I have installed VMware on a number of Linux distros without any issues. It has been a simple matter of untarring the tar.gz archive and running the install script. You will need gcc and the headers for your current kernel. Most distros have a kernel headers or kernel source package. VMware 4.5 now supports 2.6 kernels. Previous version of VMware required 3rd party patches. If you use one of the supported RedHat, SuSE, or Mandrake distros you probably can install the rpm with even less effort.
I know it is hard for geeks to understand, but there is more to law then what is written down in black and white.
In this civil suit one of the arguments that will be put forward by Air Canada is whether the use of the information was "reasonable." Their argument will probably include examples of similar agrements all in a effort to convince a judge. It is unlikely that there is any document that states how many times a person can log into the site, or what they may use the information found on the site for. These statements are unecessary.
The "reasonable" test goes far beyond what has been written on paper. It appears all over civil and criminal law in every court that has ever been influenced by the British, and probably the other European powers as well. It is a giant catch all in some respects. This test is even found at the heart of modern justice in the phrase "...beyond a reasonable doubt."
Slashdot has reported on many cases where geeks have gotten into trouble when they have assumed that an act was permitted becuase there is no statement preventing said act. This is never the case. In all laws, and in all contracts there is always an implied element of what is reasonable.
This is all fine and good, but all Linux and *BSD users are currently power users. This is not going to change in the immediate future. Gnome developers are free to target which ever market they choose, but ignoreing their only current market (power users) is not going to help in the uptake of Gnome.
Since Windows is too stubborn to run w/o swap, I say throw it back into RAM where it belongs in the first place.
But it won't work properly. Windows will try to swap out the ramdisk onto itself. In the end all that will happen is that the swap file will be full of many copies of itself. Better to have a swapfile on disk that is rarely used. I would surprised if Windows will even let you set the swapfile to be on a ram disk.
No. Here are a few tips to help when reading the reviews.
1) Most owner reviews suck. There are many sites that will post reviews made by people who recently purchased a product. These reviews are rarely objective. For one thing the author will rarely have anything to compare the new product to. For a second thing, far too often the review is merely an editorial in which the author tries to justify to themselves the purchase of some new, and expensive kit. A classic example (and sterotype) is the audiophile "reviews" extoling the features of insanely expensive hardware. However I do find such reviews of some use. If there are any problems listed, those problems should not be taken lightly. If a product is given a 8 of 10 but has <undesirable feature>, then that feature must be pretty serious to warrant mention at all.
2) Read several sources. Off the top of my head I can think of Tomshardware, Sharkyextreme, Anandtech, Arstechnica, and [H]ardOCP. These sites tend to do decent comparisons. The sites each have different methods, and don't always agree on the results. I prefer Tom's and Anand for reviews (despite Tom's past). Arstechnica tends to have good technical articles which serve as a basis to better understand and critically read the other review sites. Dansdata deserves mention as one of the best cpu cooler review site out there.
3) Trust the numbers more then the comentary. It is harder to be biased with numbers. Think critically about the results. If the review states that "product A is clearly faster then product B," but the difference is only 2%, then that comment is not justified. There is caveat to this. Sometimes the review will contain critical details that can't be expressed very easily as a number. Important information about supported features, or architechture. Such as video card X supports DirectX9 completely, whereas video card Y does not. A good review will explain why these features matter.
4) Think about what is really being tested. Read the test methodology. Look at the combination of hardware and/or software used to test the product. Again Dansdata deserves mention. Dan has built an excellent test rig to test the cpu coolers. There are flaws with the test. He acknowledges those flaws, and dicusses how the flaws, may or may not affect the validity of his tests.
5) When reading the conclusions it is very important to understand what those conclusions were based on. A prime example is ConsumerReports' (CU) review of digital cameras versus DPreview or Stevesdigicams. CU is very independant, but rates the cameras as a whole along with the bundled software, and other touchyfeely, ooh-aah features. The other sites ditch the software, and review the cold hard technical details of the camera with heavy emphasis on the image quality.
