Look at this use of a Kinect controller, haiku is in a state right now where this could be implemented. Well if haiku needs a killer app to motivate users, here you go, do this better then everyone else.
a fake video .
Its not a fake. Thats a kinect controller running custom software. its not about that video in particular, its about using modern tools in a new way to interact. Currently nothing really seperates haiku from everything else out there. Not in any meaningful way. I still have to use a keyboard and mouse.
This is all about moving into a new realm of visual computing. Which is a exciting new market. It also would work well with a low latency OS that is tunned for fast response.
but thats ok, if you don’t get it. Its not surprising. PC’s and computers in general are heading this way. Finally the digital horsepower is available to make this a reality.
I dont know if this video is fake or not
but I have seen this thing to work on a PC under Linux,
it could be nice too see kinect on BeOS just for fun.
Or money could be contributed to gettting 3d drivers etc… working. Instead of buying another product from MicroSoft one of the companies that helped bring on the demise of BeOS (Though Be Inc made bad decisions IMO bringing on thier own demise still …)
and what good are 3d driver when you have no user base and no games to run on the 3d drivers ?
See the problem ?
There has to be something new,exciting,different about haiku to bring in users. Microsoft doesn’t make the hardware, though they do resell it, also given the surge in popularity i fully expect to see someone out there competing.
Microsoft is quickly become a IP liscensing company as the software ecosystem continues to diversify.
[quote=forart.it]A real killer application would be a native professional high-definition NLE software, IMHO.
(something like Sony Vegas, for example)[/quote]
A killer application is something that the alternatives can’t do. It’s enough reason for people to choose your platform (in this case, Haiku) over all the others, regardless of other considerations. Cloning something that all the major platforms already do well, is not a “killer application”.
A real killer application would be a native professional high-definition NLE software, IMHO.
(something like Sony Vegas, for example)
1 Like
@thatguy There are acutally opengl games already ported you know the quakes …probably doom3 once released etc. And there isn’t anything stoping anyone from porting thier commercial opengl games later on.
If microsoft didn’t make the kinect who did pray tell… sure they have some company in china producing it but its still an MS product. Considering that you point out IP lisensing… that makes for an even better reason not to fuel their agendas. BTW microsoft probably makes more money off of hardward they sell than Windows itself… which is just an interface to run its Office Products which also bring in the dough.
I disagree and you can see this by looking at what BeOS did. Back in the day, Linux & Windows could play video files. BeOS could too but they wanted to show that BeOS could do it way better. So they played multiple videos without issues while Linux & Windows were staggering playing multiple videos simultaneously. They also created demos where they had different videos playing on a rotating cube and in a book where you could turn the video pages.
A killer application is one where you can do something way better than other OSes. Shows better performance and puts other OSes to shame. You must show a big difference to stand apart because slightly better won’t cut it. It is also wise to choose an application that many users want or is important to them. Something like Sony Vegas - a professional non-linear editing system (NLE) - would only attract audio professionals to Haiku. Kinect is likely bad choice too.
Biggest problem for Haiku is it lacks certain things like hardware OpenGL, hardware HD video decoding, Flash, etc. So you can’t do anything with those. Haiku excels in responsiveness, SMP and multithreading so you want an application(s) that shows those off. Either showing off multitasking between multiple intense applications like what BeOS showed off playing many videos - ie: playing a non-hd video file while converting another video + compiling Haiku (with -j4). Or a program that gets higher performance from SMP (usually conversion or 3D animation programs). Or a graphically intense program that is constantly updating the GUI with lots of changes while doing complex calculations (CPU intensive with threading).
[quote=forart.it]A real killer application would be a native professional high-definition NLE software, IMHO.
(something like Sony Vegas, for example)[/quote]
Sony vegas is nice, but it certainly isn’t comming to haiku. At least not on any timeline I can forsee, but thats still a very very limited market, and the other big OS already has Sony Vegas and I doubt sony would make a port of it. I just don’t see why. A solid bussiness case could be made for sony, but they’d have to invest in haiku to get the hardware ready for it.
