No tech has complete coverage and is unlikely to any time soon, you would be forced to support multiple options in any case. I would probably host the videos on youtube and link to them - let youtube work out what codecs to use etc.
What about suing them for patents not covered by MPEG-LA ? Don't google own a number of video encoding related patents having acquired On2? There may well be other patents out there too...
I can already have HTML5 video support in linux distros, for free, without having to wait for adobe thanks to ogg theora and webm... I can even have this support in linux/arm, 64bit windows, linux/mips, netbsd, freebsd, amigaos, solaris, irix, beos or a new platform of my own creation should i need to... With flash i am forced to use the very limited set of platforms which adobe supports.
Incidentally, HTML5 video is new which is why you have a poor experience with it, nothing has really been optimized yet... I doubt it will be long before it outperforms flash, it already does on OSX with safari by quite a considerable margin.
What Adobe are basically saying, is that a single proprietary monoculture (flash) is better than multiple slightly incompatible implementations (browsers)...
I have lots of devices here which can access HTML, but considerably less which support flash... There are open source implementations of HTML but pretty much only one closed source implementation of flash, so i can use html on niche systems like haiku, netbsd, amigaos etc while i can't use flash because adobe doesn't bother to support such niche platforms. A few years ago i used an sgi and an alphastation as my primary workstations, neither of which had flash support.
You have to make both 64 and 32bit versions for other platforms too, it's just that most mac/windows devs simply don't bother, 64bit linux is far more mature than 64bit windows or osx.
You get strange incompatibilities between different windows or osx versions too.. for instance, a portable windows app tool i used a while ago (bundles an application into a standalone binary you can run from a usb stick) works on vista but not on 7...
You could also compile an ARM version for linux or windows ce...
With linux you always have the option of distributing source code and letting third parties (ie distro maintainers) ensure that your app is compiled for and works correctly on all the different platforms they support.
Most people who are heavy users of their phones get a new one every 1-2 years, and when these new phones come out they are likely to have whatever is the current version of android... Very few people ever upgrade their existing phones, they just replace them when their contract expires.
That said, given the open source nature of android obsolete handsets could easily be opened up to third party development once the original vendor loses interest (if it doesn't happen already - some handsets aren't locked),
Such issues occur *everywhere* tho, a totally locked down platform such as apple's has other serious disadvantages... Windows has the same problem - thousands of possible hardware configurations, several different software versions - not just different generations, but multiple versions of the same generation... Various consumer oriented devices also have differences, my car differs from the model sold the previous year and there are plenty of those (and even older models) still on the roads.
By that reckoning, applications for macos should be much cheaper than applications for windows or linux (which have a wider variety of versions and run on all kinds of different hardware with a whole mishmash of drivers)...
Windows developers don't test on every possible configuration, neither do windows mobile developers... And it's quite possible to install software which doesn't work for various reasons to do with your software versions or installed hardware (even things far more arbitrary than your hardware not being powerfull enough, for example some game drm even refuses to let you run the game if you have a scsi cdrom attached)... At least android hides unusable applications from the market, i imagine user satisfaction would be decreased if people could download apps which were useless to them.
I would imagine that the vast majority of GPL code, even in cases like this, gets compiled using gcc so you already have the compiler.. It's just the build scripts you really need.
That's more to do with vxworks requiring less memory (and thus the hardware can be made cheaper), you can still try to flash linux onto those devices but they don't work very well due to the limited amount of memory/flash... They still sell linux based devices, but these are no longer the lowest and cheapest routers they offer - the vxworks ones are the new bargain bucket.
If the card is based on a broadcom chipset, why not just use generic broadcom drivers instead of the specific linksys ones?
That said, i tend to avoid cards based on broadcom chipsets because they don't release any specs for them, i have atheros, ralink and realtek based wireless cards which work extremely well with open drivers here.
Build a setup where the boxes get re-imaged every time they're power cycled, and ensure the network where the workstations sit is isolated (ie firewalled off) from any of the important systems... That way you can give the students root level access and let them experiment all they want with the worst case being they have to reboot to get back to a clean image.
If schoolkids use established programs, those programs will likely have been replaced with what are now "up and coming" by the time those kids leave school and start working (especially if they attend a university or such in between)... Wordperfect for dos was established when i went to school, where is it now? I haven't seen either dos or wordperfect for years...