Australia is one of the only countrys never to have had any wars or bloody revolutions
Not exactly. Vinegar hill was not particularly bloody, or very long. I think the Aboriginals would have a thing or two to say about bloody wars as well. I think this quote from Cam on Kuro5hin sums it up nicely. Especially the second to last sentence.
The fact remains that Australian history has a great deal of open resistance to authority, from the Aboriginal Wars when indigenous Australians fought to keep their land, to the 1804 Rebellion, to the Eureka Stockade and the 1932 face off between NSW and the Commonwealth. These historical events are an integral part of Australian history and deserve not be forgotten or brushed-off because they are unfashionable or might make Australia look bad or disloyal. Fortunately this attitude exists less and less as the Australian nation becomes more comfortable with its self-image.
I hope that you are correct in that Australians won't stand for this extradition. The above quote supports your argument. However, I am not optomistic. I believe Australia's involment in Iraq has a lot to do with the trade negotiations. If the Australian government is willing to pay for a trade agreement with blood, what is one more?
Cisco is great if you are buying the expensive stuff. You can still get the software updates for the old routers. The support has always been good as well. But like you say the support for the smaller boxes is pretty poor. It is not even just the low end stuff. I have been burned by "enterprise" access points as well. It seems end of life happens just a few months after the first production is finished.
Even if Linksys is a sperate entity from Cisco I won't buy one. Linksys has always made crap. A local ISP hates Linksys due to bad fragmentation of PPPoE packets.
Netgear can be hit or miss. Friends and I scored some underpriced Netgear switches (FS105) that have been flawless. Mine has been running nonstop for four years. On the other hand I had 10 Netgear hubs (DSx08) in the lab. A year later atleast half had intermittent failures. I would still buy Netgear over D-Link.
All the D-Link stuff I have used has been obviously cheaply made. Some stuff worked, some didn't, but it was all cheap. The NICs have cheap chips but mostly work. At the other end we have 8port desktop switches. I scrounged over a dozen switches still in the box. When I got them back to the lab all but one of the power supplies were missing. I found out later that the power supplies would get really hot and melt.
The SMC Barricade is another mini-router to consider. I know of two that have been running flawlessly for close to two years. I think there may still be Engineers working at SMC:-)
Paid for services, such as MSDN subscriptions, were down as well. The real news is not that Hotmail was down, but that all Passport based services were having problems. MS has been trying hard to sell Passport as a "single sign on solution." This indicent does not help that marketing effort. This is not the first time that Passport has been out. In the past the passport domain expired and was rescued by a very nice person who registered the domain on a weekend, reinstating the service.
You are correct. There were/are problems with MacOS9. Also the idea of the two operating systems "simply operating differently" is a good one. I should have better worded what I wrote. I added "compared to MacOS9" at the last minute in the hopes of avoiding flames for comparing MacOSX to Gnome, Kde, or WinXP.
Depending on the review site, that motorcycle she rides has a 125hp or 129hp engine with a claimed top speed of 172mph, 277km/h. She writes that her father believes that rideing in 5th and 6th gear is more dangerous then the radiation. He is probably right.
This highlights one of the problems I have with KDE GNOME. Both projects had the misguided idea that a gui that looked like Windows would be easier for users switching from Windows. Ofcourse the opposite is true. If a gui looks like Windows users are going to expect it to act exactly like Windows. When the behaviour is a little different users get frustrated and confused. It would be far better to have a completely different UI that is userfriendly.
That said, I am not sure that Apple switched the UI for reasons of useability. There are so many UI mistakes in OSX compared to MacOS9 that I not sure if Apple was ever thinking about good UI when designing OSX.;-)
how do you continue to produce quality work in the face of constant staff reductions?
You can't.
Constant staff reductions cause a high stress environment. Most people have a hard time preforming their best when stressed[1]. Some workers burn themselves out by working longer hours in a futile attempt to reach goals made unobtainable by the stressfull environment. Narturally, this leads to more stress.
