OSX seems to do this too, it will say shutdown was cancelled due to some stupid app... You can override it using the cli based shutdown command tho, then it will just kill -9 any misbehaving app like any other unix would.
Well, 'business' users typically have very little use for audio, so they will typically have a cheap audio chip and very poor shielding on the internal cabling, and the lousiest pair of speakers you have ever heard... You might be better off with a consumer grade laptop, as the fact your eee sounds better seems to suggest.
Is it just me, or is the windows 7 taskbar much taller (ie consumes more vertical space) than previous versions? That, combined with thicker titlebars, doesn't make for very efficient use of vertical space on widescreen displays and especially on small netbook displays... The Ubuntu netbook interface seems far more suited to such devices, it has no bar at the bottom, and the menu bar at the top combines with the titlebar of any open window to use very little of the very limited vertical space on the screen.
People don't expect word to work flawlessly, it just doesn't and never has, they are willing to accept as normal that it's full of bugs and work around them... These same people are not willing to accept similar bugs from any other product, many of which occur due to trying to implement a poorly designed, reverse engineered format. They demand that any other product work flawlessly in order to replace something that works very poorly indeed.
99.9999% compatibility is a lot better than you get from MS, especially if you have a mix of different versions. Most businesses accept a much larger array of problems as simply being normal costs of doing business. The number of businesses i have been to where compatibility problems, kludgy workarounds, downtime due to crashes/reboots etc are considered a day to day part of business.
Sometimes the formatting issues occur even on the same version, depending on how your printer is configured for instance, a well known issue which doesn't occur often in most companies since they typically use the same printer. Installing security updates and service packs has also been known to modify the formatting...
What's hypocritical is that such problems are considered normal, and yet when openoffice exhibits similar issues (which are clearly unavoidable) it's cited as a reason why openoffice is unusable, but considering the format some formatting errors will be unavoidable... What version of word should openoffice try to emulate, or assume files are from?
No, but being patent encumbered does unless the patent holder declares the patent is free for anyone to implement under any terms they wish (ie they use the patent totally defensively and agree never to initiate any legal action against anyone over it).
The BSD license also has restrictions, just less of them... There is plenty of code released to the public domain with no restrictions whatsoever.
That said, the GPL places a few restrictions on you so that future users will be able to enjoy the same freedoms you did. BSD and public domain allow you to make derivative versions which are not free. Society works the same way (or is meant to), it has rules which restrict your freedom, but which are designed to ensure everyone receives a certain level of freedom and one person can't take that away from someone else.
Consider tho, that the iphone comes out of the box with gpl software on it (webkit at the very least)... And uses other gpl software during the development process (gcc etc)...
You are free to download this source, and also download the iphone sdk, compile this source and then execute it on the iphone simulator provided as part of the sdk... What costs $100 is the ability to test the code on your real iphone, or to submit your compiled binaries to the iphone store.
You could also try building this on an unofficial sdk for use with a jailbroken iphone...
You can't even settle things with a battle of the benchmarks: file system workloads vary so wildly that you can make a plausible argument for why any benchmark is either totally irrelevant or crucially important.
As pointed out, filesystem workloads vary massively, which is why it's good to have a choice of different filesystems which can be chosen based on individual requirements. Only offering a single filesystem like many other OS's do is extremely inefficient. One size does not fit all.
You can kill an X11 application too, using xkill, or terminal and a top... Both of these examples assume the GUI is functional, which it might not be (some programs, eg vmware, can hijack all input, and if such a program crashes you often find yourself unable to interact with the gui even if technically it hasn't crashed)...
On Linux the gui is not an inherent part of the system, and the entire gui system (X11) can be killed and restarted... Explorer is not the equivalent of X11, it's not even the equivalent of a window manager, it's closer to nautilus.
Despite Digital's best efforts, FX!32 still ran a lot slower than native Alpha code... Alpha was a lot more expensive than x86, and when running FX!32 not much faster.
With native Alpha code, the performance difference was more than enough to cover the price difference (a multi processor x86 box that, when running appropriately threaded code, could keep up with a single cpu Alpha cost a lot more).. Alpha boxes running Linux were really great for those of us who had them, they ran rings around people using x86... I still have 5 Alpha boxes in the attic ranging from a 166mhz EV4 to a dual 750mhz EV67.