Clockwerk is a start, but very very far from being anywhere near as capable as vegas.
I disagree and you can see this by looking at what BeOS did. Back in the day, Linux & Windows could play video files. BeOS could too but they wanted to show that BeOS could do it way better. So they played multiple videos without issues while Linux & Windows were staggering playing multiple videos simultaneously. They also created demos where they had different videos playing on a rotating cube and in a book where you could turn the video pages.
A killer application is one where you can do something way better than other OSes. Shows better performance and puts other OSes to shame. You must show a big difference to stand apart because slightly better won’t cut it. It is also wise to choose an application that many users want or is important to them. Something like Sony Vegas - a professional non-linear editing system (NLE) - would only attract audio professionals to Haiku. Kinect is likely bad choice too.
Biggest problem for Haiku is it lacks certain things like hardware OpenGL, hardware HD video decoding, Flash, etc. So you can’t do anything with those. Haiku excels in responsiveness, SMP and multithreading so you want an application(s) that shows those off. Either showing off multitasking between multiple intense applications like what BeOS showed off playing many videos - ie: playing a non-hd video file while converting another video + compiling Haiku (with -j4). Or a program that gets higher performance from SMP (usually conversion or 3D animation programs). Or a graphically intense program that is constantly updating the GUI with lots of changes while doing complex calculations (CPU intensive with threading).[/quote]
To some dgeree this holds true today, the problem is that most modern hardware less then say 5 years old, can do this to with pretty much any os.
[quote=NoHaikuForMe][quote=forart.it]A real killer application would be a native professional high-definition NLE software, IMHO.
(something like Sony Vegas, for example)[/quote]
A killer application is something that the alternatives can’t do. It’s enough reason for people to choose your platform (in this case, Haiku) over all the others, regardless of other considerations. Cloning something that all the major platforms already do well, is not a “killer application”.[/quote]
this is the first time I agree with you, hence why I started the thread. Imagine having a gesture hand motion controlled OS.
[quote=tonestone57]
I disagree and you can see this by looking at what BeOS did.[/quote]
“What BeOS did” was fail in the market and lose millions of dollars of other people’s money.
There were real applications on Windows playing multiple videos smoothly in that period. But because they were real applications not demos, and they had real users, they weren’t doing presentations to kids about how they were going to change the world. The companies involved are probably still in business, unlike Be.
Are you sure? Excelling at SMP and multithreading are goals not check box items. My impression is that Haiku’s actual performance in those areas remains… lacklustre. And as usual while they were playing the goal posts have moved: 4 way multiprocessing was an exceptionally high-end workstation capability in BeOS’s day, but right now it’s something that comes with mid-range PCs, and the workstations are offering 8-16 way SMP.
“What BeOS did” was fail in the market and lose millions of dollars of other people’s money.
BeOS was selling their OS and having trouble to get buyers. Linux only succeeded in getting their 1-2% market share by giving their OS away for free and after many years of pushing it. Had Linux charged money then they would have failed also. It was the free, open-source concept that has allowed Linux to survive! Yet being fully free is still not enough for Linux to gain a large desktop market share. That must say something? Only two PAY-FOR OSes dominate x86, Windows & MacOS X because no other OS can compete with them for market share.
There were real applications on Windows playing multiple videos smoothly in that period.
Maybe on higher end hardware but not on any similar system that BeOS was running. I recall running older systems that had trouble with videos on Windows, etc. that played well on BeOS (Pentium2, 256 or 512MB RAM).
4 way multiprocessing was an exceptionally high-end workstation capability in BeOS’s day, but right now it’s something that comes with mid-range PCs, and the workstations are offering 8-16 way SMP.