A reverse approach is to dedicate more time to yourself and stress reducing activities. Pick your activities wishly. Your goal is to not simply relax, it is to reduce your stress. Try new activities. Consider trying those activities that you have put off "until you are more settled." Many people find that physcial activity is an excellent way to reduce stress. As a bonus physcial activity will improve your health which helps combat stress. Along the same lines, get a good night's sleep everynight.
By reduceing your overall stress level you will be able to better manage your working stress. You will have more focus. You will be able to produce better work. You may not perform at your peak level, but you will be closer.
[1] Some people work best under pressure. But not under this kind of stress.
Gator provides a service of "autocompleting web forms." While this service is useless, Gator can still claim that monitoring every website the user visits is an important part of insureing that Gator works. E.g to diagnose crash information. This really is not much of a stretch and any lawyer would argue this in court. Remember that logic has no place in a court of law. Gator also dosen't give a damn about "personal information" as deffined by para 4.a.1-7 of the proposed bill. Gator is mostly interested in the urlstats, and the interception of add urls. Again a lawyer will argue that urlstats don't constitute personal information under para 4.a.8. Even if Gator did want your vital stats, Gator could ask for the information and most users would fill in the form. (Especially as Gator would use sleaze ball tactics to make it look like there was no easy way to opt out.) This collection of personal stats could be easily argued as a "registration process to better serve our customers." This is not a hard arguement. A typical warranty registration card is mostly marketing questions.
Understand that I hate Gator and the people who write this kind of software. I don't wish to defend them. However through this little thought experiment we can see that any semi-compotent lawyer will be able to get an aquittal for Gator and similar spyware. If charged under this bill Gator would have as many weasels in court as they could afford. A conviction would be more then just a fine and jail sentence. A conviction would be the end of Gators slimey business model.
The problems you list are typicall of well meaning developers who simply don't understand the full scope of the problem.
And don't try using 640x480 on windows anymore
This is a common problem if you set your font size to "large" and are forced back to vga mode for any reason. I think it has been an annoyance since windows 95.
Agreed. I hate spyware. I hate programs that call home with file listings from harddrives. But a nasty law like this is not the solution. There are even enough clauses in section 3 that would allow most spyware as legal. (Read 3a and 3b.) Similar to anti-spam laws this law is not going to stop the sleaze balls. There are already laws on the books now that could be used to punish spyware.
Worst this is the kind of law that will be used as a trump card to make a criminal out of some kid who causes a little mischief like loading a key logger on someone else's computer. While such an act of mischief should not go unpunnished it is not an aggravated misdemeanor. On of the biggest threats to American freedom is the huge number of laws that can be thrown at someone when prosecutors decide that that someone must be punished.
text isn't as high quality (not that I suppose it would matter in this case)
Text quality can make a difference. I depends on the type and ammount of vision lose.
It is pretty sad that many people will run monitors at low resolutions so that they have larger fonts. MacOSX is nasty as MacOSX assumes 72 dpi. You can't change it. I find this particularly annoying as I type this on a 133dpi LCD. You can change the dpi in Windows, but beyond the "large font", "small font" options it does not seem to do much. X is a total mess as many apps do their own thing. I won't even start on the assholes who insist on useing 13 pixel fonts on web pages. Computers are amazingly powerfull, but I have yet to find a decent OS that can scale all the fonts appropriately.
I have faced a similar challenge. The very sad state of affairs is that a large high quality magnifying glass coupled with a good monitor beat all the other "solutions" I tried. Useing the magnifying glass was frustrateing for the user, but they kept comming back to the it.
I have found that text based tools can work better then GUI based. With text based tools it is easier to blow up the font to 2 inch high letters (and larger). This works well with a minimal window manager with no overlapping windows like ion or ratpoison. GUIs tend to waste too much screen space, especially when you increase the font size. With a magnifying glass the buttons and icons don't need to be huge.
Simple software with few dialogs and popups, and a consistent behaviour is key. My partially sighted client tends to use keyboard navigation and short cuts. They navigate throught the windows and screens from memory. For example to send an email they type: ALT-N "recipient" TAB "subject" TAB "message" CTRL-ENTER.
