I know, but language isn't the same as math. Three times faster means 300%.
No, it absolutely does not. If you are saying 1.4x (140%) faster when you actually mean it is 0.4x (40%) faster then your language is irrefutably wrong. By your logic saying it is 40% faster would actually mean it was slower, which is wrong.
Mantle was being used in commercial titles before Metal to prove the concept and was then offered as a basis for glNext as AMD is a contributor to Khronos. If inaction on the part of Khronos (which Apple is a part of, so I'm not sure where you're drawing the line here) then why did Apple not just offer Metal as a candidate for glNext or even release Metal as an open standard?
Do we know when Vulkan was conceived? The trademark for Vulkan was filed in February
The initiative before that was called glNext but discussions about low driver overhead APIs have been going on long before the Vulkan name was decided.
I don't know the behind the scenes dynamics at Khronos or Apple.
Apple is a member of the Khronos group.
Metal has been worked on since at least 2013, when the betas for iOS 8 were being written. Metal as an API has been ready to ship since iOS 8 came out. Vulkan's still in the proof of concept stage and might replace or augment Metal in the future.
And yet they didn't propose it for glNext, they just went and created their own incompatible proprietary API.
This. If you want to design to open standards, OpenGL/OpenGL ES gives you (we hope) the best performance you can reasonably get within the constraints of a platform-neutral system.
No, the whole reason for Vulkan is because OpenGL currently doesn't give you the best performance which is why the switch to lower-overhead APIs is happening but instead of collaborating - like AMD is doing - they are going off in their own non-standard direction.
If you care about that last few percent performance-wise, you use Metal, because that's as close as you're going to get to the hardware.
Just as long as Apple continues to maintain OpenGL and OpenGL ES with reasonable performance, this is a very useful thing to have around. Now if they start to ignore OpenGL in favor of their own API, then we should be concerned, but I doubt they will do that.
Yeah great, back to the days of dealing with OpenGL, DirectX, S3d, Glide, MSI, CIF, NV1, etc... where every vendor writes their own API.
Because anytime that you have an "open standard" that you don't control its always a slower process and not optimized for your needs.
Of course it is a slower process to achieve interoperability between vendors, that doesn't mean we should abandon open standards except as a baseline to create your own proprietary extensions otherwise we end up back in the IE6/Netscape days.
It's the same reason that Google forked WebKit.
No it isn't, WebKit is not an "open standard", it is an implementation designed to present information formatted in an open standard way.
Because Metal has been shipping since iOS 8 and usable now.
Open standards are a collaborative effort, instead of contributing to working with other vendors Apple has just created their own proprietary non-standard API. Whereas AMD created the Mantle API as a prototype implementation of a low overhead GPU API and then contributed that work to the Vulkan development as a baseline for the standard which can be used by all vendors. Instead of just being a leech, be a contributor.
1.4x faster is the same as 40% faster, 140% of the baseline.
That doesn't make any sense, 1.4x is 140% of the baseline. It's really not that complicated: If your baseline is 100% and you want to get to 140% you need to go 40% (or 0.4x) faster than 100%.
Metal: Not a new Graphics Engine. It's been on iOS for two years.
But what is the point of it? Instead of contributing their work to an open standard - like AMD has done with Mantle - they have switched from an open standard and created their own Apple-centric 3D graphics API that is just another case of NIH syndrome.
Duh. It will make it look like linux, bsd, ios, android are broken because they cant connect to a windows box.
That might fool you but most people - particularly if they are remoting into systems - are going to be smart enough to see that an update to Windows didn't break Linux, BSD, iOS, Android, etc...
Or they'll expect remote servers to implement whatever changes Microsoft will require for interoperatibility. We've been through this in the 1990-ies, when Microsoft's Internet Explorer was introduced with subtle incompatibilities in HTML-rendering...
And how has that worked out? Back then that affected personal computing - an area which Microsoft had a monopoly - and it still ultimately failed. This is across desktop, server and mobile, this conspiracy theory of yours has no chance at all, in fact you don't even posit what Microsoft would gain out of it.