The same thing happened to IA64, running native code IA64 was much faster than x86, but it was too expensive and too slow when emulating x86 code...
Expect the same thing to happen with ARM, although ARM processors will still be around because of the phone/pda market, their use in laptops and desktops is likely to only take off among Linux users where closed source apps are few and far between.
Alpha was so much faster that for a while at least, the emulated apps ran faster than anything intel could offer and native apps just made intel look like a joke.
But it was more expensive, and only ran apps designed for NT, while most people were still running win16 and dos apps which the alpha couldn't emulate.
Alpha boxes running linux were great tho, the only thing that couldn't be recompiled was netscape, and you could run the native alpha version of netscape designed for digital unix, which always seemed far more stable than other versions of netscape.
Linux also had the ability to emulate x86 apps on the alpha, but it was slower than the version windows used.. I never really used it beyond testing because i didn't have any apps that were only available for x86 linux.
OoO execution actually makes the cpu more efficient at executing less efficient code (or code which runs less efficiently by virtue of having been written for a different cpu)... If your willing to write/compile your code for a particular cpu then in-order execution works very well, thats why the ps3 and xbox 360 use in-order processors.
How bottom of the line? Do you have proper video drivers?
I was always able to play videos with mplayer which wouldn't play on windows running on the same hardware, going all the way back to 300mhz machines playing dvd video or early divx...
I wouldn't say 1/6 of the power, the p4 is slower clock for clock than the p3 was, so your 3ghz p4 will be comparable to a 2-2.5ghz p3, making it about 2.5 times the performance of a 1ghz cortex-a8... But compare the power usage, your p4 will consume something like 100W of power, if you conservatively estimate the cortex a8 to consume 1W (i believe the actual processor part consumes considerably less, and its the whole package including gpu and such that consumes 1W), then you could build a machine using 64 of these chips including interconnect circuitry which would trounce your 3ghz p4 while using a similar amount of power.
Also, vista is meant to be much slower than windows 7...
In the case of browsers tho, we have both gecko and webkit which are open and already capable of being built on many different architectures.
Portable non sourcecode is primarily of interest to those who want to make closed source apps, for people willing to publish their source porting isn't terribly difficult and usually gets done.
Windows CE has got a rather poor application base compared to windows/x86, you would end up with a decidedly sub par experience...
If you marketed it as being windows, and let customers assume compatibility with the x86 windows they're familiar with, users will be disappointed and you will get a bad reputation... If you advertise it as something else, it has to compete on a more level playing field with linux which is cheaper, has a much larger application base, is more mature on the arm architecture and is much easier to use...
Linux would also have much better compatibility with external peripherals like printers because most of the drivers are open and easily recompiled on arm, yet the windows drivers are typically shipped as binaries for x86 by the peripheral manufacturer who may be unwilling to produce arm drivers. I don't think many peripheral vendors produce drivers for windows ce...
Assuming windows was ported to arm, what about applications?
In the linux world, most apps come with sourcecode and are relatively easy to recompile for arm, you can use an existing distribution like debian which has thousands of applications already compiled for the arm processor, or if you are making such hardware it's not a huge effort to compile appropriate applications yourself. Most linux applications are coded fairly cleanly, those badly written ones which don't are likely to have been fixed already since linux runs on so many different platforms.
On windows however, most apps come without sourcecode, and only come with binaries for x86, the vast majority of apps don't even come with x86-64 versions. Only the original app vendor has the capability to port the application to arm, which brings up lots of problems...
Most windows app vendors are for-profit companies, they won't port to an architecture which doesn't have an existing user base. (conversely, a user base wont bulid around an architecture which has no apps) It's impossible to tell how portable their code is, for some it may be possible to just recompile, but my experience of commercial development is that the build environment is often quite fragile and would need quite a lot of work to target a new platform. Many people continue to run apps that aren't supported anymore by their original vendors, so no chance of porting.
