Console games are optimized for specific hardware, that is the advantage of having a concrete target knowing things like memory size, clocks and bandwidth, number of compute units, system memory size, cpu speed, number of cores, speed of cores, cache size, etc... If what you do is just increase the clocks then you can maintain perfect backwards compatibility with software optimized for that GPU while increasing performance. If you start changing things like number of cores then you need a different approach to optimizing software for that target.
This is pretty well established in the games industry for a very long time.
Applications can get the DPI settings directly from X.
It's what that value is for FFS.
DPI is only one part of it, the important other part is scaling. Do all display servers, WMs and frameworks have a common way in which DPI and scaling should be interpreted for dpi-unaware and dpi-aware applications as well as how this is dealt with in multi-monitor setups?
IMO the way Windows handles DPI and scaling in multi-monitor setups is pretty crappy.
Yes this is what I mean, there's no standard way. Everybody comes up with their own way, so if you look at Fedora you can only do integer scaling even though libraries like EFL support floating point scaling for example. Is there a specification for a standard way that all these different windowing kits and libraries handle dpi unaware, dpi system aware and multi-screen dpi aware? Otherwise everybody just goes off and does their own thing and you end up with systems that people expect to work together that result in a poor user experience.
It's good news that they are moving to address things like this that are standard in other operating systems and to hopefully standardize on a way of doing it so we don't end up with everybody coming up with their own incompatible way of doing it. The issue will really come down to application support though, with so many different application frameworks on Linux they all need to support it, do so in the same way as GNOME and also then make sure that applications are written to support it. Application support for it in Windows is still mixed and that doesn't have the range of GUI frameworks and toolkits that Linux does, on macOS it's better but still not perfect.
It will be a long time for it to become commonly available and just an expected thing rather than a special feature but you have to start somewhere. I really just hope we don't end up with 10 different ways to do it though.
Does that mean this thing - like iDevices - can't be requested to play music from services other than Apple Music?
This is not really correct. I don't use Apple Music; and when I tell my iPhone to "open Pandora", it works just fine.
It is correct, I specifically wrote "play music", you can tell Siri to open any app but you cannot tell it to play music from that app, that only works with Apple's music service and no other.
I can basically already pretty much accomplish this with my iPad or iPhone, using airplay to my own speakers connected to my relatively cheap third-gen Apple TV or relatively cheap Airport Express.
Does that mean this thing - like iDevices - can't be requested to play music from services other than Apple Music? If so that's pretty shit, Echo, Google Home and Sonos all support multiple services even though the 2 former ones both have competing music services.
In general, Apple's stance on these issues is that the anonymize everything, and encrypt everything, so that they can't tie requests to users.
But we're talking about a system like Siri so by definition Apple needs to be able to decrypt what you send them and know who to send the results back to.
By local, I meant that the commands are processed locally on the data stored on the device, as opposed to the command AND all relevant data sent to the remote server.
What it needs is context and the more context you expect it to have the more information you need to provide. Simple things that are just effectively voice commands need very little context.
So yeah, the remote server is still issuing the commands to the device. That's unavoidable. But as long as the actually data is processed locally on the device, that's what I was referring to as 'local'.
I'm not sure how the examples you listed would work any other way.
I take your point but the future is leveraging connectivity such that you can have access to all content wherever you are on whatever device you are using with the ability to download for offline access for when you may not have network connectivity (plane ride, camping, etc).
It's by no means a perfect solution to cover all use cases and is improving over time but it suits the vast majority just fine even now.
It makes sense that Siri (and Cortana, Google whatever they call it) do their voice processing in a data centre cause there is simply not enough horsepower (yet) on the device itself to do that kind of work.
Well it's not just processing the voice to translate what was said, the system needs to interpret what the intention of the action is. You don't learn that effectively by just doing it on individual devices, you need large datasets to understand it. So that it can process the voice, try and understand the intention and then get feedback from the user as to whether this was actually correct.
But... what happens after that? If the entire conversation involves the data center telling the phone what the user wanted, and the phone then delivers, it's still effectively local.
That isn't local at all, It's the complete opposite. What you described is the remote server gets the all the voice data of what you said, the remote server translates that into understanding the intention and then the remote server sends the intention back to the phone which is the remote server controlling the phone. What significant bit about that is "local"?
I'm personally inclined to believe that this is how Siri works, because Siri responds very well to localized commands, such as "play such and such song" or, "create a reminder to do blah".
Those aren't localized commands at all. In the song example your voice command is sent to the server, it interprets it to understand its meaning then sends the command back to your phone with the action to "play" and the name of the song. All your phone is doing is looking through your song list to see if you have that song, and if so playing it.
When I tried to do anything with google assistant, it just wouldn't handle those kinds of details
What kinds of details? It couldn't play a song that you had locally? It couldn't create a reminder? I doubt that.
More to the point, the reason so few people have purchased Windows phones is the phones lack the software people want to run. Android, and iOS, do not.
