Linked by Thom Holwerda on Sat 15th Jan 2011 10:40 UTC
Yesterday, the ninth Firefox 4.0 beta was released. One of the major new features in Firefox 4.0 is hardware acceleration for anything from canvas drawing to video rendering. Sadly, this feature won't make its way to the Linux version of Firefox 4.0. The reason? X' drivers are "disastrously buggy". Update: Benoit Jacob informed my via email that there's some important nuance: hardware acceleration (OpenGL only) on Linux has been implemented, but due to bugs and issues, only one driver so far has been whitelisted (the proprietary NVIDIA driver).
Thread beginning with comment 459203
To view parent comment, click here.
To read all comments associated with this story, please click here.
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[11]: Reverse engeneering sucks
by lucas_maximus on Thu 20th Jan 2011 18:46
in reply to "RE[10]: Reverse engeneering sucks"
Member since:2009-08-18
GPL or it's you problem.
Why should it be? I thought GPL gave you freedom and doesn't tie you to the developer ... except with the Linux kernel you have to bend to kernel devs whims or fork the code (ala Android).
That depends greatly how clear it is marked depreciated.
This is not really a problem with the process. My point still stands.
The whole graphics stack is going through a revolution. This is a good thing. X run as the user. Nice VTTY switching. Lots of shared code, thus shared improvements in the future. X alternatives possible (like Wayland).
Again you branch it and actually have a "unstable branch" ... so I don't understand why code with some obvious flaws is being put into a new release. BSDs do this quite well .. they have release, stable, current.
Also If I released code that had obvious flaws at work I would get pulled into my managers office, but because it is Linux it is suddenly perfectly okay.
1% on the desktop. Check out the phone market (counting Android as Linux, which it basically is), or GPS, or routers, or TVs, or web server, or super computers. You name it, Linux has a big market share, if not dominance. So the approach clearly works.
People don't see Linux, they just see Android, Chrome OS whatever it is.
It is used because it is cheaper than trying to customize something else for the purpose. No because the development model is "better".
Also on embedded devices they are usually using a specific version of kernel (which has been throughly tested to work), because this version of the kernel happens to work well with the device.
Doesn't prove anything about the success of the "kernel development method".
Graphics cards are clearly desktop only, so surprise surprise, the support isn't good. But as I said, the graphics stack is going through a revolution right now. I know you are going to come back with "it always is" but this is really massive, not just general change.
Again why is this being put into kernels that are deemed to be "stable", when there are massive and obvious flaws?
The developers know what they are doing. If Linux was like you seem to think, it wouldn't have been so massively successful as it is has (outside the desktop).
I whole heartly disagree. It because it is free and it costs less for business to customize the kernel than it is to write their own.
In the closed world that is required. In the open world, it's not. You change the API, you go and change all the drivers too. Notify the maintainers, etc etc. Things can happen much quicker if the code is in the trunk.
Why? When I build code with .NET 2, I know it will work with version 3.0, 3.5, 4.0 because the API is kept consistent ... this is because of good design of the API, it doesn't need to change constantly.
.NET 1.1 and .NET 1.0 had obvious flaws in it, and these were fixed ... after that the core API has been extended ... not reworked (which is what happened between 1.1 and 2.0).
Blame the maintainers. It's not Intel don't have the staff. They don't allocate enough, and it's the 1% desktop market share that is why. As things settle down to normal levels of churn, Intel won't need to work as much as they aren't now.
If they kept a stable ABI intel wouldn't have to keep on reworking it ... which is exactly my point. Thanks for proving it.
Also everytime someone says this to me, I point out that they work perfectly in OpenBSD ... which have an even smaller market share than linux and has less devs. Mayeb they actually know what they are doing.
