Criminals will go back to using payphones and face to face meetings to discuss their criminal activities. And stealing phones, since they're already criminals having to steal a phone isn't much of a deterrent.
Yes, the risks of incorporating open source under licenses such as the GPL into a proprietary product you distribute are valid, however... There is also the risk of incorporating closed source code or linking to / distributing a proprietary library.
But this is assuming your business distributes closed source software, which most don't.
If you do get caught using code in violation of it's license, those enforcing the GPL will usually want you to stop infringing, whereas a proprietary company will often want a cut of any profits you ever made from the software.
I know several people who work for companies that sell proprietary software, and most of them don't use that software themselves, even tho they could get it for free (without pirating it).
You really have to worry about the quality of software when even it's authors don't want to use it (and forcing them to use it doesn't count). They say programming is like an art, but there's no passion involved when you've no interest in what your working on, it becomes purely a mundane 9-5 job.
Yes, killing proprietary software would be good...
Proprietary vendors have time and again proven they cannot be trusted, getting their customers locked in to proprietary formats so they can't leave rather than competing with a better product.
We'd gain the ability to modify code, switch to other providers at will, choose who we want to provide support or even choose not to have paid support if we have the skills and save the money.
OEMs would gain the ability to customise the software as much as they wanted to for their hardware.
And development would progress faster, as anyone could reuse existing code and make incremental improvements rather than having to reinvent the wheel constantly, and people wouldn't be wasting their time trying to reverse engineer proprietary formats.
It would be better for pretty much everyone,with the exception of the fat cats at purely software companies.
The idea of a singular source of support is pretty offputting to me... A single source of support is a monopoly, they can provide half assed support at premium prices and you have no choice but to suck it up. ZenOSS is a good example here, does anyone else provide support for it? Do you think their enterprise support would improve if someone did? RedHat is also a good example, many other companies provide a supported Linux distribution, if RedHat provided lousy support they would lose customers very quickly.
The problem is that large companies are packed full of people with little or no problem solving skills... They either don't want to, or are incapable of trying to solve problems themselves, and would rather pay extra for someone else to do it... Yes, they're basically not doing their jobs, and yet these blatantly incompetent people end up being paid a lot of money.
On the other hand, those people who are smart enough to solve problems (and it really isn't that hard) can set up support consultancies and employ people to do what you're doing on behalf of other companies.
I've seen countless situations where relatively simple problems were unable to be solved internally, and the people who's responsibility it was to fix them just wanted to hand them off to a third party as quickly as possible, and simply didn't have the skill to diagnose what was wrong. The issue took a few seconds to diagnose, and a few seconds to fix once someone with the right mindset started looking at it.
Most open source products have 24/7 support available if you're willing to pay for it... If you don't want it, you pay nothing and still get to use the app. Similarly, most closed source products come with little or no support by default, and you then have to pay even more to get a decent level of support.
But more importantly, closed source typically gives you one choice for support - the original vendor, third parties don't have sufficient access to the app to provide a proper level of support. Open source apps often have multiple vendors who can provide support, so you get choice and competition which improves quality and drives down prices.
The whole mentality here is that anybody can change the source of a project, submit it, and you never know what kind of compiled binary you're going to get.
What, and all the viruses that can attach themselves to existing binaries clearly have never existed?
If you have the source code, then you have the opportunity to compile your own binary and be sure what's in it.
Yeah but it's not proper Excel, last i checked it was a massively cut down version with missing features and compatibility problems with it's bigger counterpart. People complain that OpenOffice lacks features and compatibility, but Mobile Excel is worse... A linux based device could easily have OpenOffice compiled for it.
Yeah, Google Docs needs some improvement for sure... But MS's online office attempts are not truly web based, you still need their client to tie into it, Google's can be viewed from anywhere by anyone with a browser.
The issue here is that for linux bleeding edge apps are often easily obtainable, and have higher version numbers than the stable versions making users want them.
Viruses/malware - You could indeed create a.sh file and deliver it via a browser onto grandma's machine, but you would also have to socially engineer her into setting it with execute permissions and executing it, the poor design of windows is that it makes it much easier to download and execute a malicious program since it decides if a program is executable or not based on it's filename, it also encourages you to run with greater privileges so the scope of damage is greater.. it's possible and trivially easy to make it impossible for linux users to execute anything that wasn't installed by the root user btw, windows requires extra third party apps for that. Also, windows users are conditioned to downloading and executing binaries, linux users are more likely to select software from their package manager.
Crashes - complexity is the issue, the underlying system in windows is far more complex than linux and therefore harder to maintain, plus you have a lot of third party driver code in the kernel on virtually every system and no guarantee that your mix of driver versions has been tested together.