That's the problem in any established market. Windows Phone was actually not a bad operating system but it wasn't disruptive or innovative either. If you want to be successful in an established market you need seamless compatibility (which seems unlikely) or disruptive innovation. No new OS (Windows Phone, Maemo, Meego, webOS, FirefoxOS, etc) has had either of those things in the current iOS/Android smartphone market.
There's value in doing it for the wider community, I use iOS and its initial multi-tasking UI was absolute garbage. Windows Phone had a great UI for it and then Apple went and copied it for iOS which is great! I've also never had an Android phone as my primary device but they had the notification center, something iOS lacked that proved very useful for customers so when Apple copied that for iOS too it was of benefit to me and all other iOS users even if we didn't touch Android or Windows Phone.
If they don't come out with some kind of disruptive innovation then it will fail just like pretty much every other smartphone platform that isn't iOS or Android but if they do it will be a net win for users: Whether we go from a duopoly to a 3-way competitive market or this disruptive innovation is just consumed by the incumbents.
Yes, pretty much. If you've created an image of the layout to the point that machine learning can interpret it then you might as well have just created the layout in the first place.
Well that all depends on what Bill Gates is getting out of it. A lot of his "philanthropic" endeavors are really just tax dodges which are intended to benefit himself more than who he's claiming to help.
This isn't a philanthropic effort or a tax dodge, from what I can see there it is investing in a funding round of a for-profit company.
Slashdot Mirror
Console games are optimized for specific hardware, that is the advantage of having a concrete target knowing things like memory size, clocks and bandwidth, number of compute units, system memory size, cpu speed, number of cores, speed of cores, cache size, etc... If what you do is just increase the clocks then you can maintain perfect backwards compatibility with software optimized for that GPU while increasing performance. If you start changing things like number of cores then you need a different approach to optimizing software for that target.
This is pretty well established in the games industry for a very long time.
It's very much likely given the need to maintain compatibility.
Spending $500 this close to the PS5 when I find I wouldn't benefit much from PS4 Pro seems like a bad decision.
When is the PS5 being released?
If it's the same GPU with a higher clockspeed then of course pointing out that clockspeed is relevant in terms of performance.
I don't know and I don't really care, the reality is many of them exist and are in use widely.
Applications can get the DPI settings directly from X. It's what that value is for FFS.
DPI is only one part of it, the important other part is scaling. Do all display servers, WMs and frameworks have a common way in which DPI and scaling should be interpreted for dpi-unaware and dpi-aware applications as well as how this is dealt with in multi-monitor setups?
IMO the way Windows handles DPI and scaling in multi-monitor setups is pretty crappy.
Yes this is what I mean, there's no standard way. Everybody comes up with their own way, so if you look at Fedora you can only do integer scaling even though libraries like EFL support floating point scaling for example. Is there a specification for a standard way that all these different windowing kits and libraries handle dpi unaware, dpi system aware and multi-screen dpi aware? Otherwise everybody just goes off and does their own thing and you end up with systems that people expect to work together that result in a poor user experience.
How does Enlightenment handle HiDPI UI scaling with GTK applications for example?
It's good news that they are moving to address things like this that are standard in other operating systems and to hopefully standardize on a way of doing it so we don't end up with everybody coming up with their own incompatible way of doing it. The issue will really come down to application support though, with so many different application frameworks on Linux they all need to support it, do so in the same way as GNOME and also then make sure that applications are written to support it. Application support for it in Windows is still mixed and that doesn't have the range of GUI frameworks and toolkits that Linux does, on macOS it's better but still not perfect.
It will be a long time for it to become commonly available and just an expected thing rather than a special feature but you have to start somewhere. I really just hope we don't end up with 10 different ways to do it though.
Does that mean this thing - like iDevices - can't be requested to play music from services other than Apple Music?
This is not really correct. I don't use Apple Music; and when I tell my iPhone to "open Pandora", it works just fine.
It is correct, I specifically wrote "play music", you can tell Siri to open any app but you cannot tell it to play music from that app, that only works with Apple's music service and no other.
I can basically already pretty much accomplish this with my iPad or iPhone, using airplay to my own speakers connected to my relatively cheap third-gen Apple TV or relatively cheap Airport Express.
Does that mean this thing - like iDevices - can't be requested to play music from services other than Apple Music? If so that's pretty shit, Echo, Google Home and Sonos all support multiple services even though the 2 former ones both have competing music services.
In general, Apple's stance on these issues is that the anonymize everything, and encrypt everything, so that they can't tie requests to users.
But we're talking about a system like Siri so by definition Apple needs to be able to decrypt what you send them and know who to send the results back to.
Nobody bought SGI systems to run crude 3D accelerated games.
