Bzip2 doesnt understand the filesystem atall, it's not meant to.. However if all those empty blocks are zeroed (as opposed to containing deleted data thats not yet overwritten) then it will compress the 39gig of zeroes down to about 2kb.
IRC already has an away status too, hence the/away command, unfortunately many clients and many servers don't implement it properly. mirc for instance shows you the away message every time you send a message to an away user (irritating), bitchx only shows it once.. Also the ircnet ircd atleast, only shows the away message when you do a/whois request if your on the same server, it used to show it regardless of what server you were on.. this is very annoying
Well, ANSI color has worked for years on any terminal which supports it, even without an irc client you can raw telnet to irc and ANSI works.. It took mirc _YEARS_ to support ANSI, but eventually they did.. while having their own non standard color scheme aswell. I remember back in the days, when mirc color codes started appearing, you saw garbage on the screen and it made it hard to read the text... We used to automatically ban anyone who used those codes.
Since when have microsoft been forward thinking? All their products are written with a "just good enough for now" approach, hence why they had to scrap the win9x code base and *supposedly* scrapped the iis codebase
Well, what purpose does gnome etc serve on a server anyway? But you do make the point for compiling your system from scratch, lets say you want gnome but don't want sound support... the redhat precompiled binaries have to include sound support for those who want it, a binary you compile yourself will only include what you want and nothing more. None of my gentoo servers have any support for sound whatsoever, nor for that matter, do i have gnome, or even X11 installed.
Also on their os's which theyre trying to sell for servers.. you are FORCED to have this crap (among other things.. a browser for instance) installed, even tho it isn't needed and for many server uses should be removed... Atleast on all my unix based servers i remove all the crap like that, or simply dont install it.. most unix flavors come with netscape or mozilla for instance, but none of them prevent you removing it.
Real is more cross platform infact... realplayer has been ported (poorly) to lots of unix flavors, including many running on non x86 hardware, mplayer on the other hand requires x86 dll's from windows which restricts it to x86 hardware.
I have a genuine Tulip card, DEC DE500 using the original 100mb 21140 chip.. Windows 2003 and 2000 detect the card and let me configure it during install, but once booted up fully they refuse to work and say that the card has a problem (they dont say what the problem is) This card works perfectly with Linux (on both alphA AND X86), Tru64(Digital UNIX) and OpenVMS.
The alternative to restarting gdm/X11 would be to reboot, windows forces you to do this.. linux atleast gives more knowlegeable people the choice. Aside from that, gdm/xdm is irritating, i always preferred a text based login and then starting X manually (or automatically in your login script) that way you can shut down X11 to conserve ram if your leaving a cpu/ram intensive cpu app running overnight, or if X fucks up somehow you can goto the console and fix it... I've seen systems in endless loops of trying (and failing) to start X over and over, rendering the system totally unuseable.
Even with older hardware, windows often requires you to install the drivers.. Otherwise the default drivers are very poor and don't take advantage of hardware features and consequently run very slowly.. An updated linux distribution on the other hand typically comes with the best available drivers for the card and doesn't require you to go and download additional ones... As for hardware which comes out _AFTER_ a particular os is packaged and distributed, well you cant expect an os to support a piece of hardware which doesn't yet exist, however linux being updated more regularly gives it an advantage here, the current consumer version of windows came out what... 2-3 years ago?
Gravis UltraSound, only supported by DOS.. couldn't get one of those working with win95 let alone anything later.. Linux handles it just fine, even the latest versions.
http://www.osnews.com/story.php?news_id=6136&page= 1 is a site showing benchmarks with different kinds of optimizations, yes i know it's for sparc but many things apply to other architectures aswell. OpenSSL is used by MANY users, in the web browser example for instance, openssl is often the back end used when accessing https sites.. As for word processor, openoffice seems noticeably more responsive when self-compiled than the downloadable precompiled binaries, mozilla is noticeably quicker too.. ESPECIALLY when the precompiled binaries are compiled using older compilers. As for small barely noticeable performance gains, i have 2 identical (in hardware terms) dell laptops here, one running redhat and one running gentoo. Identical versions of kde, mozilla, openoffice and many other tools run quicker on the gentoo box when the two are run side by side (this involves typing a command on both and hitting enter simultaneously on both, not very scientific but it works) Small performance gains, or slight memory usage reudction, in various different libraries and programs soon add up.. imagine how many library calls a large application like a word processor or web browser makes.. As for john the ripper, download the precompiled version and benchmark it on your duron, then download the source and compile it with mmx support and benchmark it again.. mmx extensions benefit john the ripper HUGELY.. but distributing a precompiled binary that used mmx would render it useless to any user without an mmx-supporting cpu.. you may say everyone has mmx nowadays, but consider 3dnow (amd only) sse/2/3 etc, all of which can provide performance improvements in some cases and all of which are unsupported in various currently-sold processors
Why is compiling from source a waste of time? many cpu intensive apps benefit greatly from being compiled from source, openssl for instance, and john the ripper for one.. precompiled binaries have to be as generic as possible, and thus can't make use of features present in newer cpus.. the alternative is multiple binaries, which serves to confuse users aswell as increase the program size.. Also many features can be disabled at compiletime, features which may create dependencies, increase binary size, slow down program operation.. a precompiled binary has to include everything enabled just incase someone needs it... The key is, to make compiling programs easier.. portage and ports already go a very long way towards this, and as computers become faster so compilation times will decrease
Giving away development tools would encourage people to learn to program and write their own programs.. These programs could and often would be better than MS tools, thus competing with them.. Why do you think windows has never included development tools when pretty much everything else does? Remember the days when every computer shipped with development tools and manuals?
