Actually there are plenty of innovative open source platforms out there, look at syllable, plan9 (which is open source btw), and quite a few more... There have also been innovative closed source platforms like beos. Now what do all of these have in common? The open source ones are small niche projects, while the closed source ones tend to die. There simply isn't demand for a new OS that's not compatible with anything else already on the market. Because of this, innovation in certain areas can only really happen in open source, commercial businesses won't invest time/money in creating a new innovative OS because the likelihood of profit is so low.
And sure, Linux is based on a 30 year old design, but was that design really so bad to begin with? It seems like the basic unix design has held up very well, with so many different implementations still keeping to the basic ideas. It's a very different beast to windows, where historic design flaws are still lurking and often have forced the develop[ers to come up with new incompatible ways of doing things using the old and new apis (look at win16, the way windows encrypts passwords twice with 2 different hashes, the way theres 2 or more sets of winsock libs etc).. while leaving the old apis there too, causing unnecessary bloat.
As for the slow down discussed earlier, hardware hasn't slowed down because it has competition... Software on the other hand, has slowed to a crawl largely due to microsoft's stranglehold of the market. As i said earlier, any new innovative OS doesn't really stand a chance, and a lot of this is due to microsoft. And just look at the way IE6 stagnated from 2001 to 2006 or so when they started getting competition from firefox... Never before has the advancement of the web stagnated for so long. Even when netscape was on top, they still kept coming out with new stuff, they didn't sit and stagnate for 5 years.
Windows mobile is also a completely different beast to desktop windows... Linux on the N810 is however normal linux, and pretty much anything you can compile on desktop linux can be recompiled to run on the N810. The fact that most apps come with source code just makes the process simpler.
LVDI yes... By 'droid' i assume you mean the multilink adapter which converts DVI or VGA to LVDI, and yes i have one of those. I still have the Number 9 Revolution IV videocard with the LVDI connector, tho it sits in a box somewhere. When i first got this screen i used it with an SGI O2. It's currently connected to an AMD64 based system with an Nvidia videocard.
They can sell you a server for pennies on the dollar, or you can buy software licenses that are generated for fractions of pennies. The margins on software are much higher than even the most expensive appliances using the cheapest low grade hardware.
But that's why I mentioned using an external standard file server, and storing the data there in standard formats. When the appliance is EOL, you can buy an update or switch it out for something else. I want the choice, and i have no problem using a proprietary app running on a proprietary server so long as the data it uses/creates (ie my data) is in a standard format that i will always have access to.
They ought to make a version of google apps that comes on an appliance, and can connect to external file servers... A lot of companies are worried about their data being stored on google's servers and the issue of the service disappearing as you describe.
If you have an appliance storing its data on seperate file servers, you can still use your local apps too...
Why? Add a bunch of ugly holes, that get full of dirt/dust, to support a selection of media card formats that will soon become obsolete. People typically keep their macs for considerably longer than an equivalent PC... Media card formats seem to be changing much more rapidly. There are simply too many media card formats, supporting all of them is ridiculous. Also if you only support one format, you're pushing your customers towards that format. And alienate anyone who needs other card formats.
Would you buy a mac if your bank used a non standard apple technology that didn't work with anything else? Or would you be angry at the bank trying to force your choice of technology, and complain and/or switch banks?
I'm sure most people would rather buy a machine and have it be able to access the internet securely out of the box and take full advantage of all the performance they paid for.
Nowhere in the ads for the machine did it say:
* Knowledge of how to obtain, install and maintain third party security applications required. ** Due to background security software, actual system will be slower than advertised.
No matter how "easy" it is, it's still work that end users don't want to do, and often don't even know how. And you say all the software can be obtained for free, how do you expect end users to know what software they need, and where to get it from? And how are clueless end users supposed to differentiate the legit free programs they're looking for to secure their machine, and the legit free backdoors that websites are telling them will keep their machines secure. It's not a huge step from knowing they need all this third party software, knowing where it can be obtained, knowing how to install it and keep it up to date, and knowing how to download and install ubuntu.
