India Mandates Use of Open Source Software In Government
jrepin writes The Indian government announced a policy yesterday that makes it mandatory to use open-source software in building apps and services, in an effort to "ensure efficiency, transparency and reliability of such services at affordable costs." The new policy (PDF) states that all government organizations must include a requirement for their software suppliers to consider open-source options when implementing e-governance applications and systems. The move will bring the Indian government in line with other countries including the US, UK and Germany that opt for open-source software over proprietary tools.
The devil is in the detail - a requirement to "consider" open source software is not the same as mandating open source software.
#1: AD isn't necessary except to do things the way AD wants them done. They don't have to be done that way. The only thing I've EVER had told me that AD did that wasn't replaced by SAMBA was to undo the dumbfuck way Windows treats a computer, where it wants things installed on the C: drive and locally installed with the registry set up. That's not what's needed, but it only exists to undo some of the dumb ideas that were implemented by Windows. But you don't have to use AD to undo them: you don't require it to operate outside its usage. And then AD isn't needed.
#2: Exchange doesn't scale anywhere near as well as the UNIX stalwarts. Even MS found they couldn't do it without massive and specific effort. You have no chance.
#3: UNIX manages desktops fine, better than Windows, because it was always a multi-user, non-owned user system, unlike Windows. So there's nothing to this "point" either.
#4: There's massive amounts of stuff to do backups for UNIX. Who knows what the hell you're thinking of to make you think Windows does it better.
#5: Managing and delegation of privileges are better on UNIX and Linux especially than it is on Windows. They're equal when you're talking about out-of-the-box normal privileges, excepting the Windows system wanting to execute anything, even if the execute flag isn't set, if it THINKS it should be executed. You have to do extra work to stop that, and you don't have to do that with Linux.
#6: NFS has no problems doing what you need to do with the filesystem too. Just because you don't know what they are doesn't make them nonexistent.
#7: It's easy-peasy to do what you demand on Linux. Issue a key to log on that only allows access 9-5, only from pre-configured systems, and always encrypted communication.
Not to say closed software hasn't got a place or doesn't work, it just doesn't work the same way. Between other closed source software or open sourced software. And demanding that what you do on one platform must be available on another otherwise "It just doesn't replace my system" is ridiculous and self-confirming fake incompatibility.
But demanding that just because the way you do something isn't done on another doesn't mean that the usage you want isn't available. It just isn't done with the same command as you're used to.
There's nothing stopping the copyright holders of a GPL'd project from taking it proprietary. It's not much different than if the people running a permissive licensed project (BSD/MIT) decided to take it proprietary. Everything up to that point will still be available barring any patent issues. If one of the copyright holders decides they don't want their code proprietary then the project leaders can just rewrite their portion of the code and still take it closed source.
Gee, most of what you claim is not possible truly IS possible on Linux/Unix solutions IF you know what you are doing.
With the possible exception of the Active Directory vrs OpenLDAP gripe of yours (and I'm just not prepared to argue that point with you, I think it's possible if properly managed and configured that OpenLDAP would scale just fine) everything you complain about are not issues without solutions, but solutions that you may not understand how to setup and use.
Linux generally has solutions to the above problems which do work and work well for the back of house stuff. What Linux doesn't have though is the developed "tribal knowledge" of skilled administrators who understand the solutions and how to properly implement them. (Not that the tribe of Windows administrators always understand their OS of choice either.)
Your complaints seem more born of frustration with people who didn't know what they where doing or what tools did what you are accustom to seeing than being actual limitations with Linux. Or perhaps you have had some bad experiences with Linux because somebody thought it was going to be cheaper to forget the license costs of Windows and run some half baked Unix solution instead. I can tell you that if you do the Linux thing because it is cheaper up front, you are sadly mistaken and wasting your time and money. Usually Linux requires as much budget up front as the Microsoft solution and only saves you money in years to come. Much pain is caused by half understood, improperly implemented and under funded efforts in either Microsoft or Linux worlds...
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
Yet another AC
I routinely hop onto Linux setups and do maintenance/upgrades on 10,000 systems. There are tools to do it fine, just like there are things in MS land to mitigate the pain.
The short story from any seasoned admin perspective: 'Whatever platform *I* know the tools for is better than the platform that I don't know the tools for'. This applies to all the parent posts. The Windows guy thinks Linux isn't enterprise ready because he doesn't know the tools. The Linux guy is shocked to hear this because Windows in his experience is a pain in the ass.
There's nothing stopping the copyright holders of a GPL'd project from taking it proprietary. It's not much different than if the people running a permissive licensed project (BSD/MIT) decided to take it proprietary. Everything up to that point will still be available barring any patent issues.
Two points:
1) As you said, this doesn't remove the actual project from accessibility in the FOSS-sphere. You just can't have any new versions that the copyright owners decide to publish under a proprietary license. You can still use the existing code all you want, and you can fork it too if you want.
2) This isn't really a big fear among people. Honestly, how many times has this happened? The advantage of GPL over BSD licenses is that with GPL, only the actual copyright holder can do such a thing. With BSD, anyone can do this, so you have cases like Kerberos, where a giant company with lots of resources grabs some FOSS code, "extends" it with proprietary extensions, and then pushes the new version so it's effectively been hijacked. This can't happen with GPL; there, the big corp would have to buy the copyrights from the original holders (which may be very difficult if there's a lot of contributors; with Linux (kernel), for instance, it's probably impossible to get all the contributors to agree to selling or even to a license change to GPLv3), before they could pull such a move.