Someone forgot to tell the folks at ID Software.
http://itrunsdoom.tumblr.com/post/99687965744/sgi-workstations-yeah-they-run-doom-during-the
Doom isn't 3D accelerated and - as pointed out quite explicitly in the link you posted - they did it "for funsies" because Nobody* bought SGI systems to run crude 3D accelerated games.
ftp://ftp.sgi.com/sgi/quake/intro.html
Quake indeed is 3D accelerated but this was again done for fun and not a commercial product because of the quote above.
https://www.geek.com/games/joh...
Not even sure why you think this is at all relevant, perhaps you didn't read it but yes John Carmack once used an SGI monitor.
By local, I meant that the commands are processed locally on the data stored on the device, as opposed to the command AND all relevant data sent to the remote server.
What it needs is context and the more context you expect it to have the more information you need to provide. Simple things that are just effectively voice commands need very little context.
So yeah, the remote server is still issuing the commands to the device. That's unavoidable. But as long as the actually data is processed locally on the device, that's what I was referring to as 'local'.
I'm not sure how the examples you listed would work any other way.
I take your point but the future is leveraging connectivity such that you can have access to all content wherever you are on whatever device you are using with the ability to download for offline access for when you may not have network connectivity (plane ride, camping, etc).
It's by no means a perfect solution to cover all use cases and is improving over time but it suits the vast majority just fine even now.
It makes sense that Siri (and Cortana, Google whatever they call it) do their voice processing in a data centre cause there is simply not enough horsepower (yet) on the device itself to do that kind of work.
Well it's not just processing the voice to translate what was said, the system needs to interpret what the intention of the action is. You don't learn that effectively by just doing it on individual devices, you need large datasets to understand it. So that it can process the voice, try and understand the intention and then get feedback from the user as to whether this was actually correct.
But... what happens after that? If the entire conversation involves the data center telling the phone what the user wanted, and the phone then delivers, it's still effectively local.
That isn't local at all, It's the complete opposite. What you described is the remote server gets the all the voice data of what you said, the remote server translates that into understanding the intention and then the remote server sends the intention back to the phone which is the remote server controlling the phone. What significant bit about that is "local"?
I'm personally inclined to believe that this is how Siri works, because Siri responds very well to localized commands, such as "play such and such song" or, "create a reminder to do blah".
Those aren't localized commands at all. In the song example your voice command is sent to the server, it interprets it to understand its meaning then sends the command back to your phone with the action to "play" and the name of the song. All your phone is doing is looking through your song list to see if you have that song, and if so playing it.
When I tried to do anything with google assistant, it just wouldn't handle those kinds of details
What kinds of details? It couldn't play a song that you had locally? It couldn't create a reminder? I doubt that.
Even Siri is doing less and less on Apple's servers and more locally.
Siri can't do anything without a connection to Apple's servers. Put your iPhone in Airplane Mode and then Activate Siri:
Siri not available.
You are not connected to the internet.
Why is that a problem?
Well I dont see how a company is going to break into an established market without overcoming it.
I don't want to have a different eco system appear every other year because it offers a small feature that the current one does not.
Right, that's what has been constantly happening and is why these new platforms fail, a small feature isn't going to entice anybody to change.
The platform should be an open one where anyone can interact.
We should also have world peace, among many other things.
Yeah Target has never had a massive breach of customer data...
More to the point, the reason so few people have purchased Windows phones is the phones lack the software people want to run. Android, and iOS, do not.
That's the problem in any established market. Windows Phone was actually not a bad operating system but it wasn't disruptive or innovative either. If you want to be successful in an established market you need seamless compatibility (which seems unlikely) or disruptive innovation. No new OS (Windows Phone, Maemo, Meego, webOS, FirefoxOS, etc) has had either of those things in the current iOS/Android smartphone market.
There's value in doing it for the wider community, I use iOS and its initial multi-tasking UI was absolute garbage. Windows Phone had a great UI for it and then Apple went and copied it for iOS which is great! I've also never had an Android phone as my primary device but they had the notification center, something iOS lacked that proved very useful for customers so when Apple copied that for iOS too it was of benefit to me and all other iOS users even if we didn't touch Android or Windows Phone.
If they don't come out with some kind of disruptive innovation then it will fail just like pretty much every other smartphone platform that isn't iOS or Android but if they do it will be a net win for users: Whether we go from a duopoly to a 3-way competitive market or this disruptive innovation is just consumed by the incumbents.
Yes, pretty much. If you've created an image of the layout to the point that machine learning can interpret it then you might as well have just created the layout in the first place.
Also bearing in mind the thing you are generating is a layout and it's being generated from a screenshot of that exact thing.
What might have a bit of value is generating real layouts from UI mockups.
Well that all depends on what Bill Gates is getting out of it. A lot of his "philanthropic" endeavors are really just tax dodges which are intended to benefit himself more than who he's claiming to help.
This isn't a philanthropic effort or a tax dodge, from what I can see there it is investing in a funding round of a for-profit company.
I disagree. It's good PR and it's tax deductable. Win-win ... for Gates.
I don't think investing during a funding round of a for-profit company is a valid tax deduction.