Well, a successful attempt is still an attempt: Netscape died.
But it failed, you need to learn your history: Netscape lived on thanks to Mozilla and now we have IE dying in favor of open standards, Microsoft themselves are killing IE in favor a browser that does not even support proprietary extensions like ActiveX.
Or not — depending on the nature of incompatibilities and the marketing/advertising...
Incompatibilities would make people less likely to use Microsoft's implementation, not more likely. You don't seem to understand that this isn't the 90s anymore, Microsoft doesn't even come close to dominating computing these days. Breaking their product just locks them out of the market, not everybody else in.
That, for example, in order to ssh into a remote Windows system you'll have to use Microsoft's ssh-client — because they'll use some funky cipher/digest combination or some other "extension".
In which case people would just use putty or cygwin or openssh instead, creating an incompatibility such that none of your devices can talk to Powershell is bad for Microsoft, it doesn't help them at all.
In the link I gave there is a large list of Microsoft's earlier attempts to kill a standard by first adopting it — read it up...
Firstly I can see why you had to write "attempts", because it seems none of those actually killed anything. But obviously - unless you don't understand what SSH is or its purpose - if they create an incompatibility here it is going to completely break their system making it such that Linux, BSD, iOS, Android, etc... can no longer connect to it.
If they're adapting to a changing world they're "desperate", if they're continuing on their existing course they're a sinking ship. Either way their record revenue and profits in the past year are a bad thing, Microsoft is going down hard which means the Year of the Linux Desktop is right around the corner;)
I mean, they will take OpenSSH, compile-it for Windows, and make sure Power Shell is the default login shell. Then what?
Then they have a vested interest in it so should contribute to maintaining it, even widely-used open source projects (PGP and OpenSSL for example) are often woefully under-resourced. It isn't as though OpenSSH is some bug-free, perfect product that requires no maintenance.
...of course Erwin Schrödinger gave us the thought experiment as a way to explain this lack of a binary answer to those whom it was not obvious many decades ago.
But they have been separating their various products that people depend on from Windows, for example you can run Office on pretty much any platform now through a web browser. So then encumbering the OS with subscription costs makes no sense, all those people who just need Office now don't need Windows so the good thing is Microsoft has got to concentrate on making people want Windows.
Linux does great in situations that don't specifically require running Windows-specific software.
But even then there's little reason to not just use Windows, that's the point. Linux needs to offer something to people other than extreme niches and the lowest common denominator use cases. If in fact you only need email and web browser (though I don't think that's a huge market) then any operating system - even tablet ones - is fine.
It doesn't even need to be software-based, it could be a new hardware formfactor which a desktop Linux vendor could be first to market with. Like you have the iPhone and iPad which of course were followed by Android. And even the Surface and convertible Windows laptops but where is the pioneering innovation on the desktop Linux side? Desktop Linux is still just the operating system you shoehorn in later when the hardware is old.
The tool was obviously compelling enough to prompt the user to set up and maintain a second workstation to use it. In that case, the user is more likely to replace and abandon their old toolchain if they no longer want to (or can afford to) maintain a multi-platform workflow.
But this scenario doesn't exist. That's why there's not even an example of it, that's why in the real world example I gave what happened is the exact opposite of what you're saying.
Slashdot Mirror
I know, but language isn't the same as math. Three times faster means 300%.
No, it absolutely does not. If you are saying 1.4x (140%) faster when you actually mean it is 0.4x (40%) faster then your language is irrefutably wrong. By your logic saying it is 40% faster would actually mean it was slower, which is wrong.
Mantle was being used in commercial titles before Metal to prove the concept and was then offered as a basis for glNext as AMD is a contributor to Khronos. If inaction on the part of Khronos (which Apple is a part of, so I'm not sure where you're drawing the line here) then why did Apple not just offer Metal as a candidate for glNext or even release Metal as an open standard?