That said, i still say you need to teach them general concepts that they can apply to any system, if you only shows kids one system they will think that's all there is and freak out if they ever encounter anything different (and that includes newer versions of the same system which is the very least they will encounter).
Use a cybercafe style setup where each system gets reimaged on boot, and then extend that system so that it gives you a choice of several images... If you have windows 7 you usually get downgrade rights too, so you could add xp as one of the options for when it might be needed, it also benefits the students because they get to use multiple different systems rather than erroneously learning that everything is the same (and then getting a nasty shock when they leave school).
I would say the best choice for high school students is to actually expose them to as many varied systems as possible... That way they learn how to use computers in general rather than specific systems. Teach users what the bold option in a word processor is for and how to find it in different applications rather than "its third from the left in the row of icons at the top of the screen" or similar.
The reason for this is simple, when they leave school and enter the world of work whatever system they used in school is likely to be irrelevant... When i was in school, the computers ran wordperfect 5.0 for dos, but since entering the world of work i haven't encountered any such systems, and even current versions of wordperfect are very different to the dos version... If you give the students XP or OSX 10.3 then these systems are already outdated and being replaced, even if you give them Windows 7 and OSX 10.6 by the time students enter the workplace they will be a generation or two behind anyway and if they never learned general concepts rather than specific applications they will have trouble adapting to newer versions. So you either teach kids specific systems, which they will never again use or you teach them general concepts that they can apply to whatever they might be using in the future.
As for maintenance, setup a cybercafe style system where each machine downloads a fresh image whenever its booted, have students store all their work on usb sticks or a centralised server (if you go for a server make sure you secure it well, some of the kids will be smart and try to break into it). It's not hard to have a handful of OS images available for such a system, especially if the hardware is relatively similar and if you reimage the machines after each boot you can give the students a lot more freedom to learn (ie you can give some of them root/admin rights) without it creating extra work for you (any mess they make is gone by rebooting). Build the network like an ISP instead of a corporate network, keep the servers away from the workstations and ensure that any security in place such as web access filtering is done at the network layer and doesn't depend on the workstations.
I don't even get why these arm based machines come with windows ce, that's just setting the user up for disappointment... Sure it looks like windows, but won't run any of the apps people would expect to run on it.... Linux at least doesn't create such a false impression, and has a much wider array of applications readily available for it.
Swapping the hard drive doesn't work to well if that happens to be the component that caused the failure...
Dumb terminals really are much better for a large network, you will replace the terminals far less frequently so you will have 1 model instead of 5... If a terminal fails a user can resume his session on another one while you bring a replacement for his broken one. They don't use much power, don't generate much heat, don't run complex software that needs updating regularly, don't have as many components that could go wrong and are less attractive to thieves.
How many users in a typical company really do anything that's so intensive that it wouldn't work well on a dumb terminal setup?
But instead of rendering 1080p video, benchmark the applications that you will actually be using on these machines.
Also consider other factors than performance, how much power do they consume, and how much noise or heat do they put out? I've worked in office environments which were insufferably hot due to the a/c being specced for a room containing 10 people, not 20 people plus large power hungry computers with multiple monitors on each.
Unless your in a small niche market, you will probably find that smaller lower powered machines are a much better fit than big powerful boxes capable of playing the latest games.
Collecting wireless traffic should not be illegal... The frequencies used by wifi are open for public use, so if traffic can be received from a public area like the road there should be nothing wrong with someone receiving it. So long as you don't try to send any data to someone else's network, or try to crack any encrypted data you receive. Simply by walking around my phone picks up wireless signals, it keeps track of wireless networks its used before and will try to connect to them if it needs to.
2000 was supposed to be XP, everyone was supposed to migrate from 9x to 2000... That didn't happen, so they released ME which seemed like it's sole purpose was to make the 9x series look as terrible as possible in order to convince people to move to 2000 or xp.
Slashdot Mirror
No tech has complete coverage and is unlikely to any time soon, you would be forced to support multiple options in any case. I would probably host the videos on youtube and link to them - let youtube work out what codecs to use etc.
What about suing them for patents not covered by MPEG-LA ? Don't google own a number of video encoding related patents having acquired On2? There may well be other patents out there too...
