It is common practice in the UNIX and Linux world to create seperate users and groups for specific access. For instance floppy, cdrom, ftp, mail, etc. You can use this to elevate access to a specific device or folder.
I've seen several apps in the *nix world create an associated user or group account as well.
There are several linux distros that won't let you log into gdm/kdm as root. Windows was designed for users to login as administrators.
Microsoft is trying to change that mentality with Vista and 7, except too many applications are having issues with UAC. What Microsoft should have done is said, "you're not allowed to claim your application works with Vista and 7 unless it behaves nicely with UAC."
Even better, it should be following a proper UNIX-esque security model. It could create users/groups for specific escalation. Apps shouldn't ask to escalte to administrator level. They should ask only to escalate the rights they specifically need, such as writing to C:\Program Files\Foo\.
Microsoft is happy to blame the users, but it is Microsoft who established the industry standards. They set the table. They tell the users how to use their OS, and they tell developers how to develop for their OS. If Microsoft shipped a more secure design from the get-go, we wouldn't have as many issues. I'm sure malware authors would still target the market-share king and eventually find chinks in the armor, but right now it is so easy to target Windows that every script-kiddie on the planet pulls it off with ease.
Microsoft claimed that IE 8 was the most secure browser on the planet, and that Windows 7 was the most secure OS ever. It clearly isn't their fault for making an insecure OS that is subject to malware.
I made a different bagel joke about bilking people for a 6 billion dollar upgrade. Funny is very subjective. To each their own.
However, I was under the impression that reading TFA is strictly forbidden on/. Anything that is worth knowing can be gleemed from the unerring wisdom of the discussion threads.
By having OSS drivers internal to the kernel itself, when there is any chance to the Linux driver API, they can grep their way through the code for all the drivers and make necessary changes.
What gets me is that Microsoft doesn't hold the source for all Windows drivers, but they do sign all of them. Why can't they at the very least check for proper API compatibility, and also keep them all in a central repository. If Microsoft has signed the driver, it should be available via Microsoft/Windows Update. A few drivers are available there, but not most, and often they are the wrong versions.
Microsoft doesn't want to be tied to hard deadlines of a fixed release schedule, but if people automatically knew ahead of time to budget for an OS upgrade at a set period of time, they'd come to accept it.
I also think a steady stream of quicker service packs every six months would make it easier to keep more consistent patch levels, and make support easier.
I know 2000 was the first to ship with the NT kernel (other than NT itself) but I thought ME included some NT-based dlls and was starting the shift in that direction.
Look at the kernel changes between Server 2008 and Server 2008 R2 and tell me it is the same kernel. It is easily twice the change of 2003 to 2003 R2.
Microsoft is HORRIBLY inconsistent with their naming. Windows 7 reports it runs kernel 6.1, but the entire purpose of naming it Windows _7_ is that is the 7th major iteration of the NT kernel.
Then again, Windows 2000 through Windows Home Server all are NT 5.x kernels. You're not going to suggest that XP was merely a point release and not a major release were you?
If we just went by that naming convention, then it the NT 5.x kernel line lasted 7 years.
Professional is designed to work in a domain. Home users won't be running a domain unless they have purchased Windows Server to run a domain controller. Ulimate seems really unnecessary. I say that running Windows 7 Ultimate. The only need for Windows 7 Ultimate I can imagine is getting the free VM of XP inside of 7 Ultimate (which doesn't ship on the DVD, you have to download it). And even that you can replicate with an existing XP license.
ME was a kludge relesae, with 2000 being the proper release for that goal (merging the Windows 9x GUI with the NT kernel). And you can argue that Vista was a kludge release (trying to implement UAC and a better security model, but selling the OS to the masses with Live integration and Aero interface), with 7 being the proper release of Vista's goals.
If you throw out ME and Vista, Windows has had three major releases in 9 years, and plans the next in 3 years.
openSUSE used to be a 6 month window. I think Mandriva still uses a 6 month window.
If you only count community distros, then Ubuntu, Debian, Fedora, openSUSE and Mandriva are the heavy hitters. Debian would be the exception.
However, this is a poor comparison to Windows. Windows wants you to pay money for the upgrade. Linux releases quicker, but it is a free upgrade. Many of these are more comparable to a Windows service pack.
If you count commercial distros like SLES and Red Hat, you'll see a much longer window between releases. These are more comparable to Windows releases.
They should have been working on Windows 8 nine months ago, or whever they basically put Windows 7 into freeze. It was over a year ago that they decided certain major features weren't making it into 7.
The team that develops the OS should be focused on the new version right away.
Surely Microsoft is a well-run corporation with long term planning. Surely they have a future roadmap of where they want Windows to go over the next 5 years. Surely Vista and 7 were intentional stepping stones along their master plan.
I can't fathom the possibility that Microsoft has become this un-agile behemoth that no longer innovates, but rather has knee-jerk reactions to the OS market.
Slashdot Mirror
Except *nix has had this model for decades. I could create a specific user with very specific rights, and then run a daemon as that specific user.
Windows can do this as well, but it seems to be rarely used in the Windows world.
7 is getting broad adoption and is pushing DX11. In the next few years, DX9 will start to die off.
It is common practice in the UNIX and Linux world to create seperate users and groups for specific access. For instance floppy, cdrom, ftp, mail, etc. You can use this to elevate access to a specific device or folder.
I've seen several apps in the *nix world create an associated user or group account as well.
I did a fresh install of Windows 7 recently. Windows Update didn't offer up the ATI driver I needed.
Just Google for it.
http://www.lockergnome.com/blade/2009/04/13/microsoft-says-vista-windows-7-are-the-most-secure-os-ever/
They make those statements all the time.
When performing a Windows 2000 or XP install, it prompts you to name a user, which is an administrator account.