Do we know when Vulkan was conceived? The trademark for Vulkan was filed in February
The initiative before that was called glNext but discussions about low driver overhead APIs have been going on long before the Vulkan name was decided.
I don't know the behind the scenes dynamics at Khronos or Apple.
Apple is a member of the Khronos group.
Metal has been worked on since at least 2013, when the betas for iOS 8 were being written. Metal as an API has been ready to ship since iOS 8 came out. Vulkan's still in the proof of concept stage and might replace or augment Metal in the future.
And yet they didn't propose it for glNext, they just went and created their own incompatible proprietary API.
This. If you want to design to open standards, OpenGL/OpenGL ES gives you (we hope) the best performance you can reasonably get within the constraints of a platform-neutral system.
No, the whole reason for Vulkan is because OpenGL currently doesn't give you the best performance which is why the switch to lower-overhead APIs is happening but instead of collaborating - like AMD is doing - they are going off in their own non-standard direction.
If you care about that last few percent performance-wise, you use Metal, because that's as close as you're going to get to the hardware.
Just as long as Apple continues to maintain OpenGL and OpenGL ES with reasonable performance, this is a very useful thing to have around. Now if they start to ignore OpenGL in favor of their own API, then we should be concerned, but I doubt they will do that.
Yeah great, back to the days of dealing with OpenGL, DirectX, S3d, Glide, MSI, CIF, NV1, etc... where every vendor writes their own API.
Because anytime that you have an "open standard" that you don't control its always a slower process and not optimized for your needs.
Of course it is a slower process to achieve interoperability between vendors, that doesn't mean we should abandon open standards except as a baseline to create your own proprietary extensions otherwise we end up back in the IE6/Netscape days.
It's the same reason that Google forked WebKit.
No it isn't, WebKit is not an "open standard", it is an implementation designed to present information formatted in an open standard way.
Is Vulkan ready yet?
Because Metal has been shipping since iOS 8 and usable now.
Open standards are a collaborative effort, instead of contributing to working with other vendors Apple has just created their own proprietary non-standard API. Whereas AMD created the Mantle API as a prototype implementation of a low overhead GPU API and then contributed that work to the Vulkan development as a baseline for the standard which can be used by all vendors. Instead of just being a leech, be a contributor.
1.4x faster is the same as 40% faster, 140% of the baseline.
That doesn't make any sense, 1.4x is 140% of the baseline. It's really not that complicated: If your baseline is 100% and you want to get to 140% you need to go 40% (or 0.4x) faster than 100%.
Metal: Not a new Graphics Engine. It's been on iOS for two years.
But what is the point of it? Instead of contributing their work to an open standard - like AMD has done with Mantle - they have switched from an open standard and created their own Apple-centric 3D graphics API that is just another case of NIH syndrome.
You mean someone uses Windows built-in DVD playback?
No, that's why they're deprecating it.
Duh. It will make it look like linux, bsd, ios, android are broken because they cant connect to a windows box.
That might fool you but most people - particularly if they are remoting into systems - are going to be smart enough to see that an update to Windows didn't break Linux, BSD, iOS, Android, etc...
Yes that's my point, it doesn't matter what they're doing there's always detractors trying to spin it to be a negative.
Or they'll expect remote servers to implement whatever changes Microsoft will require for interoperatibility. We've been through this in the 1990-ies, when Microsoft's Internet Explorer was introduced with subtle incompatibilities in HTML-rendering...
And how has that worked out? Back then that affected personal computing - an area which Microsoft had a monopoly - and it still ultimately failed. This is across desktop, server and mobile, this conspiracy theory of yours has no chance at all, in fact you don't even posit what Microsoft would gain out of it.
Well, a successful attempt is still an attempt: Netscape died.
But it failed, you need to learn your history: Netscape lived on thanks to Mozilla and now we have IE dying in favor of open standards, Microsoft themselves are killing IE in favor a browser that does not even support proprietary extensions like ActiveX.