Drivers - yes, add all the drivers to the kernel, that way every distro supports every device out of the box, and you know that your device will continue to be supported out of the box by linux, you don't have to hunt around for the latest drivers since updating the kernel updates the drivers too, and the drivers come as a single package thats been tested to work together properly. Drivers and hardware are not developed over night, if the hardware vendor started the driver development process early and in the open they could quite easily have working drivers filtered down by the time the hardware is made available, and they could still supply standalone drivers as a temporary hack (i've never had anything but problems with out of kernel drivers).
Applications - With a decent package manager you can have multiple repositories, and you can also install from physical media while using the package manager as an update system. The only problem with this system is the number of different package managers... It scales perfectly well, what doesn't scale is every individual application providing their own background daemon that checks for updates and clogs your machine up, or each app only checking for updates when it's started - you don't want the app to update itself when you're trying to use it... Is this what you propose for handling multi vendor applications and updating them, or can you think of a better system than a package manager?
I installed linux (redhat 5.1) on my amiga back in the day and ran enlightenment on it... It was surprisingly quick too, even with just the AGA chipset and no zorro video card.
Most people could get along just fine with a CE device, provided it had a decent web browser, Office, image and video viewing, audio playing, silly flash games and cards, and access to their MFD printers.
Yes, they would, but why microsoft? their biggest advantage is compatibility with their existing mass of apps, which would be lost using a ce-based machine... Giving them no advantage over linux, while still retaining the majority of the disadvantages.
The linux based small laptops work well, tho the manufacturers could put a lot more effort in towards making them work better... My eee for instance runs much faster with gentoo than it did with the default xandros, since it comes preinstalled on a single piece of hardware theres no reason the same performance tweaks couldn't have been made, and the available optional packages with xandros were rather limited.
As you pointed out, win32 compatibility is a powerful selling point, infact it's probably their biggest selling point...
They don't have write once run anywhere with.net yet, java on the other hand is years ahead, although in both cases the overhead of running a VM on an embedded device is far from ideal.
It seems they can't get win32 running on an embedded device, whereas the linux/osx apis do run on such devices...
Even if they did get win32 running on an embedded device, it would still be a different processor architecture requiring a recompile, and most win32 apps do not come with source making this rather difficult.
I think embedded devices should have the same APIs and run the same applications (albeit with different interfaces) as full blown desktops... I want a linux phone that can be connected to a keyboard and monitor so i can use it as a full blown computer wherever i go, and apps which seperate the backend functionality from the frontend interface - so i can have a choice of interfaces to the same app. I want to be able to do stuff on my phone, but also be able to easily use exactly the same apps on a more powerful system should i need the extra resources for whatever purpose.
Microsoft simply don't offer that at the moment, whereas others (linux, java) already do.. And i think it's ridiculous and arrogant of them to come along with something new proprietary and incompatible, rather than building on or simply using the existing functional platforms. All it does is dilute the developer pool, slow progress and diminish the quality for the end users, as developers waste their time reinventing the wheel for microsoft's not-compatible-with-anything-else platforms.
Your right that XP isn't suitable, because of it's size and complexity... And CE doesn't offer compatibility with existing apps, it needs apps specifically written for it... Here's where the problem lies tho.
Apps already exist for XP, but implementing the same APIs to make something compatible with existing apps results in XP, something rather too big for small machines... But when they try and design something new, they lose their biggest advantage - existing locked in userbase and applications. Infact, this now goes against them.
You have ce, or windows mobile etc, which have to start from scratch with new applications, compared to linux which is the same linux that can run the same apps with minimal effort.
Porting apps to windows ce/mobile requires significant effort, and many of the apps are proprietary so those interested in doing the porting don't have access to the source, and those who do have the source aren't interested in porting them. you have to do significant work to make the apps compile at all, let alone trying to tailor them for the different interface.
Porting apps to run on a linux based phone however, can be as simple as a recompile, and the majority of linux apps have source code available making this a very easy task to accomplish. most work done with embedded linux centers around changing the interface of apps to be more suited to the embedded device.
And the gear shift, automatic, manual, paddle shift, pre-selector etc... And the selector is sometimes on the steering column, and sometimes in the center console. Different fuel - some use petrol, some use higher grade petrol, some use diesel, some use LPG, some even use electric these days. Handbrake - some have a pedal you push down, some have a handle set into the dash you pull to release, some have automatic handbrake, some have a handle next to your seat...
And that's just the most basic essential controls... Things like the entertainment and navigation systems can have wildly different interfaces, controls for opening the windows can vary from a handle you have to turn manually, to buttons which could be located somewhere on the door or somewhere centrally, not to mention all the other equipment that's optional which can have very different interfaces or even not be fitted at all.