A single randomly occuring bug is devistateing. I found early versions of MS outlook express to be particularly bad at not always returning to the same state after a given action. The focus caret would sometimes move between the folder list and the email list. This was not a problem I noticed until my client complained.
When helping a partially sighted users continually ask how the display can be better. As a sighted person you really don't know what they need to be able to better read and interact with the computer. Little things like serif, sans serif, or monospaced fonts can make a big difference. As can colours and contrast.
Ofcourse remember that this experience is far more frustrateing for them then it is for you.
Those "useability" tools in OSX and XP suck for the visually imparied. Those tools were written by sighted people unable to comprehend how a person with partial sight would use a computer. Basically the whole windowing/desktop paradigm does not work well.
A large high quality magnifying glass is far better then any of the screen magnifiers. With a magnifiying glass it is possible to use some positional information from the screen because the user can determine where the magnified part of the screen is in relation to the rest.
I have found that text based tools can work better then GUI based. With text based tools it is easier to blow up the font to 2 inch high letters (and larger). This works well with a minimal window manager with no overlapping windows like ion or ratpoison. GUIs tend to waste too much screen space, especially when you increase the font size. With a magnifying glass the buttons and icons don't need to be huge.
There is a screen reader for emacs, but I have not used it. My client is partially deaf as well.
I believe there is some validity to Linus' argument.
In the examples you give the projects are BSD licensed, or the code was released by Apple or BSDi under the BSD license. Some BSD code has been merged into the Linux kernel and is now tainted with the GPL. BSD code has also been taken and used in closed source projects. In both cases the BSDs have lost the ability to merge back any improvments to their code. This is a type of code fork, and a situation Linus wanted to prevent.
The primary authors of the various BSD projects don't have a problem with this situation[1]. Their goal is to generate code that can be usefull to anyone. Linus' goals are obviously different. Hence the use of the GPL.
[1]I have found several BSD authors, and many BSD users to be resentfull towards the GPL and the way it can take BSD code and taint it. As an example see the license used by Darren Reed(??) and the packet filter project. The license attempts to prevent the code from being used in GPL projects. I find this attitude odd. If the authors did not want their code incorporated into a GPL or closed source project, they should not have used the BSD license.
Slashdot Mirror
I have installed VMware on a number of Linux distros without any issues. It has been a simple matter of untarring the tar.gz archive and running the install script. You will need gcc and the headers for your current kernel. Most distros have a kernel headers or kernel source package. VMware 4.5 now supports 2.6 kernels. Previous version of VMware required 3rd party patches. If you use one of the supported RedHat, SuSE, or Mandrake distros you probably can install the rpm with even less effort.
I know it is hard for geeks to understand, but there is more to law then what is written down in black and white.
In this civil suit one of the arguments that will be put forward by Air Canada is whether the use of the information was "reasonable." Their argument will probably include examples of similar agrements all in a effort to convince a judge. It is unlikely that there is any document that states how many times a person can log into the site, or what they may use the information found on the site for. These statements are unecessary.
The "reasonable" test goes far beyond what has been written on paper. It appears all over civil and criminal law in every court that has ever been influenced by the British, and probably the other European powers as well. It is a giant catch all in some respects. This test is even found at the heart of modern justice in the phrase "...beyond a reasonable doubt."
Slashdot has reported on many cases where geeks have gotten into trouble when they have assumed that an act was permitted becuase there is no statement preventing said act. This is never the case. In all laws, and in all contracts there is always an implied element of what is reasonable.
But audiophiles the world over will still pony up the big bucks for wood cone speakers.
I find it hard to go scuba diving indoors. Even if I did have a deep pool I think I would miss the pretty fish and stuff.
This is all fine and good, but all Linux and *BSD users are currently power users. This is not going to change in the immediate future. Gnome developers are free to target which ever market they choose, but ignoreing their only current market (power users) is not going to help in the uptake of Gnome.
Since Windows is too stubborn to run w/o swap, I say throw it back into RAM where it belongs in the first place.
But it won't work properly. Windows will try to swap out the ramdisk onto itself. In the end all that will happen is that the swap file will be full of many copies of itself. Better to have a swapfile on disk that is rarely used. I would surprised if Windows will even let you set the swapfile to be on a ram disk.
