfortunately, newer browsers, widget libraries, and window managers use a lot of resources. I used to use Redhat 7.1 with FVWM and Opera 6. Blazingly fast on my P3/450. Then, because of frustration with incompatible libraries for newer RPMs, I upgraded to Fedora/Opera 7. I still run Fvwm, but this new Opera version (with a newer Qt library, I presume) needs about 2 seconds of CPU time just for getting in and out of focus. If I look carefully, I can see that the borders of the windows inside the Opera window change a little bit depending on the focus. Emacs and xterm still run fine, but everything that has Gtk or Qt is slow as hell.
I have read recommendations against ball point pens. The argument is that they tend to glide too smoothly over the paper to give you adequate feedback. Pencils and fountain pens are better in that respect: you are really in direct contact with the paper without a ball bearing in between. A good fountain pen will not blot ink when you use something better than newspaper to write on.
Before drawing letters, first practice on smooth movements in the basic elements that occur in letters: arcades (write something like mmmmmmm fluently), guirlandes (uuuuuuuuuuuu), upward loops (like lllll or eeeee), downward loops (mirror image of upward loops), and then the hard one: interleaved up/down loops, which you need in order to write an elegant 'f'.
Then study the humananistic cursive (see some other posts). The idea is that you mainly use the motion patterns you just practice, and lift the tip of the pen at the right moment. For example (assuming that you know how cursive handwriting is supposed to look like in theory):
- When writing the combination "nd", lift the pen when you've finished "n" but continue moving in the northeast direction until you reach the starting point for the letter "d". If you have perfected your technique, the belly of the "d" will touch the tail of the "n", such that the result is connected.
This technique is in contrast to conventional cursive, where you keep contact, but you have to draw a curved line towards the starting point of the "d" and then you have to reverse direction and follow the line you just drew in backward direction. That is hard to do neatly and is unnessecary.
- Don't draw loops on the letters bfghjkl. Do the same motion, but lift the pen during the first half of the loop. That will not take any longer, but the resulting writing will be much clearer without the redundant forest of loops.
- Write the "r" like the printed r, not mirror-imaged.
Of course, learn the difference between UPPERCASE and lowercase. I don't know about USA habits, but I've seen many people who mix lowercase and uppercase letter shapes.Actually, dishwashing agent is made to taste horrible to discourage children from eating it. You'd be better of with just that magical flavour without the dishwashing stuff around it.
But which term? An earlier discussion showed that alternatives such as "programming enthousiast" and "codesmith" do not really carry the same associations.
Did you account for the fact that 200 W, 8 hours per day means $70 per year in electricity? (@ $0.12 per kWh)
The point size of a font refers to the height of the ancient lead type, where every letter is sitting on a rectangular piece of lead. Some fonts are designed to never reach the borders of that rectangle, i.e. a parenthesis "(" in a 12-point font can actually be a bit less than 12 pt high. Only if a "(" is more than the specified 12pt, then something is really wrong.
Furthermore, 12 points are NOT the same as 1/6 inch. There are actually 72.27 (American printer's) points in an inch, but someday, Adobe decided that for digital typesetting, a round number such as 72 points per inch was easier. (The number 72.27 pt/in is easy to remember, but that is pure coincidence. See point units.)
Though Joe Average might accept the default settings for the search button, webmasters will certainly get very pissed if their webserver prevents them from showing up in the #1 or (shiver) #2 search engine Google... Even Microsoft wouldn't risk loosing the IIS market.
Most keyboards have small foldable feet to decrease the sloping angle. For some reason, most people believe that it is more comfortable with the feet extended, but fold the damn things down if you have wrist problems! Get a wrist pad to decrease the angle even further, especially if you have the bad habit of supporting your wrists on the table. (I can't type without either resting my wrists on the pad/table or sitting in a chair with properly adjusted armrests)
This is particularly true with ergonomic keyboards.
Some years ago I tried the MS natural keyboard when it was new on the market. Though I didn't like it because of how the "click" felt, it was actually sloping the other way around: lifted on the front and lower on the back (where the function keys are). I'm a bit surprised to hear that modern split keyboards do have those evil feet on the back side.