I swapped to dvorak a few years ago, and it took a couple of weeks of not being able to type effecively. Once I learnt the Dvorak, I've never been able to touch type QWERTY again, but then I've never tried to maintain the skill either. I don't think that Dvorak is the cure to all RSI. It's more efficient, and you can pump more words out for a given amount of finger movement, but at the end of the day, the fingers are doing much the same types of movements.
I think the biggest improvement that the Dvorak switch gave to me was a consciousness of typing, and an opportunity to learn again properly. The fact that there are QWERTY caps on the board forces you to never look at the keys and that really builds the foundation of doing it right. I think the very fact that there were the "wrong" letters printed on the keys had the indirect effect of pushing me into a better posture and hand positioning.
So maybe the RSI benefits observed in going to Dvorak are partly indirect, and can be obtained just by better concentration on hand posture and sitting posture, without going to the new layout. If you are in a position of having to use QWERTY at times, that may offer the better compromise.
Also, don't discount the power of switching to a different make of keyboard, and also swapping which thumb you use to press the space bar - and if you use a mouse - which hand moves the mouse. Those things can also have a significant effect.
The big Metacity problem: vertical window movement
on
Gnome 2.10 Released
·
· Score: 1
If you are worried about no window snapping, Gnome has a much bigger problem in that. With the default Metacity window mananger, you can't move the top of windows up off the top of the screen at all with alt-drag, which leaves it all feeling quite claustrophobic. It's been that way for a long time now, and no sign of it getting fixed.
I've been using sawfish to get around that issue, but sawfish seems unmaintained now. Anyone got ideas of a different WM to drop in in order to get window movement in all directions?
W2K style is a real pain, and to me it seems worth installing a few hundred MB of cygwin to get the bash behaviour! I'd be more interested in knowing how to get Bash behaviour in W2K. Any ideas?
Which does a MS style (shudder) history search on F8 and F7 - it sounds like it does exactly as you mention. With a change of escape code, you could adjust the key binding.
The programming in Skype is not a huge undertaking, and is the sort of thing that can be done in the open source model. There is gnomemeeting and linphone that do similar jobs in the Linux world.
Also remember speakfreely
It has been going for 12 years or so, and is open-source cross-platform unix/windows. Recently work on it has revived, and it now uses the speex codec. (some details here: http://www.2pi.info/software/sf_speex/) It is a project in need of more developers though. So it's a project to think of where some positive work can be done rather than complaining that some other project doesn't have the source.
The thing is that Speakfreely does Linux--Windows with crypto, an efficient codec (speex), and some NAT traversal right now. I don't know of an working alternative. Do you know any other combination that will even do linuix-windows over a 33k connection now? I can only think of the huuuge open-h323, and my experience is that it doesn't perfom anywhere near as well with less-than-ideal connections.
John Walker is playing it on the safe side, and just warning users that he can no longer guarantee support as he will not be providing it himself. It is fairly mature software though, and doesn't need much updating with time, so that's why there hasn't been much development over the past few years.
Since John has withdrawn from development though, developers have been working on the NAT issue, and have a solution for many circumstances. Also the Speex codec has been added, so the quality/bitrate is now back in the league of the alternatives. So basically, it doesn't need much to keep it up to date.
I'd have to agree with the cultural aspect. Years ago, I thought that shareware was a good concept, but then migrated to the GNU world. Now it seems natural for me to freely distribute anything that I write. If ever I go looking for software utilities for windows, I then wonder if it's all a joke. Asking money for software is no problem, but it's so hard to find good quality software amongst all the junk with excessive childish graphics and advertising. In conparison, anything from the GNU world seems aimed more at getting the job done. I think that must be a cultural thing.
> Suppose we threw out all the chips - went back to pencil and paper? How many kilowatt-hours would we consume in heating and lighting the rooms full of green-hatted accountants scratching figures onto paper with pens?
The points are valid, but awareness of the envirnomentat cost must still be a good thing, mainly in terms of end-user responsibility. It can mean keeping hardware longer between upgrade cycles, but still keeping up with new technology over the long-term. It can also mean making new hardware with less power, but still using new tech. Some applications may only need a single-chip computer with 486 power, rather than throwing the "standard" latest PC at it.
>For running Office, maybe a '486 would be OK. Forget about Doom III, though. Or rendering Lord of the Rings.
Do we really need Doom III or Lord of the Rings though? Wouldn't it be sensible to to at least factor the environmental cost into the playing and production of these things? It's like "free" network bandwidth - if they don't see the cost, people will waste it on things they don't really get value out of.
I was being a bit provocative there with the EVFs. I agree that they're not generally better now, but I say "inherently" because it is a simpler and cleaner approach, and offers the ability to show what the camera is going to give you rather that what you can see. I do admit that the technology isn't really there yet though to get a good resolution. If only they would ditch those big useless LCDs, and put a really high res EVF in...