Where can I read your reviews?
No. Here are a few tips to help when reading the reviews.
1) Most owner reviews suck. There are many sites that will post reviews made by people who recently purchased a product. These reviews are rarely objective. For one thing the author will rarely have anything to compare the new product to. For a second thing, far too often the review is merely an editorial in which the author tries to justify to themselves the purchase of some new, and expensive kit. A classic example (and sterotype) is the audiophile "reviews" extoling the features of insanely expensive hardware. However I do find such reviews of some use. If there are any problems listed, those problems should not be taken lightly. If a product is given a 8 of 10 but has <undesirable feature>, then that feature must be pretty serious to warrant mention at all.
2) Read several sources. Off the top of my head I can think of Tomshardware, Sharkyextreme, Anandtech, Arstechnica, and [H]ardOCP. These sites tend to do decent comparisons. The sites each have different methods, and don't always agree on the results. I prefer Tom's and Anand for reviews (despite Tom's past). Arstechnica tends to have good technical articles which serve as a basis to better understand and critically read the other review sites. Dansdata deserves mention as one of the best cpu cooler review site out there.
3) Trust the numbers more then the comentary. It is harder to be biased with numbers. Think critically about the results. If the review states that "product A is clearly faster then product B," but the difference is only 2%, then that comment is not justified. There is caveat to this. Sometimes the review will contain critical details that can't be expressed very easily as a number. Important information about supported features, or architechture. Such as video card X supports DirectX9 completely, whereas video card Y does not. A good review will explain why these features matter.
4) Think about what is really being tested. Read the test methodology. Look at the combination of hardware and/or software used to test the product. Again Dansdata deserves mention. Dan has built an excellent test rig to test the cpu coolers. There are flaws with the test. He acknowledges those flaws, and dicusses how the flaws, may or may not affect the validity of his tests.
5) When reading the conclusions it is very important to understand what those conclusions were based on. A prime example is ConsumerReports' (CU) review of digital cameras versus DPreview or Stevesdigicams. CU is very independant, but rates the cameras as a whole along with the bundled software, and other touchyfeely, ooh-aah features. The other sites ditch the software, and review the cold hard technical details of the camera with heavy emphasis on the image quality.
My fingers have a mind of their own today. That word should be incident.
Not exactly. Vinegar hill was not particularly bloody, or very long. I think the Aboriginals would have a thing or two to say about bloody wars as well. I think this quote from Cam on Kuro5hin sums it up nicely. Especially the second to last sentence.
I hope that you are correct in that Australians won't stand for this extradition. The above quote supports your argument. However, I am not optomistic. I believe Australia's involment in Iraq has a lot to do with the trade negotiations. If the Australian government is willing to pay for a trade agreement with blood, what is one more?
Cisco is great if you are buying the expensive stuff. You can still get the software updates for the old routers. The support has always been good as well. But like you say the support for the smaller boxes is pretty poor. It is not even just the low end stuff. I have been burned by "enterprise" access points as well. It seems end of life happens just a few months after the first production is finished.
:-)
Even if Linksys is a sperate entity from Cisco I won't buy one. Linksys has always made crap. A local ISP hates Linksys due to bad fragmentation of PPPoE packets.
Netgear can be hit or miss. Friends and I scored some underpriced Netgear switches (FS105) that have been flawless. Mine has been running nonstop for four years. On the other hand I had 10 Netgear hubs (DSx08) in the lab. A year later atleast half had intermittent failures. I would still buy Netgear over D-Link.
All the D-Link stuff I have used has been obviously cheaply made. Some stuff worked, some didn't, but it was all cheap. The NICs have cheap chips but mostly work. At the other end we have 8port desktop switches. I scrounged over a dozen switches still in the box. When I got them back to the lab all but one of the power supplies were missing. I found out later that the power supplies would get really hot and melt.
