Jump to content
Posted 13 March 2013 - 15:38
Posted 13 March 2013 - 15:48
A microkernel? Microsoft has toyed with the idea in the past, but I don't think any shipping version of Windows has featured a microkernel. The Windows NT kernel architecture is inherently monolithic. Wikipedia seems to support this assertion.
Posted 13 March 2013 - 20:43
read the kernel part of that wiki document again. while the kernel has advanced from the original microkernel, the actual kernel, layered between executive and HAL is referred to as a microkernel.
Posted 14 March 2013 - 00:02
Yea I would say NT is more a hybrid kernel. Also not everything that runs in kernel mode is the kernel. Windows has the ability to run third party code in kernel mode. Linux is a better example of an actual monolithic kernel.
From what I understand, everything that runs in kernel mode is part of the kernel. What constitutes a microkernel is services running in user mode that are essential to kernel operation (i.e. services that would otherwise run in kernel mode). Since Linux internally separates the HAL, VMM, and other essential kernel components, does that make it a microkernel too? I admit, this argument depends entirely on your definition of what exactly constitutes the "kernel". The definition seems somewhat malleable considering modern abstracted software design. The introduction of the "Architecture of Windows NT" Wikipedia article seems to compromise on this point too by claiming, "The Windows NT kernel is known as a hybrid kernel."
Posted 18 March 2013 - 21:21
The X.Org Foundation hasn't firmly decided on their position of Canonical's Mir Display Server versus Wayland.
The meeting logs for an X.Org Foundation Board of Directors' IRC meeting from earlier this month have finally been published to the X.Org Wiki.
Oracle's Stuart Krietman began by asking "does anyone here take seriously the 'threat' of Mir?" Bart Massey responded, "Wayland is something we have agreed is part of the X.Org umbrella; Mir definitely is not." But he then said, "I would suggest we make a highly supportive and encouraging announcement [regarding] Mir...Emphasize that we think that it would be fantastic if Canonical could replace X on desktop in 12 months... All I meant is that we haven't yet taken a position on Mir."
Bart also added, "Given that Mir is a single-corporation effort, I don't think it's appropriate for us to be supporting it, but that's just my position..." Intel's Eric Anholt wrote, "I don't imagine many people seeing mir as the next big thing." AMD's Alex Deucher commented, "I mean if mir took off, we could support it as well."
This meeting took place two weeks ago and there was an action item for the foundation to issue a public statement about Mir, but so far that's been non-existent.
Posted 18 March 2013 - 21:26
Posted 19 March 2013 - 00:45
Posted 19 March 2013 - 04:56
I don't think I've seen any issue that enraged the Linux community like this in a long time.
Posted 19 March 2013 - 07:04
Yea...this whole thing makes very little sense to me.
Posted 19 March 2013 - 07:50
(Sorry in advance if this has be said before).
It's a strange form of vendor lock in, apps that depend on the Mir API won't be able to run on other Linux distros (Compare that to Wayland which can be implemented by anyone on any OS)
People are upset mainly because Canonical is doing their own thing with no input from anybody else while also saying the competing projects are bad.
Posted 19 March 2013 - 09:01
Posted 19 March 2013 - 10:36
It's a strange form of vendor lock in, apps that depend on the Mir API won't be able to run on other Linux distros