The only thing windows really has going for it, is the existing base of applications... if you port it to arm, you take all that away so you end up with an immature expensive os with very few applications and very few users which compares very poorly to linux/arm. the only thing that would help it sell is the windows name, but it would earn itself a very bad reputation when the customers realized they had effectively been tricked.
1/2 - ms do very little to ensure compatibility with hardware, they are big enough that they don't have to, the maker of the hardware has to ensure their hardware is compatible with whatever ms put out or very few people will buy their hardware.
3 - only recently, up to and including vista they didn't care about performance on old hardware and just tried to force you to buy new kit.
4 - again see 1/2, if you dont make your apps compatible with the latest ms system then very few people will use it.
6 - same as 1/2, if you make peripherals that dont work with windows, they wont sell very well
7 - hahaha, ms supported tcp/ip and a few other internet related technologies because they had no choice, they do very little to bother supporting anything else that would be used for interoperability... they don't support any filesystems other than their own, they only support their own file sharing tech (nfs support requires an optional addon), they dont support ssh out of the box only telnet, and their applications are even worse like outlook which only supports its own proprietary protocol for calendering (no ical/caldav).
8 - apache is the tallest poppy in the webserving field, webservers make better targets for hackers than end user workstations because they typically have a lot more bandwidth...
10 - considering how much money they make from windows, they could easily afford to make the budget a lot higher... also, a lot of issues are down to poor design (including lack of design, nasty kludges built on top of crufty code)... windows is a mess that really needs to be dropped and started again... incidentally unix is a lot older, and yet suffers from far less stupid design decisions
11 - people bitch because the status quo isnt good enough, in any other market you would have competition and somewhere to go if one supplier screws you around with a shoddy product, windows has gotten its users so downtrodden that they now expect shoddy products and consider that normal... computers are now perceived by many as inherently unreliable devices.
Slashdot Mirror
OSX seems to do this too, it will say shutdown was cancelled due to some stupid app...
You can override it using the cli based shutdown command tho, then it will just kill -9 any misbehaving app like any other unix would.
Well, 'business' users typically have very little use for audio, so they will typically have a cheap audio chip and very poor shielding on the internal cabling, and the lousiest pair of speakers you have ever heard...
You might be better off with a consumer grade laptop, as the fact your eee sounds better seems to suggest.
Try running AmigaOS... I always found MacOS quite sluggish on equivalent (same cpu) hardware to an Amiga.
Is it just me, or is the windows 7 taskbar much taller (ie consumes more vertical space) than previous versions?
That, combined with thicker titlebars, doesn't make for very efficient use of vertical space on widescreen displays and especially on small netbook displays...
The Ubuntu netbook interface seems far more suited to such devices, it has no bar at the bottom, and the menu bar at the top combines with the titlebar of any open window to use very little of the very limited vertical space on the screen.
Yes, the license terms which you must agree to before you can go on with the download...
He said "any computer at my disposal"... What if those computers aren't running recent versions of windows (which is all those free viewers run on)?
People don't expect word to work flawlessly, it just doesn't and never has, they are willing to accept as normal that it's full of bugs and work around them...
These same people are not willing to accept similar bugs from any other product, many of which occur due to trying to implement a poorly designed, reverse engineered format. They demand that any other product work flawlessly in order to replace something that works very poorly indeed.
Maybe in the US wireless connections are pricey and patchy...
In small crowded european and asian countries like the uk, holland and japan 3g is ubiquitous, fairly cheap and widely used.
As for syncing, we used to use mysql replication to sync a copy of a wiki to every user's laptop.
99.9999% compatibility is a lot better than you get from MS, especially if you have a mix of different versions. Most businesses accept a much larger array of problems as simply being normal costs of doing business.
The number of businesses i have been to where compatibility problems, kludgy workarounds, downtime due to crashes/reboots etc are considered a day to day part of business.
Sometimes the formatting issues occur even on the same version, depending on how your printer is configured for instance, a well known issue which doesn't occur often in most companies since they typically use the same printer.
Installing security updates and service packs has also been known to modify the formatting...
What's hypocritical is that such problems are considered normal, and yet when openoffice exhibits similar issues (which are clearly unavoidable) it's cited as a reason why openoffice is unusable, but considering the format some formatting errors will be unavoidable... What version of word should openoffice try to emulate, or assume files are from?