I can already have HTML5 video support in linux distros, for free, without having to wait for adobe thanks to ogg theora and webm...
I can even have this support in linux/arm, 64bit windows, linux/mips, netbsd, freebsd, amigaos, solaris, irix, beos or a new platform of my own creation should i need to... With flash i am forced to use the very limited set of platforms which adobe supports.
Incidentally, HTML5 video is new which is why you have a poor experience with it, nothing has really been optimized yet... I doubt it will be long before it outperforms flash, it already does on OSX with safari by quite a considerable margin.
What Adobe are basically saying, is that a single proprietary monoculture (flash) is better than multiple slightly incompatible implementations (browsers)...
I have lots of devices here which can access HTML, but considerably less which support flash... There are open source implementations of HTML but pretty much only one closed source implementation of flash, so i can use html on niche systems like haiku, netbsd, amigaos etc while i can't use flash because adobe doesn't bother to support such niche platforms. A few years ago i used an sgi and an alphastation as my primary workstations, neither of which had flash support.
You have to make both 64 and 32bit versions for other platforms too, it's just that most mac/windows devs simply don't bother, 64bit linux is far more mature than 64bit windows or osx.
You get strange incompatibilities between different windows or osx versions too.. for instance, a portable windows app tool i used a while ago (bundles an application into a standalone binary you can run from a usb stick) works on vista but not on 7...
You could also compile an ARM version for linux or windows ce...
With linux you always have the option of distributing source code and letting third parties (ie distro maintainers) ensure that your app is compiled for and works correctly on all the different platforms they support.
Try a guruplug, 2 gigabit ethernet ports, 512mb ram, 1.2ghz arm cpu, boots from flash... perfect for a small router and they're quite cheap.
Most people who are heavy users of their phones get a new one every 1-2 years, and when these new phones come out they are likely to have whatever is the current version of android...
Very few people ever upgrade their existing phones, they just replace them when their contract expires.
That said, given the open source nature of android obsolete handsets could easily be opened up to third party development once the original vendor loses interest (if it doesn't happen already - some handsets aren't locked),
Such issues occur *everywhere* tho, a totally locked down platform such as apple's has other serious disadvantages...
Windows has the same problem - thousands of possible hardware configurations, several different software versions - not just different generations, but multiple versions of the same generation...
Various consumer oriented devices also have differences, my car differs from the model sold the previous year and there are plenty of those (and even older models) still on the roads.
By that reckoning, applications for macos should be much cheaper than applications for windows or linux (which have a wider variety of versions and run on all kinds of different hardware with a whole mishmash of drivers)...
Windows developers don't test on every possible configuration, neither do windows mobile developers... And it's quite possible to install software which doesn't work for various reasons to do with your software versions or installed hardware (even things far more arbitrary than your hardware not being powerfull enough, for example some game drm even refuses to let you run the game if you have a scsi cdrom attached)... At least android hides unusable applications from the market, i imagine user satisfaction would be decreased if people could download apps which were useless to them.
I would imagine that the vast majority of GPL code, even in cases like this, gets compiled using gcc so you already have the compiler.. It's just the build scripts you really need.
That's more to do with vxworks requiring less memory (and thus the hardware can be made cheaper), you can still try to flash linux onto those devices but they don't work very well due to the limited amount of memory/flash...
They still sell linux based devices, but these are no longer the lowest and cheapest routers they offer - the vxworks ones are the new bargain bucket.
If the card is based on a broadcom chipset, why not just use generic broadcom drivers instead of the specific linksys ones?
That said, i tend to avoid cards based on broadcom chipsets because they don't release any specs for them, i have atheros, ralink and realtek based wireless cards which work extremely well with open drivers here.
Build a setup where the boxes get re-imaged every time they're power cycled, and ensure the network where the workstations sit is isolated (ie firewalled off) from any of the important systems...
That way you can give the students root level access and let them experiment all they want with the worst case being they have to reboot to get back to a clean image.
If schoolkids use established programs, those programs will likely have been replaced with what are now "up and coming" by the time those kids leave school and start working (especially if they attend a university or such in between)... Wordperfect for dos was established when i went to school, where is it now? I haven't seen either dos or wordperfect for years...