But your right, people shouldn't need to do this. Having all this extra third party crap installed hinders performance unnecessarily.
Linux and OSX are better for end users, why? Because they work out of the box, and don't require end users to know about installing and keeping up to date antivirus, anti spyware, defragmenting tools, registry cleaners, software firewalls etc...
6-bit? Hmm, how about the macbook and macbook pro systems? I have an SGI 1600SW which is roughly the same size/res as the screen on a 17" macbook pro (the 1680 one, not the new 1920 one) and the sgi screen seems to have richer colors...
I always used ulimits on multiuser systems, but never really bothered for my own single-user workstations. We used to have a few servers that handled a stack of old sun dumb x terminals, netscape (4.x) would often go haywire and try to consume all the ram, especially when people turned off the X terminal while it was running (netscape didn't detect the death of the xserver and shut down gracefully, it went nuts trying to allocate all available ram).
// Besides a 32bit kernel offers backwards compatibility with existing drivers.
What existing drivers? When the first x86 macs came out, there were no existing x86 osx drivers for it to be compatible with. A 32bit x86 kernel will not offer compatibility with 32bit PPC drivers. Thus, drivers will already need to be ported from PPC to x86, taking into account any endian issues that might occur. And any drivers written for PPC64 will already be 64bit clean anyway.
Mobile AMD chips (that is the Turions) may have a slightly higher TDP, but your comparing only the 2 chips, and not the motherboard chipset. Intel use a seperate memory controller on the chipset, AMD have it built in to the CPU so merely the presence of that would account for some of the difference. And tests of actual Turion systems vs Pentium M systems shows them fairly competitive from a power usage perspective (i couldnt find newer reviews for the respective dual core variants), see: http://techreport.com/reviews/2006q1/pentiumm-vs-turion64/index.x?pg=12
While i agree that the netburst pentiums were horrendous marketting-driven chips, the fact remains that they still available.
As for 64bit performance... Having used Windows and Linux in both 32 and 64bit incarnations, i have found the 64bit versions to be faster in virtually everything, especially when you have lots of memory (PAE causes quite a performance hit on 32bit). The only downside is the lack of support for older hardware in 64bit windows (x86_64 osx would be in the same boat as x86 osx here), a problem linux doesn't really suffer from since most drivers can be recompiled.
And yes, the memory consumption for 64bit is slightly higher, but not massively so. Pointers are twice the size, but most other data remains the same. Memory is ridiculously cheap these days, and adding memory over 3gb (increasingly common because it costs what, $100 these days) to a 64bit system doesn't require performance-sapping segmentation hacks like PAE...
The benefit is more than marginal and in a handful of apps, however most of the performance improvement is due to the extra registers rather than just being 64bit.
But it's not about performance, it's about only doing one architecture transition instead of two, and future proofing. Moving to an all but dead architecture for 6 months and then moving again seems wholly pointless. And i was only talking about the kernel, userland could still be a mix of 32/64, and not needing to use kludges like PAE to access more than 4GB of address space (translating to 3gb or so of ram) would increase performance. As it stands, new versions of OSX for quite some time will use a 32bit kernel just so they can be compatible with drivers written for a series of machines that only lasted a few months.
And no, Intel had no 64bit core chips, but they did have 64bit P4 based chips (yes they're power hungry, slow and get hot)... And there's always AMD who had a good line of 64bit x86 chips.
Having just tried, i found that in openoffice... Backspace deletes the contents of a cell without prompting you (you can still undo) Delete brings up a dialog allowing you to delete not just the contents but also any formatting, or to choose exactly what to remove... You can make it just remove formatting, or numbers, or text, or formulae etc... Very useful to strip numbers from a large block of cells without affecting formatting or textual content.
As for web queries i don't know, tho i know quite a few people who do the same as you with various scripts backing up to more serious databases.
I agree with you, I too object to being forced to choose software for my personal computer, not just by the government but by any third party.