No, but being patent encumbered does unless the patent holder declares the patent is free for anyone to implement under any terms they wish (ie they use the patent totally defensively and agree never to initiate any legal action against anyone over it).
The BSD license also has restrictions, just less of them... There is plenty of code released to the public domain with no restrictions whatsoever.
That said, the GPL places a few restrictions on you so that future users will be able to enjoy the same freedoms you did.
BSD and public domain allow you to make derivative versions which are not free.
Society works the same way (or is meant to), it has rules which restrict your freedom, but which are designed to ensure everyone receives a certain level of freedom and one person can't take that away from someone else.
Consider tho, that the iphone comes out of the box with gpl software on it (webkit at the very least)... And uses other gpl software during the development process (gcc etc)...
You are free to download this source, and also download the iphone sdk, compile this source and then execute it on the iphone simulator provided as part of the sdk... What costs $100 is the ability to test the code on your real iphone, or to submit your compiled binaries to the iphone store.
You could also try building this on an unofficial sdk for use with a jailbroken iphone...
You can't even settle things with a battle of the benchmarks: file system workloads vary so wildly that you can make a plausible argument for why any benchmark is either totally irrelevant or crucially important.
As pointed out, filesystem workloads vary massively, which is why it's good to have a choice of different filesystems which can be chosen based on individual requirements. Only offering a single filesystem like many other OS's do is extremely inefficient. One size does not fit all.
You can kill an X11 application too, using xkill, or terminal and a top...
Both of these examples assume the GUI is functional, which it might not be (some programs, eg vmware, can hijack all input, and if such a program crashes you often find yourself unable to interact with the gui even if technically it hasn't crashed)...
On Linux the gui is not an inherent part of the system, and the entire gui system (X11) can be killed and restarted... Explorer is not the equivalent of X11, it's not even the equivalent of a window manager, it's closer to nautilus.
Again, closed source software hampers development...
Despite Digital's best efforts, FX!32 still ran a lot slower than native Alpha code... Alpha was a lot more expensive than x86, and when running FX!32 not much faster.
With native Alpha code, the performance difference was more than enough to cover the price difference (a multi processor x86 box that, when running appropriately threaded code, could keep up with a single cpu Alpha cost a lot more).. Alpha boxes running Linux were really great for those of us who had them, they ran rings around people using x86...
I still have 5 Alpha boxes in the attic ranging from a 166mhz EV4 to a dual 750mhz EV67.
The same thing happened to IA64, running native code IA64 was much faster than x86, but it was too expensive and too slow when emulating x86 code...
Expect the same thing to happen with ARM, although ARM processors will still be around because of the phone/pda market, their use in laptops and desktops is likely to only take off among Linux users where closed source apps are few and far between.
Alpha was so much faster that for a while at least, the emulated apps ran faster than anything intel could offer and native apps just made intel look like a joke.
But it was more expensive, and only ran apps designed for NT, while most people were still running win16 and dos apps which the alpha couldn't emulate.
Alpha boxes running linux were great tho, the only thing that couldn't be recompiled was netscape, and you could run the native alpha version of netscape designed for digital unix, which always seemed far more stable than other versions of netscape.
Linux also had the ability to emulate x86 apps on the alpha, but it was slower than the version windows used.. I never really used it beyond testing because i didn't have any apps that were only available for x86 linux.
Old versions of office (97 i believe) were ported to Alpha back in the days.
OoO execution actually makes the cpu more efficient at executing less efficient code (or code which runs less efficiently by virtue of having been written for a different cpu)... If your willing to write/compile your code for a particular cpu then in-order execution works very well, thats why the ps3 and xbox 360 use in-order processors.
How bottom of the line?
Do you have proper video drivers?
I was always able to play videos with mplayer which wouldn't play on windows running on the same hardware, going all the way back to 300mhz machines playing dvd video or early divx...
I wouldn't say 1/6 of the power, the p4 is slower clock for clock than the p3 was, so your 3ghz p4 will be comparable to a 2-2.5ghz p3, making it about 2.5 times the performance of a 1ghz cortex-a8...