The SMC Barricade is another mini-router to consider. I know of two that have been running flawlessly for close to two years. I think there may still be Engineers working at SMC
Paid for services, such as MSDN subscriptions, were down as well. The real news is not that Hotmail was down, but that all Passport based services were having problems. MS has been trying hard to sell Passport as a "single sign on solution." This indicent does not help that marketing effort. This is not the first time that Passport has been out. In the past the passport domain expired and was rescued by a very nice person who registered the domain on a weekend, reinstating the service.
You are correct. There were/are problems with MacOS9. Also the idea of the two operating systems "simply operating differently" is a good one. I should have better worded what I wrote. I added "compared to MacOS9" at the last minute in the hopes of avoiding flames for comparing MacOSX to Gnome, Kde, or WinXP.
Depending on the review site, that motorcycle she rides has a 125hp or 129hp engine with a claimed top speed of 172mph, 277km/h. She writes that her father believes that rideing in 5th and 6th gear is more dangerous then the radiation. He is probably right.
This highlights one of the problems I have with KDE GNOME. Both projects had the misguided idea that a gui that looked like Windows would be easier for users switching from Windows. Ofcourse the opposite is true. If a gui looks like Windows users are going to expect it to act exactly like Windows. When the behaviour is a little different users get frustrated and confused. It would be far better to have a completely different UI that is userfriendly.
;-)
That said, I am not sure that Apple switched the UI for reasons of useability. There are so many UI mistakes in OSX compared to MacOS9 that I not sure if Apple was ever thinking about good UI when designing OSX.
how do you continue to produce quality work in the face of constant staff reductions?
You can't.
Constant staff reductions cause a high stress environment. Most people have a hard time preforming their best when stressed[1]. Some workers burn themselves out by working longer hours in a futile attempt to reach goals made unobtainable by the stressfull environment. Narturally, this leads to more stress.
A reverse approach is to dedicate more time to yourself and stress reducing activities. Pick your activities wishly. Your goal is to not simply relax, it is to reduce your stress. Try new activities. Consider trying those activities that you have put off "until you are more settled." Many people find that physcial activity is an excellent way to reduce stress. As a bonus physcial activity will improve your health which helps combat stress. Along the same lines, get a good night's sleep everynight.
By reduceing your overall stress level you will be able to better manage your working stress. You will have more focus. You will be able to produce better work. You may not perform at your peak level, but you will be closer.
[1] Some people work best under pressure. But not under this kind of stress.
That is a good point.
Gator provides a service of "autocompleting web forms." While this service is useless, Gator can still claim that monitoring every website the user visits is an important part of insureing that Gator works. E.g to diagnose crash information. This really is not much of a stretch and any lawyer would argue this in court. Remember that logic has no place in a court of law. Gator also dosen't give a damn about "personal information" as deffined by para 4.a.1-7 of the proposed bill. Gator is mostly interested in the urlstats, and the interception of add urls. Again a lawyer will argue that urlstats don't constitute personal information under para 4.a.8. Even if Gator did want your vital stats, Gator could ask for the information and most users would fill in the form. (Especially as Gator would use sleaze ball tactics to make it look like there was no easy way to opt out.) This collection of personal stats could be easily argued as a "registration process to better serve our customers." This is not a hard arguement. A typical warranty registration card is mostly marketing questions.
Understand that I hate Gator and the people who write this kind of software. I don't wish to defend them. However through this little thought experiment we can see that any semi-compotent lawyer will be able to get an aquittal for Gator and similar spyware. If charged under this bill Gator would have as many weasels in court as they could afford. A conviction would be more then just a fine and jail sentence. A conviction would be the end of Gators slimey business model.
The problems you list are typicall of well meaning developers who simply don't understand the full scope of the problem.
And don't try using 640x480 on windows anymore
This is a common problem if you set your font size to "large" and are forced back to vga mode for any reason. I think it has been an annoyance since windows 95.
Agreed. I hate spyware. I hate programs that call home with file listings from harddrives. But a nasty law like this is not the solution. There are even enough clauses in section 3 that would allow most spyware as legal. (Read 3a and 3b.) Similar to anti-spam laws this law is not going to stop the sleaze balls. There are already laws on the books now that could be used to punish spyware.