That said, i still say you need to teach them general concepts that they can apply to any system, if you only shows kids one system they will think that's all there is and freak out if they ever encounter anything different (and that includes newer versions of the same system which is the very least they will encounter).
If the same incorrect results are independently reproduced using the same flawed black box systems, does that make the results any less incorrect?
Use a cybercafe style setup where each system gets reimaged on boot, and then extend that system so that it gives you a choice of several images...
If you have windows 7 you usually get downgrade rights too, so you could add xp as one of the options for when it might be needed, it also benefits the students because they get to use multiple different systems rather than erroneously learning that everything is the same (and then getting a nasty shock when they leave school).
I would say the best choice for high school students is to actually expose them to as many varied systems as possible... That way they learn how to use computers in general rather than specific systems.
Teach users what the bold option in a word processor is for and how to find it in different applications rather than "its third from the left in the row of icons at the top of the screen" or similar.
The reason for this is simple, when they leave school and enter the world of work whatever system they used in school is likely to be irrelevant... When i was in school, the computers ran wordperfect 5.0 for dos, but since entering the world of work i haven't encountered any such systems, and even current versions of wordperfect are very different to the dos version... If you give the students XP or OSX 10.3 then these systems are already outdated and being replaced, even if you give them Windows 7 and OSX 10.6 by the time students enter the workplace they will be a generation or two behind anyway and if they never learned general concepts rather than specific applications they will have trouble adapting to newer versions.
So you either teach kids specific systems, which they will never again use or you teach them general concepts that they can apply to whatever they might be using in the future.
As for maintenance, setup a cybercafe style system where each machine downloads a fresh image whenever its booted, have students store all their work on usb sticks or a centralised server (if you go for a server make sure you secure it well, some of the kids will be smart and try to break into it). It's not hard to have a handful of OS images available for such a system, especially if the hardware is relatively similar and if you reimage the machines after each boot you can give the students a lot more freedom to learn (ie you can give some of them root/admin rights) without it creating extra work for you (any mess they make is gone by rebooting). Build the network like an ISP instead of a corporate network, keep the servers away from the workstations and ensure that any security in place such as web access filtering is done at the network layer and doesn't depend on the workstations.
I don't even get why these arm based machines come with windows ce, that's just setting the user up for disappointment... Sure it looks like windows, but won't run any of the apps people would expect to run on it....
Linux at least doesn't create such a false impression, and has a much wider array of applications readily available for it.
Hence the benchmarking...
Screwless maintenence also makes it much easier for people to steal components...
Pop open the lid and take out the ram sticks or cpu
Swapping the hard drive doesn't work to well if that happens to be the component that caused the failure...
Dumb terminals really are much better for a large network, you will replace the terminals far less frequently so you will have 1 model instead of 5...
If a terminal fails a user can resume his session on another one while you bring a replacement for his broken one.
They don't use much power, don't generate much heat, don't run complex software that needs updating regularly, don't have as many components that could go wrong and are less attractive to thieves.
How many users in a typical company really do anything that's so intensive that it wouldn't work well on a dumb terminal setup?
Consider power use, heat dissipation and noise too... Spread out over hundreds of machines the power use and heat output will add up.
But instead of rendering 1080p video, benchmark the applications that you will actually be using on these machines.
Also consider other factors than performance, how much power do they consume, and how much noise or heat do they put out? I've worked in office environments which were insufferably hot due to the a/c being specced for a room containing 10 people, not 20 people plus large power hungry computers with multiple monitors on each.
Unless your in a small niche market, you will probably find that smaller lower powered machines are a much better fit than big powerful boxes capable of playing the latest games.
Collecting wireless traffic should not be illegal...
The frequencies used by wifi are open for public use, so if traffic can be received from a public area like the road there should be nothing wrong with someone receiving it. So long as you don't try to send any data to someone else's network, or try to crack any encrypted data you receive.
Simply by walking around my phone picks up wireless signals, it keeps track of wireless networks its used before and will try to connect to them if it needs to.
2000 was supposed to be XP, everyone was supposed to migrate from 9x to 2000...
That didn't happen, so they released ME which seemed like it's sole purpose was to make the 9x series look as terrible as possible in order to convince people to move to 2000 or xp.