Or not — depending on the nature of incompatibilities and the marketing/advertising...
Incompatibilities would make people less likely to use Microsoft's implementation, not more likely. You don't seem to understand that this isn't the 90s anymore, Microsoft doesn't even come close to dominating computing these days. Breaking their product just locks them out of the market, not everybody else in.
That, for example, in order to ssh into a remote Windows system you'll have to use Microsoft's ssh-client — because they'll use some funky cipher/digest combination or some other "extension".
In which case people would just use putty or cygwin or openssh instead, creating an incompatibility such that none of your devices can talk to Powershell is bad for Microsoft, it doesn't help them at all.
In the link I gave there is a large list of Microsoft's earlier attempts to kill a standard by first adopting it — read it up...
Firstly I can see why you had to write "attempts", because it seems none of those actually killed anything. But obviously - unless you don't understand what SSH is or its purpose - if they create an incompatibility here it is going to completely break their system making it such that Linux, BSD, iOS, Android, etc... can no longer connect to it.
Are they forgetting who runs OpenSSH?
Most people who have any valuable contributions are more interested in code than historical politicking.
You guys are so jaded, it's insufferable.
If they're adapting to a changing world they're "desperate", if they're continuing on their existing course they're a sinking ship. Either way their record revenue and profits in the past year are a bad thing, Microsoft is going down hard which means the Year of the Linux Desktop is right around the corner ;)
I mean, they will take OpenSSH, compile-it for Windows, and make sure Power Shell is the default login shell. Then what?
Then they have a vested interest in it so should contribute to maintaining it, even widely-used open source projects (PGP and OpenSSL for example) are often woefully under-resourced. It isn't as though OpenSSH is some bug-free, perfect product that requires no maintenance.
Now I'm scared... We may, once again, see Microsoft's approach of Embrace, Extend, and Extinguish in action...
What exactly are you scared about? Assuming you understand what's going on here what part of that concerns you?
So, BB has the patent on little keyboards?
I don't think so, probably a design patent on their particular design.
Who patented little buttons?
Nobody, there are various keyboard cases for touchscreen smartphones that don't have patent issues.
...of course Erwin Schrödinger gave us the thought experiment as a way to explain this lack of a binary answer to those whom it was not obvious many decades ago.
But they have been separating their various products that people depend on from Windows, for example you can run Office on pretty much any platform now through a web browser. So then encumbering the OS with subscription costs makes no sense, all those people who just need Office now don't need Windows so the good thing is Microsoft has got to concentrate on making people want Windows.
Linux does great in situations that don't specifically require running Windows-specific software.
But even then there's little reason to not just use Windows, that's the point. Linux needs to offer something to people other than extreme niches and the lowest common denominator use cases. If in fact you only need email and web browser (though I don't think that's a huge market) then any operating system - even tablet ones - is fine.
It doesn't even need to be software-based, it could be a new hardware formfactor which a desktop Linux vendor could be first to market with. Like you have the iPhone and iPad which of course were followed by Android. And even the Surface and convertible Windows laptops but where is the pioneering innovation on the desktop Linux side? Desktop Linux is still just the operating system you shoehorn in later when the hardware is old.
It also gets released to the manufacturers of new PCs before release.
I wouldn't disagree with that. Though it is what the "Year of the Linux Desktop" crew have been saying for nearly 20 years.
Huh, interesting... "Can I [x]?" questions, lacking any other constraints, actually do have binary answers. Yes you can or no you can not. "
Ok well I just tried to run AutoCAD using Wine and it didn't work, so you're wrong, the answer is no.
The tool was obviously compelling enough to prompt the user to set up and maintain a second workstation to use it. In that case, the user is more likely to replace and abandon their old toolchain if they no longer want to (or can afford to) maintain a multi-platform workflow.
But this scenario doesn't exist. That's why there's not even an example of it, that's why in the real world example I gave what happened is the exact opposite of what you're saying.