No need to worry folks, us handful of BeOS users will switch off the lights and the internet on our way out, since we'll be the last ones to leave. Every now and then I'm actually relieved to be running a non mainstream OS.
Dont forget that BeInc threw in the towel after R4.5, released in July 1999. The focus shift was in September 1999, after which a majority of BeInc engineers left for greener pastures (Apple, Google and Microsoft). R5, which was released a few months after that, was a farewell present to the loyal fanbase.
Now, in September 1999, what was the state of the competition. Windows Me still wasn't released, and neither was Windows 2000. Mac OS was still version 8.5. BeOS 4.5 at this stage was the most advanced consumer desktop OS in existance, ready to rock on your dual CPU box. Lack of apps in 1999? It had hardware accelerated Quake 2, Cinema4D, Opera 3, SimCity was being worted, Neundo was being ported, and Gobe Productive was kick ass resposive compared to the competition. Hell, my Canopus DV raptor had codecs running under BeOS.
It is now almost 8 years since the focus shift. Haiku is on the horizon, but alas, its just too little too late.
The algorithm is based on the frequency of letters in the English language, and assigns a bit pattern (from smallest to largest) based on the order these letters appear in the main text. For odd letter combinations which never appear together (eg. qx, tz, vz etc), you will actually get no compression, but for most dictionary words, you could compress common words into a few bits (eg. compress and into 10).
The particular compression lookup-table is useless in non english language, but I guess another table should be used for other languages.
The challenge is actually related to how best to design a look-up table, and how to fill the table. This has absolutely nothing to do with AI.
A degree is nothing more than a piece of paper which certifies that you can get boring shit assigned to you done. In essence, this is all an employer cares about when hiring graduates. It does not guarantee a minimum level of knowledge or skills.
At the same time, education facilities are running a business. They want to maximise profit, which is where students come in. However, they are also competing against other education facilites, so they dont want to squeeze too hard, otherwise you will take your money elsewhere.
Having looked back at my 'academic' life, all I really needed to have is the minimum 2-3 year tertiary diploma / degree (which is called differently from country to country). This provides the above mentioned certificate (get boring shit done). After a year in the industry, degrees no longer matter, it's all based on experience and specialisation. Shit, I should know, I'm an electronic engineer by education, and 7 years later, I'm a software architect in a company with 120 software engineers. I've advanced faster in this company than people with masters degrees and excellent academic marks.
If you wish to work in academia, its a different story. But then again, if you specialise in a new field untouched by academia, guess who'll be knocking at your door once the 'education business' decides it needs celebrity names to entice a new generation of students.
do your job and don't kill people (unless it's a gun, which is meant to kill people anyway.)
In the military, the purpose of using guns is to stop an enemy combatant. Wounding is even better than killing, since now his buddies have to look after him, which means that there are less bullets being fired your way.
And every computer in existance still has a reset button, and at some point in time that reset button has been exercised.
You've obviously never used a Commodore machine (Amiga, C=64 etc). No reset buttons here. I actually had to build a reset button manually by drilling the plastic dase, installing a push button, soldering two leads to the motherboard (ground and reset signal). Thinking back to the good old C=64, I even had to upgrade the OS by installing TornadoDOS. Back then, installing ment yanking out the EPROM storing the OS and inserting a new EPROM. And updating the floppy interface from serial to parallel. (disk transfer rates went from 300Bps to 3kBps). The Operating system back in those days took a full 8Kb back then.
It's interesting how people cannot visualise the most efficient method to utilise multiple CPU's present in next gen consoles. You do not allocate one CPU for physics, one for sound, one for graphics and so on, since these are not independant parallel tasks. Only when the physics calculations are complete can you do the AI, and only then do you draw/play world objects (graphics / sound). This is relatively sequential, since for drawing all you're really doing is queueing commands to a graphics processor, which runs things at its own pace anyway.
The ideal method to utilise multiple CPU's is to create a schedular (on main CPU, or part of game OS) which waits for job requests, and then is allocates a free CPU to work on the job. Each job will run for a while, and then store the result and terminate. Therefore, each of the 7 CELL units in the PS3 will be running different jobs whenever its free (sometimes audio, sometimes video, sometimes AI). The beauty of this design allows you to have everything cranking on all cylinders) when a new frame is calculated (ie. generate new vertex lists, AI pathfinding etc). Once we've issued rendering commands, we have all the time in the world on the CELL to precalculate the next frame while the GPU is busy actually drawing one frame.