Feedback from the car can vary wildly too, many cars have a tachometer but some do not, some have an analog dial to show you engine rpm and speed, some have digital readouts or both, some have a trip computer to calculate your fuel economy and average speed etc, and the warning lights, some are standardized but some vary wildly between models... some cars even display manufacturer or model specific numeric fault codes. And all of these readouts may be behind the steering wheel, or perhaps located closer to the center of the car...
The interfaces presented by cars vary wildly, you can't have driven too many if you think the location of the indicator controls are the only difference.
Personally i want hardware accelerated graphics, not to look pretty but to offload work from the CPU... Most computers these days come with a reasonably powerful GPU, makes no sense for it to sit idle while your CPU could be better occupied doing other things.
Trains vary in speed greatly, but in the UK at least are generally quite slow... The top end ones travel at 125mph or so, cars can easily travel quicker than that and if a driver is travelling at such a speed you really don't them talking on the phone at the same time, and laws about phone use in cars make no difference since they're already breaking the law to drive at that speed.
On the other hand, many of the smaller local trains and london underground trains travel between 40 and 60mph, which is a perfectly reasonable and common speed for someone to be driving too.
The preachers look for places where lots of people are standing around waiting for something, it means they will be there for longer to hear (if not listen to) more of the sermon.
I could be moving at "car" speeds when riding in a taxi, or being a passenger in a friends car, or riding a bus or train...
What i do think would be useful tho, is for the "profiles" system on most phones be able to be tied to a system of recorded messages, the "i am currently driving and cannot answer the phone" is just one of them, how about "i am currently in a meeting, i will be out at 4pm" or "i am on a flight to new york and won't have access to my phone for at least 6 hours". Far more useful than the default voicemail message most people get...
Indeed, lists of requirements, notices that should be displayed in the program, diagrams, text strings, even graphical elements and all kinds of other shit are sent to developers in ridiculous formats like that... Many of those things would make a lot more sense to extract programatically, especially when management keeps changing their minds.
Slashdot Mirror
Criminals will go back to using payphones and face to face meetings to discuss their criminal activities.
And stealing phones, since they're already criminals having to steal a phone isn't much of a deterrent.
Yes, the risks of incorporating open source under licenses such as the GPL into a proprietary product you distribute are valid, however...
There is also the risk of incorporating closed source code or linking to / distributing a proprietary library.
But this is assuming your business distributes closed source software, which most don't.
If you do get caught using code in violation of it's license, those enforcing the GPL will usually want you to stop infringing, whereas a proprietary company will often want a cut of any profits you ever made from the software.
I know several people who work for companies that sell proprietary software, and most of them don't use that software themselves, even tho they could get it for free (without pirating it).
You really have to worry about the quality of software when even it's authors don't want to use it (and forcing them to use it doesn't count). They say programming is like an art, but there's no passion involved when you've no interest in what your working on, it becomes purely a mundane 9-5 job.
Yes, killing proprietary software would be good...
Proprietary vendors have time and again proven they cannot be trusted, getting their customers locked in to proprietary formats so they can't leave rather than competing with a better product.
We'd gain the ability to modify code, switch to other providers at will, choose who we want to provide support or even choose not to have paid support if we have the skills and save the money.
OEMs would gain the ability to customise the software as much as they wanted to for their hardware.
And development would progress faster, as anyone could reuse existing code and make incremental improvements rather than having to reinvent the wheel constantly, and people wouldn't be wasting their time trying to reverse engineer proprietary formats.
It would be better for pretty much everyone,with the exception of the fat cats at purely software companies.
The idea of a singular source of support is pretty offputting to me...
A single source of support is a monopoly, they can provide half assed support at premium prices and you have no choice but to suck it up.
ZenOSS is a good example here, does anyone else provide support for it? Do you think their enterprise support would improve if someone did?
RedHat is also a good example, many other companies provide a supported Linux distribution, if RedHat provided lousy support they would lose customers very quickly.
The problem is that large companies are packed full of people with little or no problem solving skills...
They either don't want to, or are incapable of trying to solve problems themselves, and would rather pay extra for someone else to do it...
Yes, they're basically not doing their jobs, and yet these blatantly incompetent people end up being paid a lot of money.
On the other hand, those people who are smart enough to solve problems (and it really isn't that hard) can set up support consultancies and employ people to do what you're doing on behalf of other companies.
I've seen countless situations where relatively simple problems were unable to be solved internally, and the people who's responsibility it was to fix them just wanted to hand them off to a third party as quickly as possible, and simply didn't have the skill to diagnose what was wrong.
