Had a look on my MacBook Pro. I could definitely see faint colour casts in some of the grey bars. As for dithering I was not convinced, but after staring at it for a while I thought I could perhaps see some regular speckled patterns in the shape of regularly sized diamonds. Might have been my imagination for staring for too long...
Parent poster describes how with a restricted number of colours, though the human eye can't see banding in the original, if you then process that image the quantisation of colours in the original can lead to banding in the processed version.
Yes, but that has nothing at all to do with this story.
There would be an issue if a 24-bit image was downsampled to 16-bit (for example) in order to display it on a 16-bit screen and then resaved (or processed) at the lower bit-depth. But it's not. Any transformations done in the image editing program are at the bit depth of the image, not of the display.
So if the eye can't see the deficiencies of the display before manipulating the image, it won't see them afterwards either.
My understanding is that CDs go up to 20 kHz... but that the bit-depth is somewhat course at that range.
You are probably thinking of 'one-bit' (or bitstream) digital to analogue converters. (Wikipedia article.) It gets around the problem of producing 16 bits of resolution with a single bit by switching at a frequency many times that of the sampling frequency and averaging over time.
In its purest form, it would switch at 2^16, or 65536 times the CD sample frequency. If one CD sample value is 0, the DAC would be off for 65536 DAC output samples. If the CD sample value is 65535 it is on for 65536 DAC output samples. For intermediate values it is on for the given proportion of the time. In other words, the CD sample value determines the duty-cycle of the output from the DAC. The one-bit on/off output is then averaged over time. This results in a conversion with almost no non-linear distortion of the signal.
Unfortunately a frequency of 65536 * 44.1 kHz would be in the THz range, so the actual frequency that a 1 bit DAC operates at is somewhat lower. For lower frequency audio signals the averaging process is still very accurate, but it loses some accuracy for the highest frequency audio tones mostly when there are rapid transients in the high frequencies. You might refer to this as a 'coarsening of the bit-depth'.
A full 16-bit DAC doesn't suffer from this problem because each sample from the CD is converted straight into a voltage proportional to that sample value in a single step. But it is very difficult to make a completely linear 16-bit DAC, so the non-lineararity of the DAC introduces its own distortions. But these distortions do not depend on frequency as they do with a 1-bit DAC.
Tell me this is a parody, someone please! Please, won't someone? It is a parody, isn't it? I mean, surely not even in America... Come on, someone... it's gotta be a parody, right?...
I had finished with my account on 'Stokes', and had logged in remotely from 'Hamilton' to delete everything in my home directory. I was just about to type 'rm -rf *' when a collegue called me. I turned round and had a short discussion, turned back to the terminal, and typed the fatal command.
Milliseconds later I re-read what was on the screen:
That won't work for the same reason that the current Windows security model doesn't work.
It's too much trouble.
I believe this is one of the main reasons why UNIX applications generally do not play fast and loose with permissions. The security model is very simple. A process is owned by the account most suitable for the role it will perform. There's no need for complicated LPSECURITY_ATTRIBUTES structures. (And yes, I do think that even those are too complicated for most purposes, so you can guess what I think of the more esoteric aspects of Windows security tokens.)
Be honest, if you program for Win32, how many times have you just passed NULL as the first parameter of CreateEvent()?
If you want to make people do the Right Thing, make the Right Thing easier to do than the Wrong Thing.
It's not that complicated. As the gap between two cars gets smaller so the risk of the following car running into the back of the leading car increases. If I am driving the car behind, I get to choose how big to make that gap, so I can control the risk of collision. If I make the gap big enough, say 2 or 3 seconds, then I can all but eliminate the risk.
If I am in front then I have no control over the size of the gap. If I have to brake unexpectedly I'm trusting to luck that the nut job behind won't run into me.
Since most collisions on the roads are caused by drivers not leaving a big enough gap, taking control of the size of that gap seems like a pretty smart move to me.
That assumes that you can go faster to open up a gap between you and nut job behind. In real life there are at least three reasons why you usually can't do this:
1. You are already driving at the speed limit 2. There is a procession of traffic in front of you 3. When you speed up, nut job behind also speeds up
My car doesn't have cruise control, but I have heard from a number of people whose cars do have crusie control that quite often their fuel consumption comes down slightly when they turn off the cruise control and use their own foot to regulate the fuel. I can't think of a good reason why this should be so. Has anyone else noticed this?
They will honk, swerve in and out of traffic to get around you
Then let them. I'd much rather have a dangerous driver in front of me so that I am in control of the gap than have them behind me where there is very little I can do. When I wave someone past to overtake me, it's usually a big vote of 'no confidence' in their driving.
However, I am worried that outlawing non-neutrality would also squash the ISPs' ability to do legitimate traffic shaping.