So it is designed by default to log you in as an administrator.
There are several linux distros that won't let you log into gdm/kdm as root. Windows was designed for users to login as administrators.
Microsoft is trying to change that mentality with Vista and 7, except too many applications are having issues with UAC. What Microsoft should have done is said, "you're not allowed to claim your application works with Vista and 7 unless it behaves nicely with UAC."
Even better, it should be following a proper UNIX-esque security model. It could create users/groups for specific escalation. Apps shouldn't ask to escalte to administrator level. They should ask only to escalate the rights they specifically need, such as writing to C:\Program Files\Foo\.
Microsoft is happy to blame the users, but it is Microsoft who established the industry standards. They set the table. They tell the users how to use their OS, and they tell developers how to develop for their OS. If Microsoft shipped a more secure design from the get-go, we wouldn't have as many issues. I'm sure malware authors would still target the market-share king and eventually find chinks in the armor, but right now it is so easy to target Windows that every script-kiddie on the planet pulls it off with ease.
openSUSE 11.2 ships with AppArmor and is stable today.
It is a very capable OS and extremely secure.
I'll even give you a free copy. Just don't tell anyone.
http://download.opensuse.org/distribution/11.2/iso/
Microsoft claimed that IE 8 was the most secure browser on the planet, and that Windows 7 was the most secure OS ever. It clearly isn't their fault for making an insecure OS that is subject to malware.
That is simply impossible.
The article made one bagel joke about the future.
I made a different bagel joke about bilking people for a 6 billion dollar upgrade. Funny is very subjective. To each their own.
However, I was under the impression that reading TFA is strictly forbidden on /. Anything that is worth knowing can be gleemed from the unerring wisdom of the discussion threads.
By having OSS drivers internal to the kernel itself, when there is any chance to the Linux driver API, they can grep their way through the code for all the drivers and make necessary changes.
What gets me is that Microsoft doesn't hold the source for all Windows drivers, but they do sign all of them. Why can't they at the very least check for proper API compatibility, and also keep them all in a central repository. If Microsoft has signed the driver, it should be available via Microsoft/Windows Update. A few drivers are available there, but not most, and often they are the wrong versions.
Microsoft doesn't want to be tied to hard deadlines of a fixed release schedule, but if people automatically knew ahead of time to budget for an OS upgrade at a set period of time, they'd come to accept it.
I also think a steady stream of quicker service packs every six months would make it easier to keep more consistent patch levels, and make support easier.
I know 2000 was the first to ship with the NT kernel (other than NT itself) but I thought ME included some NT-based dlls and was starting the shift in that direction.
http://maps.google.com/maps?&q=higgs+boson
Does this mean I get $6 billion?
I suspect Kilgore Trout is somehow responsible.
Okay. Color me stupid, but aren't there existing hardon colliders that function just fine, like the one in Brookhaven?
What they do right to get theirs to launch without all these issues?
As far as I know, the major difference with the LHC is scale.
Was it another bird with a bagel crumb?
The first 6 billion dollars doesn't include bagel protection. That is an additional 6 billion.
This is /.
I'm just warming up for my Trash-80 jokes.
Look at the kernel changes between Server 2008 and Server 2008 R2 and tell me it is the same kernel. It is easily twice the change of 2003 to 2003 R2.
Microsoft is HORRIBLY inconsistent with their naming. Windows 7 reports it runs kernel 6.1, but the entire purpose of naming it Windows _7_ is that is the 7th major iteration of the NT kernel.
Then again, Windows 2000 through Windows Home Server all are NT 5.x kernels. You're not going to suggest that XP was merely a point release and not a major release were you?
If we just went by that naming convention, then it the NT 5.x kernel line lasted 7 years.
Professional is designed to work in a domain. Home users won't be running a domain unless they have purchased Windows Server to run a domain controller. Ulimate seems really unnecessary. I say that running Windows 7 Ultimate. The only need for Windows 7 Ultimate I can imagine is getting the free VM of XP inside of 7 Ultimate (which doesn't ship on the DVD, you have to download it). And even that you can replicate with an existing XP license.
Isn't this late 2009? How is late 2012 four years later?
Are you doing math on a Pentium?
ME was a kludge relesae, with 2000 being the proper release for that goal (merging the Windows 9x GUI with the NT kernel). And you can argue that Vista was a kludge release (trying to implement UAC and a better security model, but selling the OS to the masses with Live integration and Aero interface), with 7 being the proper release of Vista's goals.
If you throw out ME and Vista, Windows has had three major releases in 9 years, and plans the next in 3 years.
openSUSE used to be a 6 month window. I think Mandriva still uses a 6 month window.
If you only count community distros, then Ubuntu, Debian, Fedora, openSUSE and Mandriva are the heavy hitters. Debian would be the exception.
However, this is a poor comparison to Windows. Windows wants you to pay money for the upgrade. Linux releases quicker, but it is a free upgrade. Many of these are more comparable to a Windows service pack.
If you count commercial distros like SLES and Red Hat, you'll see a much longer window between releases. These are more comparable to Windows releases.
Don't update your OS then. I'm sure Window 2000 does everything you'll ever need. Why upgrade past that?
They should have been working on Windows 8 nine months ago, or whever they basically put Windows 7 into freeze. It was over a year ago that they decided certain major features weren't making it into 7.
The team that develops the OS should be focused on the new version right away.
Surely Microsoft is a well-run corporation with long term planning. Surely they have a future roadmap of where they want Windows to go over the next 5 years. Surely Vista and 7 were intentional stepping stones along their master plan.
I can't fathom the possibility that Microsoft has become this un-agile behemoth that no longer innovates, but rather has knee-jerk reactions to the OS market.
What about offline caching?
I thought that was part of Chrome OS so that it can still be used offline?