The issue took a few seconds to diagnose, and a few seconds to fix once someone with the right mindset started looking at it.
Most open source products have 24/7 support available if you're willing to pay for it... If you don't want it, you pay nothing and still get to use the app.
Similarly, most closed source products come with little or no support by default, and you then have to pay even more to get a decent level of support.
But more importantly, closed source typically gives you one choice for support - the original vendor, third parties don't have sufficient access to the app to provide a proper level of support. Open source apps often have multiple vendors who can provide support, so you get choice and competition which improves quality and drives down prices.
The whole mentality here is that anybody can change the source of a project, submit it, and you never know what kind of compiled binary you're going to get.
What, and all the viruses that can attach themselves to existing binaries clearly have never existed?
If you have the source code, then you have the opportunity to compile your own binary and be sure what's in it.
Yeah but it's not proper Excel, last i checked it was a massively cut down version with missing features and compatibility problems with it's bigger counterpart. People complain that OpenOffice lacks features and compatibility, but Mobile Excel is worse... A linux based device could easily have OpenOffice compiled for it.
Yeah, Google Docs needs some improvement for sure...
But MS's online office attempts are not truly web based, you still need their client to tie into it, Google's can be viewed from anywhere by anyone with a browser.
Windows 7 will be no better than vista, but by then the hardware will be even more powerful so it won't seem so bloated.
The issue here is that for linux bleeding edge apps are often easily obtainable, and have higher version numbers than the stable versions making users want them.
Viruses/malware - You could indeed create a .sh file and deliver it via a browser onto grandma's machine, but you would also have to socially engineer her into setting it with execute permissions and executing it, the poor design of windows is that it makes it much easier to download and execute a malicious program since it decides if a program is executable or not based on it's filename, it also encourages you to run with greater privileges so the scope of damage is greater.. it's possible and trivially easy to make it impossible for linux users to execute anything that wasn't installed by the root user btw, windows requires extra third party apps for that. Also, windows users are conditioned to downloading and executing binaries, linux users are more likely to select software from their package manager.
Crashes - complexity is the issue, the underlying system in windows is far more complex than linux and therefore harder to maintain, plus you have a lot of third party driver code in the kernel on virtually every system and no guarantee that your mix of driver versions has been tested together.
Drivers - yes, add all the drivers to the kernel, that way every distro supports every device out of the box, and you know that your device will continue to be supported out of the box by linux, you don't have to hunt around for the latest drivers since updating the kernel updates the drivers too, and the drivers come as a single package thats been tested to work together properly. Drivers and hardware are not developed over night, if the hardware vendor started the driver development process early and in the open they could quite easily have working drivers filtered down by the time the hardware is made available, and they could still supply standalone drivers as a temporary hack (i've never had anything but problems with out of kernel drivers).
Applications - With a decent package manager you can have multiple repositories, and you can also install from physical media while using the package manager as an update system. The only problem with this system is the number of different package managers... It scales perfectly well, what doesn't scale is every individual application providing their own background daemon that checks for updates and clogs your machine up, or each app only checking for updates when it's started - you don't want the app to update itself when you're trying to use it... Is this what you propose for handling multi vendor applications and updating them, or can you think of a better system than a package manager?
I installed linux (redhat 5.1) on my amiga back in the day and ran enlightenment on it... It was surprisingly quick too, even with just the AGA chipset and no zorro video card.
Most people could get along just fine with a CE device, provided it had a decent web browser, Office, image and video viewing, audio playing, silly flash games and cards, and access to their MFD printers.
Yes, they would, but why microsoft? their biggest advantage is compatibility with their existing mass of apps, which would be lost using a ce-based machine... Giving them no advantage over linux, while still retaining the majority of the disadvantages.
The linux based small laptops work well, tho the manufacturers could put a lot more effort in towards making them work better... My eee for instance runs much faster with gentoo than it did with the default xandros, since it comes preinstalled on a single piece of hardware theres no reason the same performance tweaks couldn't have been made, and the available optional packages with xandros were rather limited.
As you pointed out, win32 compatibility is a powerful selling point, infact it's probably their biggest selling point...
They don't have write once run anywhere with .net yet, java on the other hand is years ahead, although in both cases the overhead of running a VM on an embedded device is far from ideal.
It seems they can't get win32 running on an embedded device, whereas the linux/osx apis do run on such devices...
Even if they did get win32 running on an embedded device, it would still be a different processor architecture requiring a recompile, and most win32 apps do not come with source making this rather difficult.