I think you are confusing net neutrality with traffic shaping. They are not the same, although there are plenty of interested parties who would like us to think that they are.
I believe that the argument for traffic shaping is accepted by most people. Prioritise real-time traffic, throttle bandwidth-sucking low-priority traffic such as file-sharing. The only people who object to this are the few who really expect to be able to suck up their entire 8Mbps connection with 24-hours-a-day BitTorrents.
The real debate now (and the important one that must be won for net neutrality) is over prioritising traffic according to its source (rather than its type). The reason this is so insidious is that networks could do deals with large content providers to ensure that streaming video downloads fast from those providers, but is so slow as to be unusable from small independents. This directly undermines one of the most important values of the internet.
There is certainly one thing that you do get, that many other people are missing, and the ISPs and others with a big financial stake would dearly love everyone else to miss.
That is that when people talk about net neutrality there are two different things they might be talking about:
1. Differentiating between packets based on packet type/protocol. This is already done and most people think it is a good thing.
2. Differentiating between packets based on where they came from, or where they are going to.
The big companies who argue against net neutrality say that we can't have net neutrality, because (1) is absolutely essential to keep your VOIP calls glitch-free when capacity is limited.
What they don't like to mention is that actually the reason they don't like net neutrality is because they want to make deals with selected networks and content providers to extract money from them in return for giving their data higher priority.
Also fixed-width is very important too. It's harder to read wide columns.
Better is max-width. Reading wide columns is tiring, but sideways scrolling to read a fixed width column on a very narrow screen is much worse.
Most browsers accept a max-width CSS rule. For IE you can either write a special fixed-width rule, or use a tiny bit of Javascript:
// This script makes the 'page' element have a maximum width // of 55ems. We use this instead of max-width because // Internet Explorer doesn't yet support max-width. var d = document; function resizeBox() {
sObj = d.getElementById("page").style;
wem = d.getElementById("em").offsetWidth;
if (d.body.clientWidth) (d.body.clientWidth/wem>55) ? sObj.width="53em" : sObj.width="96%"; } // Gecko-based browsers understand max-width, so don't need this. if (navigator.userAgent.toLowerCase().indexOf("gecko" )==-1) {
onload = resizeBox;
onresize = resizeBox; }
(Sorry, my JavaScript skills are minimal to say the least. All suggested improvements very welcome.)
Well, I heard the interview this evening on the PM programme, and the police representative said that basically if your wireless network is wide open and as a consequence is used by someone else for criminal purposes, then you will suffer a great deal of inconvenience at the very least.
Your equipment will be confiscated for forensic analysis and it will be a while before you get it back. The analysis will probably show that it wasn't you - but do you really want to take that risk, and do you want to be without your computer and broadband access for six months while they establish that?
What we really need is either a language that can express things in such a way that the compiler can easily make good decisions about what can be parallelized
You mean Fortran 90?
Seriously, several constructs in Fortran are designed specifically for parallel execution. The language itself makes it hard to write code that the compiler can't heavily optimise. There's a reason why variable aliasing is strongly controlled in Fortran and why function parameters have an 'intent' attribute. Then there are constructs such as WHERE, which is by its very nature implicitly a parallel set of operations.
Of course, what you are pointing out is the basic flaw with the whole 'net neutrality' argument. It's not a public network, per se. It's owned and opperated by someone. They have the right and privledge to impose what ever restrictions they want on people.
Yes - but that's missing the point of the net neutrality argument. Rather than just accept the status quo, shouldn't we be asking: "Is it a good thing (that the operators can make their own rules)?"
Presumably, those who want net neutrality think this is not a good thing, and that there should be some kind of regulation to stop the owners and operators from favouring one content provider over another.
Net Neutrality isn't really about prioritization... it's about money. ISPs QoS the traffic, they just don't (yet) charge for certain tiers.
True, but it is a little more than that. The interests who are opposed to net neutrality are mainly concerned with being able to prioritise the traffic from partner sites and networks - regardless of the type of traffic. So their own streaming video, or streaming video from a partner will get a much higher priority than the same streaming video if it is coming from a small independent content provider. This is the pernicious thing about opposition to net neutrality. And they cover it in FUD by deliberately conflating the issue of QoS traffic shaping based on traffic type (which most people don't have a problem with) and traffic shaping based on source.
But the sonic boom comes from the displacement of air when that displacement is faster than a pressure wave can be propegated. Since the rotation of the wheels does not displace any air, there's no shock wave.
There already exists a car that can get 65 mpg - routinely - not on some secret test track. It has been available for about six years, and there are millions already on the roads in Europe. It's called the Renault Clio dCI and I have one. It is about the same size as any other super-mini and has a four star (out of five) European safety rating. It's quite lively - pulls away quickly and goes well over 80 mph. The Clio is not the only car of its class - there are others with similar performance and specification.