I thought Haiku was aiming for the desktop. I’m certainly not running 16 core system. Maybe you are? Haiku can handle total of 8 cores which today is very good because most systems are 2, 4 or 6 cores. I doubt this will change anytime soon for desktops because we don’t need all that computing power and every core adds to power consumption.
[quote=tonestone57]“What BeOS did” was fail in the market and lose millions of dollars of other people’s money.
BeOS was selling their OS and having trouble to get buyers. Linux only succeeded in getting their 1-2% market share by giving their OS away for free and after many years of pushing it. Had Linux charged money then they would have failed also. It was the free, open-source concept that has allowed Linux to survive! Yet being fully free is still not enough for Linux to gain a large desktop market share. That must say something? Only two PAY-FOR OSes dominate x86, Windows & MacOS X because no other OS can compete with them for market share.
There were real applications on Windows playing multiple videos smoothly in that period.
Maybe on higher end hardware but not on any similar system that BeOS was running. I recall running older systems that had trouble with videos on Windows, etc. that played well on BeOS (Pentium2, 256 or 512MB RAM).
4 way multiprocessing was an exceptionally high-end workstation capability in BeOS’s day, but right now it’s something that comes with mid-range PCs, and the workstations are offering 8-16 way SMP.
I thought Haiku was aiming for the desktop. I’m certainly not running 16 core system. Maybe you are? Haiku can handle total of 8 cores which today is very good because most systems are 2, 4 or 6 cores. I doubt this will change anytime soon for desktops because we don’t need all that computing power and every core adds to power consumption.[/quote]
AMD is shipping bulldozer 8 core client parts in april or may. Q2 of 2011. Those will be 8 core parts. They are 32nm which should mean asub 3mm dies size which should mean somewhere in the neighborhood of 100w or less.
Though they will be high and midrange parts. They will see fiarly wide market adoption and quickly as the represent or should represent a serious bargin in terms of computer power. I am thinking $300 cpu at intorduction. I know I am waiting on reviews but I am fairly certian I will be upgrading my current rig to Bulldozer AM3+ socket and ddr3. From my current thuban phenomII 6 core ddr2 buuild.
I still think most people (largest market share) will be using 2, 4 or 6 core systems for some years more. Only higher end desktops will use 8 cores and have few users to start with. Anyways, Haiku can handle up to 8 virtual cores (8 physical cores without HyperThreading or 4 cores with HT ).
For SMP I was talking about efficiency not that other OSes can’t do it. You run benchmark with 1 core on and then turn 2nd core on and compare the results. Compute the percentage increase. It can be run for 1, 2, 4, etc. cores and do analysis on the results to see how well multithreading is handled on multiple cores.
Early versions of BeOS were free with the heavily subsidised hardware from Be, and the software, a work in progress, was needed to get it to do anything at all. At this time Be thought it might be a computer company like Apple, not a software company. After a few years that idea didn’t work out.
Before Be Inc. finally went out of business, you might remember there was speculation that they’d be bought by Red Hat. What does Red Hat do? Well, it sells Linux subscriptions. Takes hundreds of millions of dollars a year (if 2011 is a good year they’ll break the billion dollar mark) that way.
BeOS R5 (the version almost all the people who call themselves BeOS fans actually started with) was free as-in beer, Be was desperate to put it onto PCs, calling around magazine companies to get it mounted on cover CDs, uploading it to the same free FTP archive sites where you’d get Slackware or Debian. The pay for “Pro” version was almost identical except for the widely-ignored prohibition on commercial use, in reality (though it would be a while before they came out and said it) BeOS as a commercial proposition was dead by the time BeOS R5 was delivered.
[quote]
Had Linux charged money then they would have failed also. It was the free, open-source concept that has allowed Linux to survive! Yet being fully free is still not enough for Linux to gain a large desktop market share. That must say something? Only two PAY-FOR OSes dominate x86, Windows & MacOS X because no other OS can compete with them for market share.[/quote]
Linux people charged money almost from the outset. Still do. Red Hat Linux dates back to before Linux 1.0, and way before BeOS DR 1.