I'm really suprised by the comments here at Slashdot. Slashdot used to be the refugee of technology enthusiasts, who would flock to newer/interesting/alterantive technologies. This used to be the meeting place of people running various Operating Systems, from the exotic (Plan 9 / Elate) to the antique (Amiga...), and everything in between. The community used to be proud of not running a mainstream OS. Having read the article and the comments, I expected a "MEH" and "not a problem on my system" type of response. Instead, I see lots wannabe devotees, but in their hearts and their desktops, they run the Mainstream OS (*gasp*). Typical case of of do as i say, not as i do.
Well if the reader base of Slashdot alone used a non-mainstream OS, Google statistics would shold an increase in the number of alternative OS's, and this would in a round-about way remove this root-kit problem all together, since at the end of the day, there would be no financial incentive by Sony (or others) to finance a DRM and copy protection scheme which would only be effective for
Want to stop this nonsense from happening in the future? Actually run a non-mainstream OS. That shouldn't be hard for most of the visitors of this website, shouldn't it?
I'm really suprised by the comments here at Slashdot. Slashdot used to be the refugee of technology enthusiasts, who would flock to newer/interesting/alterantive technologies. This used to be the meeting place of people running various Operating Systems, from the exotic (Plan 9 / Elate) to the antique (Amiga...), and everything in between. The community used to be proud of not running a mainstream OS. Having read the article and the comments, I expected a "MEH" and "not a problem on my system" type of response. Instead, I see lots wannabe devotees, but in their hearts (and their desktops), they run the Mainstream OS (*gasp*). Tipicle case of of do as i say, not as i do.
Well if the reader base of Slashdot alone used a non-mainstream OS, Google statistics would shold an increase in the number of alternative OS's, and this would in a round-about way remove this root-kit problem all together, since at the end of the day, there would be no financial incentive by Sony (or others) to finance a DRM and copy protection scheme which would only be effective for
Want to stop this nonsense from happening in the future? Actually run a non-mainstream OS. That shouldn't be hard for most of the visitors of this website.
Re:Sad end to a Sad story - One developer's view
on
The End of PalmOS?
·
· Score: 1
Tsk tsk, the problem is not with the slimey executive, the problem is with the stupid board of directors who appoints the executive. And who puts these people up? The majority shareholders, thats who. Big financial institutions (banks, pension funds, insurance funds, etc). They put Carly there to make a quick buck. And make bucks they did. Who looses? Day traders, and mom and pop. Think of it as a form of Darwins Evolution theory, but for making money. A new sucker is born every minute.
Yes, and they didn't have to recompile the kernel, fiddle with conf files, resolve dependancy issues, and can basically play multiple sounds at the same time. All this while spinning teapots.
Opera 8.01 works fine. But what else would you expect from Scandinavians.
Re:For the Nth time, can someone explain the jargo
on
Zeta Goes Gold
·
· Score: 1
Frankly, I couldn't see anything in it that made it a "Media OS". It didn't run faster or handle video better than NT4 or Windows 2000 on the same hardware, and I had to add RAM to the box that had been running NT4 to make BeOS happy.
I've downloaded a lovely DivX from the net (bittorrent rules), and I find it almost impossible to seek using VideoLan/WindowsMediaPlayer/WMPClassic/NeroVision/ Quicktime under Windows. Using BeOS, every single native video playing app (ie. everything except VideoLan and MPlayer) could seek anywhere without any issues. Plus, the MediaKit in BeOS is one of the jewels of the OS. With Linux today I still have latency with audio and have trouble mixing the sound. With BeOS, it just works, and works exceptionally well. Media translators (part of the media kit codec) allow any app to play videos as long as they use the OS interface. Ever wonder why there are 1000 amateur video players for BeOS? It's dead easy to do. Now that is my definition of a Media OS
...Does that make it a microkernel?...