This is why i support standard formats such as ODF, as a standard format provides the user the greatest freedom to choose their software.
While i strongly object to the government saying "Here is a binary file, we don't know the structure of the format but to read it you must buy a program for $400 (plus the os and hardware it requires if you dont have them already) to read it"... I object much less to "Here is a file, the format specifications are available from http://blah/ and you can choose from one of several applications at different price points (including free) and for different platforms. If your lazy, we can provide you with an app for free"
The push for ODF comes from individuals who value the freedom which you claim to.
As for "living with multiple formats", you mean living with interoperability problems due to multiple formats. When Commodore released the Amiga they realised what a hassle all the various formats were, and create a standard set that were documented and supported by libraries in the Amiga devkit. This format was called IFF, and stood for Interchangeable File Format, and supported sound, music, text, bitmap graphics etc. Because of this, people never had any issues opening files created by other amiga users, regardless of what applications they used. Some apps supported other formats as well, but virtually everything used the appropriate IFF format as it's default.
For what it's worth, i can't place such a huge block of numbers to slashdoc (lameness filter will block it) but 2^8388608 is very large indeed, the result is posted at: http://www.ev4.org/2-powerof-8388608.txt
Strange, Java web start worked just fine on 64bit Solaris. Agreed for flash tho, flash is the biggest thorn in the side of 64bit linux users. Virtually everything else can be recompiled for 64bit these days.
It's not an artificial restriction.. They simply only include drivers for the hardware that they ship. Perhaps some enterprising asian hardware manufacturer will start producing systems using the same components as Apple, and ship it with EFI firmware.
I assume your talking about the first core2 based macbooks... Apple didn't introduce hardware limitations, they used an Intel chipset that had such limitations... They officially quote these machines as supporting 3GB, you can put 4GB in them and it will be detected but not used. Other manufacturers who ship the same chipset quote it as capable of supporting 4GB, but in actual fact you can't use more than 3 on those systems either, even if your OS supports PAE or is 64bit because the limit is the motherboard chipset, not the CPU.
Well, what incentive do these OEMs have to keep memory prices high? They buy memory, surely they want to buy it as cheaply as possible. It looks more like these manufacturers are taking advantage of the low prices to boost the perceived value of their systems, and also offset some of the slowness associated with vista.
MS have to play at being an expensive premium option, they can't play the cheap option anymore, as they will never be able to undercut linux. It is also being the cheap option (not so much the software, but the hardware it ran on) that made ms.
Some "drivers" are actually user land, and will work on intel even tho the were intended for PPC... This is often the case with printer drivers (which are typically little more than a PPD file) and thus the guy claiming to have used ppc drivers on his Intel mac may be telling the truth, but simply unaware of the actual facts. When microsoft made an alpha (as in the cpu) version of windows, you could use x86 printer drivers on them under emulation.
Secondly, leopard's use of a 32bit kernel on intel macs is a bug-bear for me... There was only a very short lived series of 32bit intel macs, which lasted what? less than a year? So now they are limited to compatibility with such a short lived machine, and a future transition to 64bit. They should have used the architecture switch as an opportunity to switch to pure 64bit at the same time. Compatibility wouldn't have been any more of a problem than it already was and it would have set them up for a less bumpy future.
While your right that 32bit drivers won't work on a 64bit OS... The vast majority of drivers for linux are distributed as source code which can safely be compiled as either 64 or 32. Linux has had 64bit capability for a lot longer than windows, and thus plenty more time for non 64bit clean code to be weeded out. Because of linux's drivers i am able to use all kinds of pci cards in alpha/ultrasparc hardware that would normally not be supported on these platforms, simply by compiling the appropriate linux driver.
Your example about nvidia drivers is one of the biggest reasons why people don't want closed source drivers like nvidia's.
Slashdot Mirror
Actually there are plenty of innovative open source platforms out there, look at syllable, plan9 (which is open source btw), and quite a few more... There have also been innovative closed source platforms like beos. Now what do all of these have in common?