On an EVF it's great to have all the camera setting info available in the display, histograms, image review, etc True, the downsides are battery life and resolution, but they should be shrinking obstacles.
"Big-tuff-me can do manual with my SLR".. Really, most digital cameras in the $US500+ range do do all those manual modes: manual exposure, aperture/shutter priority, manual white balance, manual iso choice, manual focus. There are a lot, Canon G2, Olympus c730... And why reflex? - surely an electronic viewfinder is inherently a better alternative, and the quality levels are getting closer.
The problem with your idea of lenses is that to use a lens off a 35mm film camera would lose one of the big advantages of digital: the small sensor area. The small sensors mean that they can use much smaller and lighter lenses, but have the same effective focal length and aperture, and even a better depth-of-field.
Then when people start using lenses designed for 35mm they would find the quality lousy anyhow because they are not suited to the smaller ccd. Want a big ccd then? Why lose this great opportunity to shrink the standard and make lenses smaller?
No, you can't hide behind the "I'm a legend because I know how to use an SLR" front.
The main problem with digital, at least in the consumera and "prosumer" lever equipment is the dynamic range. 8-bit depth doesn't give much room to move in high contrast situations, and you tend to end up stuck with clipping to white.
The solution will be 12 or 16 bit colour depths, but what do we use at the consumer end for a file format? Is there any good alternative to jpg with higher bit depths but without all sorts of patent woes?
I must disagree about the claim that you can't throw away half the data from the fft of real data. All signals that come out of my microphone are real in the time domain so the frequency domain spectrum is symmetric. In other words, you can perfectly reconstruct the signal just from the frequency information between 0Hz and the Nyquist Frequency. So if you start off with N real data points, you only need N/2 complex points to perfectly reconstruct it.
This property is used to speed up FFTs, by performing a real N point transform using a complex N/2 point transform algorithm. Again there is no loss of information at all - except due to the roundoff errors inherent in the transform.
I agree that an FFT is not a magic solution to voice compression, but it can be a start to some form of compression. For voiced speech, the signal can be well represented as the sum of a relatively small number of sinewaves (20-40) which are effectively constant over a 10ms temporal window. Using an FFT to calculate this information can lead to effective compression techniques. That's only an example, not a suggestion that that is a good technique.
Slashdot Mirror
I swapped to dvorak a few years ago, and it took a couple of weeks of not being able to type effecively. Once I learnt the Dvorak, I've never been able to touch type QWERTY again, but then I've never tried to maintain the skill either. I don't think that Dvorak is the cure to all RSI. It's more efficient, and you can pump more words out for a given amount of finger movement, but at the end of the day, the fingers are doing much the same types of movements.
I think the biggest improvement that the Dvorak switch gave to me was a consciousness of typing, and an opportunity to learn again properly. The fact that there are QWERTY caps on the board forces you to never look at the keys and that really builds the foundation of doing it right. I think the very fact that there were the "wrong" letters printed on the keys had the indirect effect of pushing me into a better posture and hand positioning.
So maybe the RSI benefits observed in going to Dvorak are partly indirect, and can be obtained just by better concentration on hand posture and sitting posture, without going to the new layout. If you are in a position of having to use QWERTY at times, that may offer the better compromise.
Also, don't discount the power of switching to a different make of keyboard, and also swapping which thumb you use to press the space bar - and if you use a mouse - which hand moves the mouse. Those things can also have a significant effect.
If you are worried about no window snapping, Gnome has a much bigger problem in that. With the default Metacity window mananger, you can't move the top of windows up off the top of the screen at all with alt-drag, which leaves it all feeling quite claustrophobic. It's been that way for a long time now, and no sign of it getting fixed.
I've been using sawfish to get around that issue, but sawfish seems unmaintained now. Anyone got ideas of a different WM to drop in in order to get window movement in all directions?
W2K style is a real pain, and to me it seems worth installing a few hundred MB of cygwin to get the bash behaviour! I'd be more interested in knowing how to get Bash behaviour in W2K. Any ideas?
in my /etc/inputrc (or ~/.inputrc) I have:
"\e[18~": history-search-forward
"\e[19~": history-search-backward
Which does a MS style (shudder) history search on F8 and F7 - it sounds like it does exactly as you mention. With a change of escape code, you could adjust the key binding.
Also remember speakfreely It has been going for 12 years or so, and is open-source cross-platform unix/windows. Recently work on it has revived, and it now uses the speex codec. (some details here: http://www.2pi.info/software/sf_speex/) It is a project in need of more developers though. So it's a project to think of where some positive work can be done rather than complaining that some other project doesn't have the source.