It’s about the applications. Windows ISVs, though the bane of Microsoft’s existence in other ways, ensure Windows is going to cover all the bases. On OS X it’s a bit trickier, and so on. Haiku does very badly on this basis.
[quote]
For SMP I was talking about efficiency not that other OSes can’t do it. You run benchmark with 1 core on and then turn 2nd core on and compare the results. Compute the percentage increase. It can be run for 1, 2, 4, etc. cores and do analysis on the results to see how well multithreading is handled on multiple cores.[/quote]
Even though this is a rather crude test, I encourage you to try it and compare against other systems.
[quote=tonestone57]I still think most people (largest market share) will be using 2, 4 or 6 core systems for some years more. Only higher end desktops will use 8 cores and have few users to start with. Anyways, Haiku can handle up to 8 virtual cores (8 physical cores without HyperThreading or 4 cores with HT ).
For SMP I was talking about efficiency not that other OSes can’t do it. You run benchmark with 1 core on and then turn 2nd core on and compare the results. Compute the percentage increase. It can be run for 1, 2, 4, etc. cores and do analysis on the results to see how well multithreading is handled on multiple cores.[/quote]
I am looking forward to 8 cores with haiku, the performance difference can be handled with properly written applications however. You do thing serialy when the overhead for multithreading outwieghs the advantage of using it.
Its really very application specific. Overall I have tested haiku throughput, its actually pretty good on average. Could be better but some of the differences have to do with instructions used and the fact that 64b instructions will trump 32b instructions in a number of tasks.
So its all gonna be ok down the road.I’d really like to see a GPGPU model built into teh kernel. Harnessin’g that for haiku " better then other systems in thoery" could really give haiku alot of punching power.
BTW I expect to see a big growth in 6 and 8 core systems mid 2011, its going to be a slow roll uphill but if haiku does 6-8 core better then linux and windows, it might have a chance. But thats depdant on R1.
How much longer can it wiat is the question as the competition ups the game more and more by the day.
[quote=NoHaikuForMe]
Early versions of BeOS were free with the heavily subsidised hardware from Be, and the software, a work in progress, was needed to get it to do anything at all.[/quote]
So you’re comparing to an early version when no users cared about the OS. I wonder how many users Linux had with version 1.0. That’s the comparison you’re giving here. You should never expect a software company to get traction early on.
There was a stripped down BeOS 5 PE edition but that was to attract users to the OS and get them to buy the PRO version. You can’t get users to a new OS without letting them try it out for free to build the user-base first.
Be Inc was a software company that required money to exist. Linux is lots of open-source developers working for free in their spare time and lots of companies (not just one) that hire developers to code for Linux because they use it for their servers.
Um, you’re using a very small # of distros to represent the norm. I’ve downloaded Ubuntu, Debian, Redhat, Fedora, PCLinuxOS, CentOS, VidaLinux, and other distros and never been charged for them. You’re trying to use Enterprise Redhat to say Linux is charged for. Well, desktop Linux versions I’ve used have never charged for them. I believe all the desktop distros are free (ie: can download an ISO for free). Redhat was free too and for desktop until 9.0 and after that switched over to commercial. I should know because I used it.
Linux has tons of applications yet does very badly here too. They can’t convince people to switch over. Linux gives people a free OS + many free applications and yet people resist and still stay with Windows & OS X. So, BeOS failed but Linux isn’t doing any better on the desktop from what I see.
That is actually not crude. A multithreaded application should be able to use multiple cores. You can tell if multithreaded if all CPUs get maximized during operation & # of threads. Then you would use some performance measure like fps or time to completion running under 1 core and multiple cores. Haiku allows turning cores on/off very easy. (ie: using multithreaded program like Handbrake for testing is one example) Of course Haiku is still Alpha & better to wait for R1 to run a test like this for fairness.