The open source ones are small niche projects, while the closed source ones tend to die. There simply isn't demand for a new OS that's not compatible with anything else already on the market. Because of this, innovation in certain areas can only really happen in open source, commercial businesses won't invest time/money in creating a new innovative OS because the likelihood of profit is so low.
And sure, Linux is based on a 30 year old design, but was that design really so bad to begin with? It seems like the basic unix design has held up very well, with so many different implementations still keeping to the basic ideas. It's a very different beast to windows, where historic design flaws are still lurking and often have forced the develop[ers to come up with new incompatible ways of doing things using the old and new apis (look at win16, the way windows encrypts passwords twice with 2 different hashes, the way theres 2 or more sets of winsock libs etc).. while leaving the old apis there too, causing unnecessary bloat.
As for the slow down discussed earlier, hardware hasn't slowed down because it has competition... Software on the other hand, has slowed to a crawl largely due to microsoft's stranglehold of the market. As i said earlier, any new innovative OS doesn't really stand a chance, and a lot of this is due to microsoft. And just look at the way IE6 stagnated from 2001 to 2006 or so when they started getting competition from firefox... Never before has the advancement of the web stagnated for so long. Even when netscape was on top, they still kept coming out with new stuff, they didn't sit and stagnate for 5 years.
Windows mobile is also a completely different beast to desktop windows...
Linux on the N810 is however normal linux, and pretty much anything you can compile on desktop linux can be recompiled to run on the N810.
The fact that most apps come with source code just makes the process simpler.
LVDI yes...
By 'droid' i assume you mean the multilink adapter which converts DVI or VGA to LVDI, and yes i have one of those.
I still have the Number 9 Revolution IV videocard with the LVDI connector, tho it sits in a box somewhere. When i first got this screen i used it with an SGI O2.
It's currently connected to an AMD64 based system with an Nvidia videocard.
They can sell you a server for pennies on the dollar, or you can buy software licenses that are generated for fractions of pennies. The margins on software are much higher than even the most expensive appliances using the cheapest low grade hardware.
But that's why I mentioned using an external standard file server, and storing the data there in standard formats. When the appliance is EOL, you can buy an update or switch it out for something else. I want the choice, and i have no problem using a proprietary app running on a proprietary server so long as the data it uses/creates (ie my data) is in a standard format that i will always have access to.
They ought to make a version of google apps that comes on an appliance, and can connect to external file servers... A lot of companies are worried about their data being stored on google's servers and the issue of the service disappearing as you describe.
If you have an appliance storing its data on seperate file servers, you can still use your local apps too...
Why? Add a bunch of ugly holes, that get full of dirt/dust, to support a selection of media card formats that will soon become obsolete.
People typically keep their macs for considerably longer than an equivalent PC... Media card formats seem to be changing much more rapidly.
There are simply too many media card formats, supporting all of them is ridiculous.
Also if you only support one format, you're pushing your customers towards that format. And alienate anyone who needs other card formats.
Would you buy a mac if your bank used a non standard apple technology that didn't work with anything else? Or would you be angry at the bank trying to force your choice of technology, and complain and/or switch banks?
And considerably slower than it should be...
I'm sure most people would rather buy a machine and have it be able to access the internet securely out of the box and take full advantage of all the performance they paid for.
Nowhere in the ads for the machine did it say:
* Knowledge of how to obtain, install and maintain third party security applications required.
** Due to background security software, actual system will be slower than advertised.
No matter how "easy" it is, it's still work that end users don't want to do, and often don't even know how.
And you say all the software can be obtained for free, how do you expect end users to know what software they need, and where to get it from? And how are clueless end users supposed to differentiate the legit free programs they're looking for to secure their machine, and the legit free backdoors that websites are telling them will keep their machines secure.
It's not a huge step from knowing they need all this third party software, knowing where it can be obtained, knowing how to install it and keep it up to date, and knowing how to download and install ubuntu.