Worst this is the kind of law that will be used as a trump card to make a criminal out of some kid who causes a little mischief like loading a key logger on someone else's computer. While such an act of mischief should not go unpunnished it is not an aggravated misdemeanor. On of the biggest threats to American freedom is the huge number of laws that can be thrown at someone when prosecutors decide that that someone must be punished.
text isn't as high quality (not that I suppose it would matter in this case)
Text quality can make a difference. I depends on the type and ammount of vision lose.
It is pretty sad that many people will run monitors at low resolutions so that they have larger fonts. MacOSX is nasty as MacOSX assumes 72 dpi. You can't change it. I find this particularly annoying as I type this on a 133dpi LCD. You can change the dpi in Windows, but beyond the "large font", "small font" options it does not seem to do much. X is a total mess as many apps do their own thing. I won't even start on the assholes who insist on useing 13 pixel fonts on web pages. Computers are amazingly powerfull, but I have yet to find a decent OS that can scale all the fonts appropriately.
I have faced a similar challenge. The very sad state of affairs is that a large high quality magnifying glass coupled with a good monitor beat all the other "solutions" I tried. Useing the magnifying glass was frustrateing for the user, but they kept comming back to the it.
I have found that text based tools can work better then GUI based. With text based tools it is easier to blow up the font to 2 inch high letters (and larger). This works well with a minimal window manager with no overlapping windows like ion or ratpoison. GUIs tend to waste too much screen space, especially when you increase the font size. With a magnifying glass the buttons and icons don't need to be huge.
Simple software with few dialogs and popups, and a consistent behaviour is key. My partially sighted client tends to use keyboard navigation and short cuts. They navigate throught the windows and screens from memory. For example to send an email they type: ALT-N "recipient" TAB "subject" TAB "message" CTRL-ENTER.
A single randomly occuring bug is devistateing. I found early versions of MS outlook express to be particularly bad at not always returning to the same state after a given action. The focus caret would sometimes move between the folder list and the email list. This was not a problem I noticed until my client complained.
When helping a partially sighted users continually ask how the display can be better. As a sighted person you really don't know what they need to be able to better read and interact with the computer. Little things like serif, sans serif, or monospaced fonts can make a big difference. As can colours and contrast.
Ofcourse remember that this experience is far more frustrateing for them then it is for you.
Those "useability" tools in OSX and XP suck for the visually imparied. Those tools were written by sighted people unable to comprehend how a person with partial sight would use a computer. Basically the whole windowing/desktop paradigm does not work well.
A large high quality magnifying glass is far better then any of the screen magnifiers. With a magnifiying glass it is possible to use some positional information from the screen because the user can determine where the magnified part of the screen is in relation to the rest.
I have found that text based tools can work better then GUI based. With text based tools it is easier to blow up the font to 2 inch high letters (and larger). This works well with a minimal window manager with no overlapping windows like ion or ratpoison. GUIs tend to waste too much screen space, especially when you increase the font size. With a magnifying glass the buttons and icons don't need to be huge.
There is a screen reader for emacs, but I have not used it. My client is partially deaf as well.
Insure that ECN is not enabled on the Linux box.
I believe there is some validity to Linus' argument.
In the examples you give the projects are BSD licensed, or the code was released by Apple or BSDi under the BSD license. Some BSD code has been merged into the Linux kernel and is now tainted with the GPL. BSD code has also been taken and used in closed source projects. In both cases the BSDs have lost the ability to merge back any improvments to their code. This is a type of code fork, and a situation Linus wanted to prevent.
The primary authors of the various BSD projects don't have a problem with this situation[1]. Their goal is to generate code that can be usefull to anyone. Linus' goals are obviously different. Hence the use of the GPL.
[1]I have found several BSD authors, and many BSD users to be resentfull towards the GPL and the way it can take BSD code and taint it. As an example see the license used by Darren Reed(??) and the packet filter project. The license attempts to prevent the code from being used in GPL projects. I find this attitude odd. If the authors did not want their code incorporated into a GPL or closed source project, they should not have used the BSD license.