People exaggerate when they call the BeOS kernel a microkernel. It is smaller than the Windows NT and Linux kernels, but bigger than QNX and Mach. A large portion of the OS is outside of the kernel proper (like the filesystem, networking pre bone, the application server etc) and can be swapped and replaced on the fly. Other components require a reboot (video 2D drivers, BONE) when the componets are replaced. The components are very modular, so a lot of people are fooled into believing that all components can be replaced, hence it seems like a micro kernel. Heck, even Haiku (former OpenBeOS) are working in parallel on multiple components, this is only possible due to the well organised framework/components of the core OS.
So to answer your question, it strickly isn't a microkernel, but it's architecture and components are like one. Most kits use messages for interprocessor communication, but for performance reasons a pure microkernelish approach was not taken.
The OSXexpert has a hard on for bashing yellowTab, but I can understand his bitterness since he has been bitten by BeInc (rendering his investment of 2 PPC BeBoxs irrelevant). Now Mr OSXexpert, how does it feel that your lovely Dual PPC G5 PowerMac is also being rendered obsolete. Apple RDF (Reality Distortion Field) still clouding your judgement? Once you've stopped counting the 200 improvements in OSX 10.4 Tiger, let us know how well OSX handles spawning and switching between multiple threads, how fast it releases system resources, how quickly can it send messages from one thread to another etc? I've given away an iMac to my niece since the spinning beach ball was driving me up the wall. I've returned to x86 land with Windows, Zeta and a partition with Ubuntu (which only serves to refresh my faith in Haiku).
Re:I don't get it! Who's going to buy it?
on
Zeta Goes Gold
·
· Score: 4, Interesting
Some people like to support the underdog. I have purchased the Release Candidate of Zeta, knowing more than likely that I'd be throwing money away. I think of it more like a donation, to ensure that an alternative to Windows and Linux continues surviving. Bernd seems to be managing the company quite well, so it looks like yellowTab just might make it. Once they have the 70 employees Bernd has been talking about, expect a full on revival of BeOS (awaken from the dead:-).
Plus, Haiku is getting closer, so by supporting the successor to BeOS, I am indirectly supporting Haiku. By showing that there is money to be made with BeOS, developers are more than likely to start offering software for another viable OS.
Re:What's the compelling reason to switch?
on
Zeta Goes Gold
·
· Score: 3, Interesting
What can Zeta do on the average 2005 desktop machine that Windows or Linux can't? Everything I've seen in terms of software offerings (CD player, CD burner, video editor, AIM client, e-mail, Firefox, etc...) are things that already exist in Windows and Linux. What's the compelling reason to switch?
BeOS / Zeta has a certain feel which cannot be benchmarked or reviewed, it can only be experienced. It's one of those mythical quantaties which stirs a type of passion which is missing in both Windows and Linux. It's weird, kind of like falling in love - for everyone else, the lady in question is but another female, yet once you've tasted the forbidden fruit, you love every aspect of her - her smell, her smile, her hair, her skin...
BeOS feels smoother and more responsive than any other system we've tried (and we've tried a lot of systems). It's like a sports car which handles curves very well. It's useless for family people since the sports car has no back seat for the kids, its useless for the tradesmen since it has no space in the trunk, but man, for us sports car geeks, it gets the heart pumping...
Re:Hardware Support Lacking
on
Zeta Goes Gold
·
· Score: 1
Mainstream hardware is supported, while exotic cards are probably not. The situation is probably similar to Linux. They have NDIS for network drivers, CUPS for printers, they have active nVidia and ATI video drivers (hardware OpenGL only available on earlier nVidia cards, but that will probably change). All in all, it is no worse than what BeInc faced in 1999. I'd probably state that yellowTab has even better driver support than what BeOS R5 did back in its day mainly thanks to the work from the Haiku developers.
Re:Do they or do they not have the source legally?
on
Zeta Goes Gold
·
· Score: 3, Interesting
yellowTab were under NDA with Palm in regards to the source code - at one stage they were unable to modify any of the 5.1d generation code, but lately it seems that the restrictions have been lifted, since they've hired kernel engineers to address bugs in the original BeOS. They have fixed the memory limit issues (32Mb addon, 1Gb memory limit, few other issues), and they've added processor specific optimisations (memcpy, memset etc), so that it works on newer CPU's (like the AMD x86-64 in 32 bit mode, the hyperthreaded P4's and such).
They've obviously got the licence, since it would be suicide to have coorporate offices, 35 employees etc for a non-legal product. yellowTab are legit, they just were not able to publicly speak about the nature of the licence when they first announced Zeta. Now that the restrictions have expired, it would be foolish for them to say anything else (since it would cast doubt on a previous era).