But your right, people shouldn't need to do this. Having all this extra third party crap installed hinders performance unnecessarily.
Linux and OSX are better for end users, why? Because they work out of the box, and don't require end users to know about installing and keeping up to date antivirus, anti spyware, defragmenting tools, registry cleaners, software firewalls etc...
6-bit? Hmm, how about the macbook and macbook pro systems?
I have an SGI 1600SW which is roughly the same size/res as the screen on a 17" macbook pro (the 1680 one, not the new 1920 one) and the sgi screen seems to have richer colors...
I agree, i still have an SGI 1600SW that i've had for several years, and it's been connected to several different computers in it's time.
I always used ulimits on multiuser systems, but never really bothered for my own single-user workstations. We used to have a few servers that handled a stack of old sun dumb x terminals, netscape (4.x) would often go haywire and try to consume all the ram, especially when people turned off the X terminal while it was running (netscape didn't detect the death of the xserver and shut down gracefully, it went nuts trying to allocate all available ram).
// Besides a 32bit kernel offers backwards compatibility with existing drivers.
What existing drivers? When the first x86 macs came out, there were no existing x86 osx drivers for it to be compatible with. A 32bit x86 kernel will not offer compatibility with 32bit PPC drivers. Thus, drivers will already need to be ported from PPC to x86, taking into account any endian issues that might occur. And any drivers written for PPC64 will already be 64bit clean anyway.
Mobile AMD chips (that is the Turions) may have a slightly higher TDP, but your comparing only the 2 chips, and not the motherboard chipset. Intel use a seperate memory controller on the chipset, AMD have it built in to the CPU so merely the presence of that would account for some of the difference. And tests of actual Turion systems vs Pentium M systems shows them fairly competitive from a power usage perspective (i couldnt find newer reviews for the respective dual core variants), see:
http://techreport.com/reviews/2006q1/pentiumm-vs-turion64/index.x?pg=12
While i agree that the netburst pentiums were horrendous marketting-driven chips, the fact remains that they still available.
As for 64bit performance... Having used Windows and Linux in both 32 and 64bit incarnations, i have found the 64bit versions to be faster in virtually everything, especially when you have lots of memory (PAE causes quite a performance hit on 32bit). The only downside is the lack of support for older hardware in 64bit windows (x86_64 osx would be in the same boat as x86 osx here), a problem linux doesn't really suffer from since most drivers can be recompiled.
And yes, the memory consumption for 64bit is slightly higher, but not massively so. Pointers are twice the size, but most other data remains the same. Memory is ridiculously cheap these days, and adding memory over 3gb (increasingly common because it costs what, $100 these days) to a 64bit system doesn't require performance-sapping segmentation hacks like PAE...
The benefit is more than marginal and in a handful of apps, however most of the performance improvement is due to the extra registers rather than just being 64bit.
But it's not about performance, it's about only doing one architecture transition instead of two, and future proofing. Moving to an all but dead architecture for 6 months and then moving again seems wholly pointless.
And i was only talking about the kernel, userland could still be a mix of 32/64, and not needing to use kludges like PAE to access more than 4GB of address space (translating to 3gb or so of ram) would increase performance. As it stands, new versions of OSX for quite some time will use a 32bit kernel just so they can be compatible with drivers written for a series of machines that only lasted a few months.
And no, Intel had no 64bit core chips, but they did have 64bit P4 based chips (yes they're power hungry, slow and get hot)... And there's always AMD who had a good line of 64bit x86 chips.
Having just tried, i found that in openoffice...
Backspace deletes the contents of a cell without prompting you (you can still undo)
Delete brings up a dialog allowing you to delete not just the contents but also any formatting, or to choose exactly what to remove... You can make it just remove formatting, or numbers, or text, or formulae etc... Very useful to strip numbers from a large block of cells without affecting formatting or textual content.
As for web queries i don't know, tho i know quite a few people who do the same as you with various scripts backing up to more serious databases.