I think embedded devices should have the same APIs and run the same applications (albeit with different interfaces) as full blown desktops... I want a linux phone that can be connected to a keyboard and monitor so i can use it as a full blown computer wherever i go, and apps which seperate the backend functionality from the frontend interface - so i can have a choice of interfaces to the same app.
I want to be able to do stuff on my phone, but also be able to easily use exactly the same apps on a more powerful system should i need the extra resources for whatever purpose.
Microsoft simply don't offer that at the moment, whereas others (linux, java) already do.. And i think it's ridiculous and arrogant of them to come along with something new proprietary and incompatible, rather than building on or simply using the existing functional platforms. All it does is dilute the developer pool, slow progress and diminish the quality for the end users, as developers waste their time reinventing the wheel for microsoft's not-compatible-with-anything-else platforms.
Your right that XP isn't suitable, because of it's size and complexity...
And CE doesn't offer compatibility with existing apps, it needs apps specifically written for it...
Here's where the problem lies tho.
Apps already exist for XP, but implementing the same APIs to make something compatible with existing apps results in XP, something rather too big for small machines...
But when they try and design something new, they lose their biggest advantage - existing locked in userbase and applications. Infact, this now goes against them.
You have ce, or windows mobile etc, which have to start from scratch with new applications, compared to linux which is the same linux that can run the same apps with minimal effort.
Porting apps to windows ce/mobile requires significant effort, and many of the apps are proprietary so those interested in doing the porting don't have access to the source, and those who do have the source aren't interested in porting them. you have to do significant work to make the apps compile at all, let alone trying to tailor them for the different interface.
Porting apps to run on a linux based phone however, can be as simple as a recompile, and the majority of linux apps have source code available making this a very easy task to accomplish. most work done with embedded linux centers around changing the interface of apps to be more suited to the embedded device.
And the gear shift, automatic, manual, paddle shift, pre-selector etc... And the selector is sometimes on the steering column, and sometimes in the center console.
Different fuel - some use petrol, some use higher grade petrol, some use diesel, some use LPG, some even use electric these days.
Handbrake - some have a pedal you push down, some have a handle set into the dash you pull to release, some have automatic handbrake, some have a handle next to your seat...
And that's just the most basic essential controls... Things like the entertainment and navigation systems can have wildly different interfaces, controls for opening the windows can vary from a handle you have to turn manually, to buttons which could be located somewhere on the door or somewhere centrally, not to mention all the other equipment that's optional which can have very different interfaces or even not be fitted at all.
Feedback from the car can vary wildly too, many cars have a tachometer but some do not, some have an analog dial to show you engine rpm and speed, some have digital readouts or both, some have a trip computer to calculate your fuel economy and average speed etc, and the warning lights, some are standardized but some vary wildly between models... some cars even display manufacturer or model specific numeric fault codes. And all of these readouts may be behind the steering wheel, or perhaps located closer to the center of the car...
The interfaces presented by cars vary wildly, you can't have driven too many if you think the location of the indicator controls are the only difference.
Personally i want hardware accelerated graphics, not to look pretty but to offload work from the CPU...
Most computers these days come with a reasonably powerful GPU, makes no sense for it to sit idle while your CPU could be better occupied doing other things.
But by that reckoning, the Wii could have been the Nintendo Entertainment System V5..
Yeah, it's always amusing when you encounter such people..
Pointing out to them that Linus has only got up to 2.6.27 can be quite amusing.
Trains vary in speed greatly, but in the UK at least are generally quite slow...
The top end ones travel at 125mph or so, cars can easily travel quicker than that and if a driver is travelling at such a speed you really don't them talking on the phone at the same time, and laws about phone use in cars make no difference since they're already breaking the law to drive at that speed.
On the other hand, many of the smaller local trains and london underground trains travel between 40 and 60mph, which is a perfectly reasonable and common speed for someone to be driving too.
The preachers look for places where lots of people are standing around waiting for something, it means they will be there for longer to hear (if not listen to) more of the sermon.
I could be moving at "car" speeds when riding in a taxi, or being a passenger in a friends car, or riding a bus or train...
What i do think would be useful tho, is for the "profiles" system on most phones be able to be tied to a system of recorded messages, the "i am currently driving and cannot answer the phone" is just one of them, how about "i am currently in a meeting, i will be out at 4pm" or "i am on a flight to new york and won't have access to my phone for at least 6 hours".
Far more useful than the default voicemail message most people get...
Indeed, lists of requirements, notices that should be displayed in the program, diagrams, text strings, even graphical elements and all kinds of other shit are sent to developers in ridiculous formats like that...
Many of those things would make a lot more sense to extract programatically, especially when management keeps changing their minds.