I've been following the progress yellowTab have been making, ever since they licenced the rights to distribute BeOS R5PE and Pro (when they acquired the rights from Koch Media). yellowTab have some small mistakes as they were trying to learn how to stand on their own two feet, but lo-and-behold, they are now a company with 35 employees (and rising). Unlike BeInc, yellowTab know a thing or two about marketing, and are slowly generating enough revenue to employ 70 employees. They have a few of the old BeInc engineers who originally worked on the BeOS, and they have managed to hire / contract some of the Haiku (former OpenBeOS) developers to work on some of the Zeta components.
If yellowTab play their cards right, they will have enough finances to employ the targetted 70 engineers, and work on Zeta R2, which for all intentional purposes can be regarded as BeOS R7.
The more the merrier, I say, and I wish them luck.
Back in the army, I used to man an anti aircraft radar installation. The missile launch panel was protected by a series of locks / protocol codes for that day. The designers figured that this level of security (2 keys and a keypad) should be enough to protect the latch mechanism on the launch button, thus preventing soldiers from accidently launching missiles. Stick your fingers through a maintainance panel in the back, and you could manually unlock the latch protecting the launch buttons. So much for secure control panels.
Dont fool yourself, if someone maliscious wanted to bypass the security of "the button", they would. I'm 100% confident that there are workarounds for the regular launch process.
Slashdot Mirror
No need to worry folks, us handful of BeOS users will switch off the lights and the internet on our way out, since we'll be the last ones to leave. Every now and then I'm actually relieved to be running a non mainstream OS.
Dont forget that BeInc threw in the towel after R4.5, released in July 1999. The focus shift was in September 1999, after which a majority of BeInc engineers left for greener pastures (Apple, Google and Microsoft). R5, which was released a few months after that, was a farewell present to the loyal fanbase.
Now, in September 1999, what was the state of the competition. Windows Me still wasn't released, and neither was Windows 2000. Mac OS was still version 8.5. BeOS 4.5 at this stage was the most advanced consumer desktop OS in existance, ready to rock on your dual CPU box. Lack of apps in 1999? It had hardware accelerated Quake 2, Cinema4D, Opera 3, SimCity was being worted, Neundo was being ported, and Gobe Productive was kick ass resposive compared to the competition. Hell, my Canopus DV raptor had codecs running under BeOS.
It is now almost 8 years since the focus shift. Haiku is on the horizon, but alas, its just too little too late.
The particular compression lookup-table is useless in non english language, but I guess another table should be used for other languages.
The challenge is actually related to how best to design a look-up table, and how to fill the table. This has absolutely nothing to do with AI.
A degree is nothing more than a piece of paper which certifies that you can get boring shit assigned to you done. In essence, this is all an employer cares about when hiring graduates. It does not guarantee a minimum level of knowledge or skills.
At the same time, education facilities are running a business. They want to maximise profit, which is where students come in. However, they are also competing against other education facilites, so they dont want to squeeze too hard, otherwise you will take your money elsewhere.
Having looked back at my 'academic' life, all I really needed to have is the minimum 2-3 year tertiary diploma / degree (which is called differently from country to country). This provides the above mentioned certificate (get boring shit done). After a year in the industry, degrees no longer matter, it's all based on experience and specialisation. Shit, I should know, I'm an electronic engineer by education, and 7 years later, I'm a software architect in a company with 120 software engineers. I've advanced faster in this company than people with masters degrees and excellent academic marks.
If you wish to work in academia, its a different story. But then again, if you specialise in a new field untouched by academia, guess who'll be knocking at your door once the 'education business' decides it needs celebrity names to entice a new generation of students.
In the military, the purpose of using guns is to stop an enemy combatant. Wounding is even better than killing, since now his buddies have to look after him, which means that there are less bullets being fired your way.
Back in my day, teenagers were smarter...
It is a northbridge chipset limitation. 3Gb is the maximum you can use, even if you plug in 4 Gb.