RE[12]: Reverse engeneering sucks
by jabjoe on Thu 20th Jan 2011 20:33
in reply to "RE[11]: Reverse engeneering sucks"
Member since:2009-05-06
Agh too much to answer point by point. Look, if the world was like you thought, someone would do stable wrapper round the unstable API, and everyone would use it. There would be pressure for it to be rolled into Linux itself. This hasn't happened, or shown any signs of happening. You point about .NET is completely unrelated. That's userland. Linux's userland interface is very stable. Christ, I've grabbed old Unix apps and found they just compile under Linux and run, which shocked even me. I think the thing is you are seeing the kernel and drivers as separate things, so you are thinking of a interface across them. But the developers decided that was a bad model, they decided best have it one thing (the trunk) so the interfaces have no need to be stable, and can change to what ever is best. You say there is no proof that it's wide and varied use is to do with this model working, and it's used because it's free/cheap. Then why don't they use one of the other free OS? Why has none of them even close to the device support or Linux? Even though they existed before Linux? I'll tell you why, because Linux went critical mass in the way they didn't. This is because of the stickiness of the GPL. The GPL does tie you in and that is deliberate. It's BSD that doesn't. Forking is not only allowed but encouraged. Forked get merged, forks take over, fork die. It's an organic process. I'm not saying Linux is the best in all ways, I hate the un-Unix like parts (ALSA and lack of /dev/eth0), and no generic fs device sharing like Plan9, but you can't fault it on hardware support or speed. Graphics is sticky, but is being delt with, even though the drivers are finished, these new drivers are already taking over, and you can't stop that because it's open source, it's the distros' choice.
News
Linked by Thom Holwerda on 06/19/13 23:02 UTC, submitted by M.Onty
"Microsoft has sensationally abandoned its controversial plans to restrict the sharing of XBox One games, and has also removed daily online authentication requirements for its forthcoming console", reports The Guardian. They had no choice. Still a good move.
Linked by Thom Holwerda on 06/19/13 22:28 UTC
"A snippet of code in the Steam digital distribution platform has revealed that Valve may be planning to let users easily share their games with friends in the future. The Verge has verified the code's authenticity, which was originally spotted by a member of the NeoGAF gaming forum; the code references a 'shared game library' and a notification that would alert a user when their games are currently in use by a borrower." This would change the, uh, game.
Linked by Thom Holwerda on 06/18/13 22:33 UTC
Official Apple statement on PRISM and privacy: "Regardless of the circumstances, our Legal team conducts an evaluation of each request and, only if appropriate, we retrieve and deliver the narrowest possible set of information to the authorities. In fact, from time to time when we see inconsistencies or inaccuracies in a request, we will refuse to fulfill it." This is basically Apple re-publishing their earlier statement in a more official manner. You either believe it, or you don't.
Linked by Anonymous on 06/18/13 22:26 UTC
The open source Contiki operating system and its commercial Thingsquare distribution are making strides towards the connected home. According to a Computerworld article, a new LED light bulb manufacturer is putting small computers into their light bulbs. Each computer runs the Contiki OS, which gives each bulb an IP address and allows them to be controlled with a smartphone application. To connect the bulbs to the application, the bulbs use both Wi-Fi and the much lower power IEEE 802.15.4 mesh technology.
Linked by Thom Holwerda on 06/18/13 22:25 UTC
"Google asked the secretive Foreign Intelligence Surveillance Court on Tuesday to ease long-standing gag orders over data requests it makes, arguing that the company has a constitutional right to speak about information it's forced to give the government. The legal filing, which cites the First Amendment's guarantee of free speech, is the latest move by the California-based tech giant to protect its reputation in the aftermath of news reports about sweeping National Security Agency surveillance of Internet traffic." Draining the ditch after the cow has drowned in it.
Linked by Thom Holwerda on 06/18/13 17:45 UTC
"I can't find one person who has been using the Nexus 7 for an extended period of time, and hasn't seen a massive downgrade in performance. Just what kind of downgrade are we talking here? I cannot pick up my Nexus 7 without experiencing problems like a lag of ten seconds, or more, just to rotate the display; touches refusing to acknowledged; stuttering notification panel actions; and unresponsive apps." Fully and utterly agreed. My Nexus 7 was blazing-fast and awesome for a few months, and at some point, it just started sucking. Just like that. I've tried loads of ROMs, and nothing helps.