So what your saying is that the process while not perfect, is better than the alternative offered by microsoft...
I agree with you, I too object to being forced to choose software for my personal computer, not just by the government but by any third party.
This is why i support standard formats such as ODF, as a standard format provides the user the greatest freedom to choose their software.
While i strongly object to the government saying "Here is a binary file, we don't know the structure of the format but to read it you must buy a program for $400 (plus the os and hardware it requires if you dont have them already) to read it"... I object much less to "Here is a file, the format specifications are available from http://blah/ and you can choose from one of several applications at different price points (including free) and for different platforms. If your lazy, we can provide you with an app for free"
The push for ODF comes from individuals who value the freedom which you claim to.
As for "living with multiple formats", you mean living with interoperability problems due to multiple formats. When Commodore released the Amiga they realised what a hassle all the various formats were, and create a standard set that were documented and supported by libraries in the Amiga devkit. This format was called IFF, and stood for Interchangeable File Format, and supported sound, music, text, bitmap graphics etc. Because of this, people never had any issues opening files created by other amiga users, regardless of what applications they used. Some apps supported other formats as well, but virtually everything used the appropriate IFF format as it's default.
For what it's worth, i can't place such a huge block of numbers to slashdoc (lameness filter will block it) but 2^8388608 is very large indeed, the result is posted at:
http://www.ev4.org/2-powerof-8388608.txt
And is certainly much higher than 524TB!
Strange, Java web start worked just fine on 64bit Solaris.
Agreed for flash tho, flash is the biggest thorn in the side of 64bit linux users. Virtually everything else can be recompiled for 64bit these days.
It's not an artificial restriction.. They simply only include drivers for the hardware that they ship.
Perhaps some enterprising asian hardware manufacturer will start producing systems using the same components as Apple, and ship it with EFI firmware.
I assume your talking about the first core2 based macbooks...
Apple didn't introduce hardware limitations, they used an Intel chipset that had such limitations... They officially quote these machines as supporting 3GB, you can put 4GB in them and it will be detected but not used. Other manufacturers who ship the same chipset quote it as capable of supporting 4GB, but in actual fact you can't use more than 3 on those systems either, even if your OS supports PAE or is 64bit because the limit is the motherboard chipset, not the CPU.
Well, what incentive do these OEMs have to keep memory prices high? They buy memory, surely they want to buy it as cheaply as possible.
It looks more like these manufacturers are taking advantage of the low prices to boost the perceived value of their systems, and also offset some of the slowness associated with vista.
MS have to play at being an expensive premium option, they can't play the cheap option anymore, as they will never be able to undercut linux.
It is also being the cheap option (not so much the software, but the hardware it ran on) that made ms.
Some "drivers" are actually user land, and will work on intel even tho the were intended for PPC... This is often the case with printer drivers (which are typically little more than a PPD file) and thus the guy claiming to have used ppc drivers on his Intel mac may be telling the truth, but simply unaware of the actual facts. When microsoft made an alpha (as in the cpu) version of windows, you could use x86 printer drivers on them under emulation.
Secondly, leopard's use of a 32bit kernel on intel macs is a bug-bear for me... There was only a very short lived series of 32bit intel macs, which lasted what? less than a year? So now they are limited to compatibility with such a short lived machine, and a future transition to 64bit. They should have used the architecture switch as an opportunity to switch to pure 64bit at the same time. Compatibility wouldn't have been any more of a problem than it already was and it would have set them up for a less bumpy future.
While your right that 32bit drivers won't work on a 64bit OS... The vast majority of drivers for linux are distributed as source code which can safely be compiled as either 64 or 32. Linux has had 64bit capability for a lot longer than windows, and thus plenty more time for non 64bit clean code to be weeded out.
Because of linux's drivers i am able to use all kinds of pci cards in alpha/ultrasparc hardware that would normally not be supported on these platforms, simply by compiling the appropriate linux driver.
Your example about nvidia drivers is one of the biggest reasons why people don't want closed source drivers like nvidia's.