You've obviously never used a Commodore machine (Amiga, C=64 etc). No reset buttons here. I actually had to build a reset button manually by drilling the plastic dase, installing a push button, soldering two leads to the motherboard (ground and reset signal). Thinking back to the good old C=64, I even had to upgrade the OS by installing TornadoDOS. Back then, installing ment yanking out the EPROM storing the OS and inserting a new EPROM. And updating the floppy interface from serial to parallel. (disk transfer rates went from 300Bps to 3kBps). The Operating system back in those days took a full 8Kb back then.
It's interesting how people cannot visualise the most efficient method to utilise multiple CPU's present in next gen consoles. You do not allocate one CPU for physics, one for sound, one for graphics and so on, since these are not independant parallel tasks. Only when the physics calculations are complete can you do the AI, and only then do you draw/play world objects (graphics / sound). This is relatively sequential, since for drawing all you're really doing is queueing commands to a graphics processor, which runs things at its own pace anyway.
The ideal method to utilise multiple CPU's is to create a schedular (on main CPU, or part of game OS) which waits for job requests, and then is allocates a free CPU to work on the job. Each job will run for a while, and then store the result and terminate. Therefore, each of the 7 CELL units in the PS3 will be running different jobs whenever its free (sometimes audio, sometimes video, sometimes AI). The beauty of this design allows you to have everything cranking on all cylinders) when a new frame is calculated (ie. generate new vertex lists, AI pathfinding etc). Once we've issued rendering commands, we have all the time in the world on the CELL to precalculate the next frame while the GPU is busy actually drawing one frame.
A typical example of a student who didn't attend a US College. I salute you.
Well if the reader base of Slashdot alone used a non-mainstream OS, Google statistics would shold an increase in the number of alternative OS's, and this would in a round-about way remove this root-kit problem all together, since at the end of the day, there would be no financial incentive by Sony (or others) to finance a DRM and copy protection scheme which would only be effective for
Want to stop this nonsense from happening in the future? Actually run a non-mainstream OS. That shouldn't be hard for most of the visitors of this website, shouldn't it?
Well if the reader base of Slashdot alone used a non-mainstream OS, Google statistics would shold an increase in the number of alternative OS's, and this would in a round-about way remove this root-kit problem all together, since at the end of the day, there would be no financial incentive by Sony (or others) to finance a DRM and copy protection scheme which would only be effective for
Want to stop this nonsense from happening in the future? Actually run a non-mainstream OS. That shouldn't be hard for most of the visitors of this website.
Fred fish, is that you?
Tsk tsk, the problem is not with the slimey executive, the problem is with the stupid board of directors who appoints the executive. And who puts these people up? The majority shareholders, thats who. Big financial institutions (banks, pension funds, insurance funds, etc). They put Carly there to make a quick buck. And make bucks they did. Who looses? Day traders, and mom and pop. Think of it as a form of Darwins Evolution theory, but for making money. A new sucker is born every minute.
Yes, and they didn't have to recompile the kernel, fiddle with conf files, resolve dependancy issues, and can basically play multiple sounds at the same time. All this while spinning teapots.
Opera 8.01 works fine. But what else would you expect from Scandinavians.
I've downloaded a lovely DivX from the net (bittorrent rules), and I find it almost impossible to seek using VideoLan/WindowsMediaPlayer/WMPClassic/NeroVision
People exaggerate when they call the BeOS kernel a microkernel. It is smaller than the Windows NT and Linux kernels, but bigger than QNX and Mach. A large portion of the OS is outside of the kernel proper (like the filesystem, networking pre bone, the application server etc) and can be swapped and replaced on the fly. Other components require a reboot (video 2D drivers, BONE) when the componets are replaced. The components are very modular, so a lot of people are fooled into believing that all components can be replaced, hence it seems like a micro kernel. Heck, even Haiku (former OpenBeOS) are working in parallel on multiple components, this is only possible due to the well organised framework/components of the core OS.
So to answer your question, it strickly isn't a microkernel, but it's architecture and components are like one. Most kits use messages for interprocessor communication, but for performance reasons a pure microkernelish approach was not taken.
The OSXexpert has a hard on for bashing yellowTab, but I can understand his bitterness since he has been bitten by BeInc (rendering his investment of 2 PPC BeBoxs irrelevant). Now Mr OSXexpert, how does it feel that your lovely Dual PPC G5 PowerMac is also being rendered obsolete. Apple RDF (Reality Distortion Field) still clouding your judgement? Once you've stopped counting the 200 improvements in OSX 10.4 Tiger, let us know how well OSX handles spawning and switching between multiple threads, how fast it releases system resources, how quickly can it send messages from one thread to another etc? I've given away an iMac to my niece since the spinning beach ball was driving me up the wall. I've returned to x86 land with Windows, Zeta and a partition with Ubuntu (which only serves to refresh my faith in Haiku).