Why is this remarkable? It is not.
The only remarkable thing is that more people don't seem to know about this. Until fuel prices start to reflect the true cost of motoring, many people seem to prefer to bury their heads in the sand and continue to drive their gas-guzzling monsters.
And the X-prize? It sounds as though it shouldn't be too hard to hit that 100 mpg figure. The real challenge is the change of perception required from the public.
the smallest computer is only limited by the size of the keyboard
I agree (although it's nice to have a decent-sized screen too).
But I would have thought that this might finally provide the impetus for alternative keyboards. I can imagine a handheld PC with sculpted handgrips containing switches for a chord keyboard. After the initial effort of learning a new type of keyboard, data entry could be nearly as fast with a full-sized qwerty keyboard. And there would be no danger of dropping the PC while reaching for awkward keystrokes.
Would people go for this sort of thing? Of course many would be put off by the initial effort of learning the key combinations. The problem with this kind of keyboard is you really have to learn it. You can't 'hunt and peck'. But I think that anyone seriously interested in a miniature PC might just bite.
Slashdot Mirror
Had a look on my MacBook Pro. I could definitely see faint colour casts in some of the grey bars. As for dithering I was not convinced, but after staring at it for a while I thought I could perhaps see some regular speckled patterns in the shape of regularly sized diamonds. Might have been my imagination for staring for too long...
Parent poster describes how with a restricted number of colours, though the human eye can't see banding in the original, if you then process that image the quantisation of colours in the original can lead to banding in the processed version.
Yes, but that has nothing at all to do with this story.
There would be an issue if a 24-bit image was downsampled to 16-bit (for example) in order to display it on a 16-bit screen and then resaved (or processed) at the lower bit-depth. But it's not. Any transformations done in the image editing program are at the bit depth of the image, not of the display.
So if the eye can't see the deficiencies of the display before manipulating the image, it won't see them afterwards either.
I think parent is referring to the Decoy Effect. It's a bit more subtle - the article is well-worth reading.
You are probably thinking of 'one-bit' (or bitstream) digital to analogue converters. (Wikipedia article.) It gets around the problem of producing 16 bits of resolution with a single bit by switching at a frequency many times that of the sampling frequency and averaging over time.
In its purest form, it would switch at 2^16, or 65536 times the CD sample frequency. If one CD sample value is 0, the DAC would be off for 65536 DAC output samples. If the CD sample value is 65535 it is on for 65536 DAC output samples. For intermediate values it is on for the given proportion of the time. In other words, the CD sample value determines the duty-cycle of the output from the DAC. The one-bit on/off output is then averaged over time. This results in a conversion with almost no non-linear distortion of the signal.
Unfortunately a frequency of 65536 * 44.1 kHz would be in the THz range, so the actual frequency that a 1 bit DAC operates at is somewhat lower. For lower frequency audio signals the averaging process is still very accurate, but it loses some accuracy for the highest frequency audio tones mostly when there are rapid transients in the high frequencies. You might refer to this as a 'coarsening of the bit-depth'.
A full 16-bit DAC doesn't suffer from this problem because each sample from the CD is converted straight into a voltage proportional to that sample value in a single step. But it is very difficult to make a completely linear 16-bit DAC, so the non-lineararity of the DAC introduces its own distortions. But these distortions do not depend on frequency as they do with a 1-bit DAC.
Tell me this is a parody, someone please! Please, won't someone? It is a parody, isn't it? I mean, surely not even in America... Come on, someone... it's gotta be a parody, right?...
I had finished with my account on 'Stokes', and had logged in remotely from 'Hamilton' to delete everything in my home directory. I was just about to type 'rm -rf *' when a collegue called me. I turned round and had a short discussion, turned back to the terminal, and typed the fatal command.
Milliseconds later I re-read what was on the screen:
That won't work for the same reason that the current Windows security model doesn't work.
It's too much trouble.
I believe this is one of the main reasons why UNIX applications generally do not play fast and loose with permissions. The security model is very simple. A process is owned by the account most suitable for the role it will perform. There's no need for complicated LPSECURITY_ATTRIBUTES structures. (And yes, I do think that even those are too complicated for most purposes, so you can guess what I think of the more esoteric aspects of Windows security tokens.)
Be honest, if you program for Win32, how many times have you just passed NULL as the first parameter of CreateEvent()?
If you want to make people do the Right Thing, make the Right Thing easier to do than the Wrong Thing.
It's not that complicated. As the gap between two cars gets smaller so the risk of the following car running into the back of the leading car increases. If I am driving the car behind, I get to choose how big to make that gap, so I can control the risk of collision. If I make the gap big enough, say 2 or 3 seconds, then I can all but eliminate the risk.