Linked by Thom Holwerda on 06/18/13 17:32 UTC, submitted by poundsmack
A new release of the hobby operating system MenuetOS! The release notes are a bit non-descript, but you'll have to take what you can get: "updates and improvements (picview, httpc, ehci, ...)".
Linked by Thom Holwerda on 06/17/13 17:58 UTC
"The Internet is one of the most transformative technologies of our lifetimes. But for 2 out of every 3 people on earth, a fast, affordable Internet connection is still out of reach. And this is far from being a solved problem. There are many terrestrial challenges to Internet connectivity - jungles, archipelagos, mountains. There are also major cost challenges. Right now, for example, in most of the countries in the southern hemisphere, the cost of an Internet connection is more than a month's income. Solving these problems isn't simply a question of time: it requires looking at the problem of access from new angles. So today we're unveiling our latest moonshot from Google[x]: balloon-powered Internet access." Insane.
Linked by Thom Holwerda on 06/17/13 17:52 UTC
"MineAssemble is a tiny bootable Minecraft clone written partly in x86 assembly. I made it first and foremost because a university assignment required me to implement a game in assembly for a computer systems course. Because I had never implemented anything more complex than a 'Hello World' bootloader before, I decided I wanted to learn about writing my own kernel code at the same time. Note that the goal of this project was not to write highly efficient hand-optimized assembly code, but rather to have fun and write code that balances readability and speed. This is primarily accomplished by proper commenting and consistent code structuring." Just cool.
Linked by Thom Holwerda on 06/14/13 21:03 UTC
Jon Rubinstein, former CEO of Palm: "Well, I'm not sure I would have sold the company to HP. That's for sure. Talk about a waste. Not that I had any choice because when you sell a company you don't get to decide that. Obviously, the board and shareholders decide that. If we had known they were just going to shut it down and never really give it a chance to flourish, what would have been the point of selling the company? I think the deal we had with Verizon really hurt us, but who knew that at the time? These things are all hindsight."More News »
Sponsored Links
Member since:
2009-05-06
GPL or it's you problem.
That depends greatly how clear it is marked depreciated.
The whole graphics stack is going through a revolution. This is a good thing. X run as the user. Nice VTTY switching. Lots of shared code, thus shared improvements in the future. X alternatives possible (like Wayland).
This public big deal noise from Mozilla may even speed the changes up. Lots of other project have managed OpenGL fine in these times of transition, so Mozilla might be making more noise then is justified.
1% on the desktop. Check out the phone market (counting Android as Linux, which it basically is), or GPS, or routers, or TVs, or web server, or super computers. You name it, Linux has a big market share, if not dominance. So the approach clearly works.
Graphics cards are clearly desktop only, so surprise surprise, the support isn't good. But as I said, the graphics stack is going through a revolution right now. I know you are going to come back with "it always is" but this is really massive, not just general change.
Of course they should fix things ... but not in a manner that is likely to cause more bugs.
The developers know what they are doing. If Linux was like you seem to think, it wouldn't have been so massively successful as it is has (outside the desktop).
Functional Testing will not catch all the bugs.
I tried to be clear that won't be the only testing.
In the closed world that is required. In the open world, it's not. You change the API, you go and change all the drivers too. Notify the maintainers, etc etc. Things can happen much quicker if the code is in the trunk.
Blame the maintainers. It's not Intel don't have the staff. They don't allocate enough, and it's the 1% desktop market share that is why. As things settle down to normal levels of churn, Intel won't need to work as much as they aren't now.
Just today, there was some more good news on the Gallium3D front:
http://www.phoronix.com/scan.php?page=article&item=amd_r500_expande...