The thing is that Speakfreely does Linux--Windows with crypto, an efficient codec (speex), and some NAT traversal right now. I don't know of an working alternative. Do you know any other combination that will even do linuix-windows over a 33k connection now? I can only think of the huuuge open-h323, and my experience is that it doesn't perfom anywhere near as well with less-than-ideal connections.
John Walker is playing it on the safe side, and just warning users that he can no longer guarantee support as he will not be providing it himself. It is fairly mature software though, and doesn't need much updating with time, so that's why there hasn't been much development over the past few years.
Since John has withdrawn from development though, developers have been working on the NAT issue, and have a solution for many circumstances. Also the Speex codec has been added, so the quality/bitrate is now back in the league of the alternatives. So basically, it doesn't need much to keep it up to date.
http://www.fourmilab.ch/wb/speak-freely.pl?read=50 1
http://www.fourmilab.ch/wb/speak-freely.pl?read=50 9
I'd have to agree with the cultural aspect. Years ago, I thought that shareware was a good concept, but then migrated to the GNU world. Now it seems natural for me to freely distribute anything that I write.
If ever I go looking for software utilities for windows, I then wonder if it's all a joke. Asking money for software is no problem, but it's so hard to find good quality software amongst all the junk with excessive childish graphics and advertising. In conparison, anything from the GNU world seems aimed more at getting the job done. I think that must be a cultural thing.
> Suppose we threw out all the chips - went back to pencil and paper? How many kilowatt-hours would we consume in heating and lighting the rooms full of green-hatted accountants scratching figures onto paper with pens?
The points are valid, but awareness of the envirnomentat cost must still be a good thing, mainly in terms of end-user responsibility. It can mean keeping hardware longer between upgrade cycles, but still keeping up with new technology over the long-term. It can also mean making new hardware with less power, but still using new tech. Some applications may only need a single-chip computer with 486 power, rather than throwing the "standard" latest PC at it.
>For running Office, maybe a '486 would be OK. Forget about Doom III, though. Or rendering Lord of the Rings.
Do we really need Doom III or Lord of the Rings though? Wouldn't it be sensible to to at least factor the environmental cost into the playing and production of these things? It's like "free" network bandwidth - if they don't see the cost, people will waste it on things they don't really get value out of.
I was being a bit provocative there with the EVFs. I agree that they're not generally better now, but I say "inherently" because it is a simpler and cleaner approach, and offers the ability to show what the camera is going to give you rather that what you can see. I do admit that the technology isn't really there yet though to get a good resolution. If only they would ditch those big useless LCDs, and put a really high res EVF in...
On an EVF it's great to have all the camera setting info available in the display, histograms, image review, etc True, the downsides are battery life and resolution, but they should be shrinking obstacles.
"Big-tuff-me can do manual with my SLR" .. Really, most digital cameras in the $US500+ range do do all those manual modes: manual exposure, aperture/shutter priority, manual white balance, manual iso choice, manual focus. There are a lot, Canon G2, Olympus c730... And why reflex? - surely an electronic viewfinder is inherently a better alternative, and the quality levels are getting closer.
The problem with your idea of lenses is that to use a lens off a 35mm film camera would lose one of the big advantages of digital: the small sensor area. The small sensors mean that they can use much smaller and lighter lenses, but have the same effective focal length and aperture, and even a better depth-of-field.
Then when people start using lenses designed for 35mm they would find the quality lousy anyhow because they are not suited to the smaller ccd. Want a big ccd then? Why lose this great opportunity to shrink the standard and make lenses smaller?
No, you can't hide behind the "I'm a legend because I know how to use an SLR" front.
The main problem with digital, at least in the consumera and "prosumer" lever equipment is the dynamic range. 8-bit depth doesn't give much room to move in high contrast situations, and you tend to end up stuck with clipping to white.
The solution will be 12 or 16 bit colour depths, but what do we use at the consumer end for a file format? Is there any good alternative to jpg with higher bit depths but without all sorts of patent woes?
I must disagree about the claim that you can't throw away half the data from the fft of real data. All signals that come out of my microphone are real in the time domain so the frequency domain spectrum is symmetric. In other words, you can perfectly reconstruct the signal just from the frequency information between 0Hz and the Nyquist Frequency. So if you start off with N real data points, you only need N/2 complex points to perfectly reconstruct it.
This property is used to speed up FFTs, by performing a real N point transform using a complex N/2 point transform algorithm. Again there is no loss of information at all - except due to the roundoff errors inherent in the transform.
I agree that an FFT is not a magic solution to voice compression, but it can be a start to some form of compression. For voiced speech, the signal can be well represented as the sum of a relatively small number of sinewaves (20-40) which are effectively constant over a 10ms temporal window. Using an FFT to calculate this information can lead to effective compression techniques. That's only an example, not a suggestion that that is a good technique.