If I am in front then I have no control over the size of the gap. If I have to brake unexpectedly I'm trusting to luck that the nut job behind won't run into me.
Since most collisions on the roads are caused by drivers not leaving a big enough gap, taking control of the size of that gap seems like a pretty smart move to me.
That assumes that you can go faster to open up a gap between you and nut job behind. In real life there are at least three reasons why you usually can't do this:
1. You are already driving at the speed limit
2. There is a procession of traffic in front of you
3. When you speed up, nut job behind also speeds up
My car doesn't have cruise control, but I have heard from a number of people whose cars do have crusie control that quite often their fuel consumption comes down slightly when they turn off the cruise control and use their own foot to regulate the fuel. I can't think of a good reason why this should be so. Has anyone else noticed this?
I believe that the argument for traffic shaping is accepted by most people. Prioritise real-time traffic, throttle bandwidth-sucking low-priority traffic such as file-sharing. The only people who object to this are the few who really expect to be able to suck up their entire 8Mbps connection with 24-hours-a-day BitTorrents.
The real debate now (and the important one that must be won for net neutrality) is over prioritising traffic according to its source (rather than its type). The reason this is so insidious is that networks could do deals with large content providers to ensure that streaming video downloads fast from those providers, but is so slow as to be unusable from small independents. This directly undermines one of the most important values of the internet.
There is certainly one thing that you do get, that many other people are missing, and the ISPs and others with a big financial stake would dearly love everyone else to miss.
That is that when people talk about net neutrality there are two different things they might be talking about:
1. Differentiating between packets based on packet type/protocol. This is already done and most people think it is a good thing.
2. Differentiating between packets based on where they came from, or where they are going to.
The big companies who argue against net neutrality say that we can't have net neutrality, because (1) is absolutely essential to keep your VOIP calls glitch-free when capacity is limited.
What they don't like to mention is that actually the reason they don't like net neutrality is because they want to make deals with selected networks and content providers to extract money from them in return for giving their data higher priority.
Most browsers accept a max-width CSS rule. For IE you can either write a special fixed-width rule, or use a tiny bit of Javascript: (Sorry, my JavaScript skills are minimal to say the least. All suggested improvements very welcome.)
The second story (the new one) concerned two people who were cautioned for using people's wi-fi broadband internet connections without permission.
Well, I heard the interview this evening on the PM programme, and the police representative said that basically if your wireless network is wide open and as a consequence is used by someone else for criminal purposes, then you will suffer a great deal of inconvenience at the very least.
Your equipment will be confiscated for forensic analysis and it will be a while before you get it back. The analysis will probably show that it wasn't you - but do you really want to take that risk, and do you want to be without your computer and broadband access for six months while they establish that?
Seriously, several constructs in Fortran are designed specifically for parallel execution. The language itself makes it hard to write code that the compiler can't heavily optimise. There's a reason why variable aliasing is strongly controlled in Fortran and why function parameters have an 'intent' attribute. Then there are constructs such as WHERE, which is by its very nature implicitly a parallel set of operations.
Presumably, those who want net neutrality think this is not a good thing, and that there should be some kind of regulation to stop the owners and operators from favouring one content provider over another.
I'm surprised no one has mentioned the AOL Throne yet.
But the sonic boom comes from the displacement of air when that displacement is faster than a pressure wave can be propegated. Since the rotation of the wheels does not displace any air, there's no shock wave.
Except that it isn't.
There already exists a car that can get 65 mpg - routinely - not on some secret test track. It has been available for about six years, and there are millions already on the roads in Europe. It's called the Renault Clio dCI and I have one. It is about the same size as any other super-mini and has a four star (out of five) European safety rating. It's quite lively - pulls away quickly and goes well over 80 mph. The Clio is not the only car of its class - there are others with similar performance and specification.
Why is this remarkable? It is not.
The only remarkable thing is that more people don't seem to know about this. Until fuel prices start to reflect the true cost of motoring, many people seem to prefer to bury their heads in the sand and continue to drive their gas-guzzling monsters.
And the X-prize? It sounds as though it shouldn't be too hard to hit that 100 mpg figure. The real challenge is the change of perception required from the public.
But I would have thought that this might finally provide the impetus for alternative keyboards. I can imagine a handheld PC with sculpted handgrips containing switches for a chord keyboard. After the initial effort of learning a new type of keyboard, data entry could be nearly as fast with a full-sized qwerty keyboard. And there would be no danger of dropping the PC while reaching for awkward keystrokes.
Would people go for this sort of thing? Of course many would be put off by the initial effort of learning the key combinations. The problem with this kind of keyboard is you really have to learn it. You can't 'hunt and peck'. But I think that anyone seriously interested in a miniature PC might just bite.