But compare the power usage, your p4 will consume something like 100W of power, if you conservatively estimate the cortex a8 to consume 1W (i believe the actual processor part consumes considerably less, and its the whole package including gpu and such that consumes 1W), then you could build a machine using 64 of these chips including interconnect circuitry which would trounce your 3ghz p4 while using a similar amount of power.
Also, vista is meant to be much slower than windows 7...
In the case of browsers tho, we have both gecko and webkit which are open and already capable of being built on many different architectures.
Portable non sourcecode is primarily of interest to those who want to make closed source apps, for people willing to publish their source porting isn't terribly difficult and usually gets done.
Windows CE has got a rather poor application base compared to windows/x86, you would end up with a decidedly sub par experience...
If you marketed it as being windows, and let customers assume compatibility with the x86 windows they're familiar with, users will be disappointed and you will get a bad reputation...
If you advertise it as something else, it has to compete on a more level playing field with linux which is cheaper, has a much larger application base, is more mature on the arm architecture and is much easier to use...
Linux would also have much better compatibility with external peripherals like printers because most of the drivers are open and easily recompiled on arm, yet the windows drivers are typically shipped as binaries for x86 by the peripheral manufacturer who may be unwilling to produce arm drivers. I don't think many peripheral vendors produce drivers for windows ce...
Assuming windows was ported to arm, what about applications?
In the linux world, most apps come with sourcecode and are relatively easy to recompile for arm, you can use an existing distribution like debian which has thousands of applications already compiled for the arm processor, or if you are making such hardware it's not a huge effort to compile appropriate applications yourself.
Most linux applications are coded fairly cleanly, those badly written ones which don't are likely to have been fixed already since linux runs on so many different platforms.
On windows however, most apps come without sourcecode, and only come with binaries for x86, the vast majority of apps don't even come with x86-64 versions.
Only the original app vendor has the capability to port the application to arm, which brings up lots of problems...
Most windows app vendors are for-profit companies, they won't port to an architecture which doesn't have an existing user base. (conversely, a user base wont bulid around an architecture which has no apps)
It's impossible to tell how portable their code is, for some it may be possible to just recompile, but my experience of commercial development is that the build environment is often quite fragile and would need quite a lot of work to target a new platform.
Many people continue to run apps that aren't supported anymore by their original vendors, so no chance of porting.
The only thing windows really has going for it, is the existing base of applications... if you port it to arm, you take all that away so you end up with an immature expensive os with very few applications and very few users which compares very poorly to linux/arm.
the only thing that would help it sell is the windows name, but it would earn itself a very bad reputation when the customers realized they had effectively been tricked.
1/2 - ms do very little to ensure compatibility with hardware, they are big enough that they don't have to, the maker of the hardware has to ensure their hardware is compatible with whatever ms put out or very few people will buy their hardware.
3 - only recently, up to and including vista they didn't care about performance on old hardware and just tried to force you to buy new kit.
4 - again see 1/2, if you dont make your apps compatible with the latest ms system then very few people will use it.
6 - same as 1/2, if you make peripherals that dont work with windows, they wont sell very well
7 - hahaha, ms supported tcp/ip and a few other internet related technologies because they had no choice, they do very little to bother supporting anything else that would be used for interoperability... they don't support any filesystems other than their own, they only support their own file sharing tech (nfs support requires an optional addon), they dont support ssh out of the box only telnet, and their applications are even worse like outlook which only supports its own proprietary protocol for calendering (no ical/caldav).
8 - apache is the tallest poppy in the webserving field, webservers make better targets for hackers than end user workstations because they typically have a lot more bandwidth...
10 - considering how much money they make from windows, they could easily afford to make the budget a lot higher... also, a lot of issues are down to poor design (including lack of design, nasty kludges built on top of crufty code)... windows is a mess that really needs to be dropped and started again... incidentally unix is a lot older, and yet suffers from far less stupid design decisions
11 - people bitch because the status quo isnt good enough, in any other market you would have competition and somewhere to go if one supplier screws you around with a shoddy product, windows has gotten its users so downtrodden that they now expect shoddy products and consider that normal... computers are now perceived by many as inherently unreliable devices.