Some people like to support the underdog. I have purchased the Release Candidate of Zeta, knowing more than likely that I'd be throwing money away. I think of it more like a donation, to ensure that an alternative to Windows and Linux continues surviving. Bernd seems to be managing the company quite well, so it looks like yellowTab just might make it. Once they have the 70 employees Bernd has been talking about, expect a full on revival of BeOS (awaken from the dead :-).
Plus, Haiku is getting closer, so by supporting the successor to BeOS, I am indirectly supporting Haiku. By showing that there is money to be made with BeOS, developers are more than likely to start offering software for another viable OS.
BeOS / Zeta has a certain feel which cannot be benchmarked or reviewed, it can only be experienced. It's one of those mythical quantaties which stirs a type of passion which is missing in both Windows and Linux. It's weird, kind of like falling in love - for everyone else, the lady in question is but another female, yet once you've tasted the forbidden fruit, you love every aspect of her - her smell, her smile, her hair, her skin...
BeOS feels smoother and more responsive than any other system we've tried (and we've tried a lot of systems). It's like a sports car which handles curves very well. It's useless for family people since the sports car has no back seat for the kids, its useless for the tradesmen since it has no space in the trunk, but man, for us sports car geeks, it gets the heart pumping...
Mainstream hardware is supported, while exotic cards are probably not. The situation is probably similar to Linux. They have NDIS for network drivers, CUPS for printers, they have active nVidia and ATI video drivers (hardware OpenGL only available on earlier nVidia cards, but that will probably change). All in all, it is no worse than what BeInc faced in 1999. I'd probably state that yellowTab has even better driver support than what BeOS R5 did back in its day mainly thanks to the work from the Haiku developers.
yellowTab were under NDA with Palm in regards to the source code - at one stage they were unable to modify any of the 5.1d generation code, but lately it seems that the restrictions have been lifted, since they've hired kernel engineers to address bugs in the original BeOS. They have fixed the memory limit issues (32Mb addon, 1Gb memory limit, few other issues), and they've added processor specific optimisations (memcpy, memset etc), so that it works on newer CPU's (like the AMD x86-64 in 32 bit mode, the hyperthreaded P4's and such).
They've obviously got the licence, since it would be suicide to have coorporate offices, 35 employees etc for a non-legal product. yellowTab are legit, they just were not able to publicly speak about the nature of the licence when they first announced Zeta. Now that the restrictions have expired, it would be foolish for them to say anything else (since it would cast doubt on a previous era).
I've been following the progress yellowTab have been making, ever since they licenced the rights to distribute BeOS R5PE and Pro (when they acquired the rights from Koch Media). yellowTab have some small mistakes as they were trying to learn how to stand on their own two feet, but lo-and-behold, they are now a company with 35 employees (and rising). Unlike BeInc, yellowTab know a thing or two about marketing, and are slowly generating enough revenue to employ 70 employees. They have a few of the old BeInc engineers who originally worked on the BeOS, and they have managed to hire / contract some of the Haiku (former OpenBeOS) developers to work on some of the Zeta components.
If yellowTab play their cards right, they will have enough finances to employ the targetted 70 engineers, and work on Zeta R2, which for all intentional purposes can be regarded as BeOS R7.
The more the merrier, I say, and I wish them luck.
http://www.haiku-os.org/
How about I send Mario with a baseball bat? He's got a batting average of 12 heads and 16 kneecaps. Will he do?
Back in the army, I used to man an anti aircraft radar installation. The missile launch panel was protected by a series of locks / protocol codes for that day. The designers figured that this level of security (2 keys and a keypad) should be enough to protect the latch mechanism on the launch button, thus preventing soldiers from accidently launching missiles. Stick your fingers through a maintainance panel in the back, and you could manually unlock the latch protecting the launch buttons. So much for secure control panels.
Dont fool yourself, if someone maliscious wanted to bypass the security of "the button", they would. I'm 100% confident that there are workarounds for the regular launch process.