Per CPU Hour? Well then theres an easy way to make money here... Just buy lots of machines with very slow cpus, or buy some ancient multiprocessor machines that contain lots of slow cpus.. Or perhaps run vmware, and create 500 virtual ultra-slow cpus on your one machine
Actually an xbox targetted game would work better if it hit the hardware directly, this would improve performance by not needing the extra overhead of the directx api
Well, i create subdomains under my domain for everything i sign up to, and if one of them gets spammed i will remove the dns for that subdomain.. The problem with the tags is that a spammer can remove or change the tags
Re:Nobody said MS invented computing...
on
Why PHBs Fear Linux
·
· Score: 2, Insightful
If windows never existed, then Apple, perhaps even Commodore/Amiga or Atari would have a place in business alongside sun, ibm, dec etc... x86 machines would still be around and mostly running os/2.. A diversity of systems coupled with the internet would FORCE vendors to adopt cross-platform standards, such standards did not exist in the past because there was little demand for data portability between systems, microsoft is still living in the old days when each vendor created their own systems which weren`t compatible with anything else. There are no standards, the sitation with microsoft is no different than if everyone used macs or everyone used amigas etc, the only sense of standardization comes from dominance.. The world would be a much better place if all the players had been forced to adopt standards to compete in the modern interconnected world.. Unix did this to some extent atleast, but not far enough in some ways.. That way we could choose the best tool for the job and be guaranteed compatibility with standards, rather than being forced to use the only tool for the job. Look at the auto industry, how would you like being forced to drive a yugo because no other cars are allowed on the road you have to drive to work?
Well Motorola hoped the PPC would be the successor to the M68k, a very successfull processor that was very widely used, easy to program for and very good for learning assembly on.
Vmware is partly emulation and partly virtualization, the x86 architecture isn`t really designed to virtualize like this.. hence the performance hit from vmware as it has to trap and emulate priveleged instructions.. the x86 architecture can virtualise as an original 8086 tho, but thats fairly useless nowadays
Slashdot Mirror
Bzip2 doesnt understand the filesystem atall, it's not meant to.. However if all those empty blocks are zeroed (as opposed to containing deleted data thats not yet overwritten) then it will compress the 39gig of zeroes down to about 2kb.
IRC already has an away status too, hence the /away command, unfortunately many clients and many servers don't implement it properly. mirc for instance shows you the away message every time you send a message to an away user (irritating), bitchx only shows it once.. Also the ircnet ircd atleast, only shows the away message when you do a /whois request if your on the same server, it used to show it regardless of what server you were on.. this is very annoying
Well, ANSI color has worked for years on any terminal which supports it, even without an irc client you can raw telnet to irc and ANSI works.. It took mirc _YEARS_ to support ANSI, but eventually they did.. while having their own non standard color scheme aswell.
I remember back in the days, when mirc color codes started appearing, you saw garbage on the screen and it made it hard to read the text... We used to automatically ban anyone who used those codes.
Since when have microsoft been forward thinking? All their products are written with a "just good enough for now" approach, hence why they had to scrap the win9x code base and *supposedly* scrapped the iis codebase
Well, what purpose does gnome etc serve on a server anyway? But you do make the point for compiling your system from scratch, lets say you want gnome but don't want sound support... the redhat precompiled binaries have to include sound support for those who want it, a binary you compile yourself will only include what you want and nothing more. None of my gentoo servers have any support for sound whatsoever, nor for that matter, do i have gnome, or even X11 installed.
Also on their os's which theyre trying to sell for servers.. you are FORCED to have this crap (among other things.. a browser for instance) installed, even tho it isn't needed and for many server uses should be removed... Atleast on all my unix based servers i remove all the crap like that, or simply dont install it.. most unix flavors come with netscape or mozilla for instance, but none of them prevent you removing it.
Real is more cross platform infact...
realplayer has been ported (poorly) to lots of unix flavors, including many running on non x86 hardware, mplayer on the other hand requires x86 dll's from windows which restricts it to x86 hardware.
Most people in holland use american keyboard layouts for some reason tho..
I have a genuine Tulip card, DEC DE500 using the original 100mb 21140 chip.. Windows 2003 and 2000 detect the card and let me configure it during install, but once booted up fully they refuse to work and say that the card has a problem (they dont say what the problem is)
This card works perfectly with Linux (on both alphA AND X86), Tru64(Digital UNIX) and OpenVMS.
The alternative to restarting gdm/X11 would be to reboot, windows forces you to do this.. linux atleast gives more knowlegeable people the choice.
Aside from that, gdm/xdm is irritating, i always preferred a text based login and then starting X manually (or automatically in your login script) that way you can shut down X11 to conserve ram if your leaving a cpu/ram intensive cpu app running overnight, or if X fucks up somehow you can goto the console and fix it... I've seen systems in endless loops of trying (and failing) to start X over and over, rendering the system totally unuseable.
Even with older hardware, windows often requires you to install the drivers.. Otherwise the default drivers are very poor and don't take advantage of hardware features and consequently run very slowly.. An updated linux distribution on the other hand typically comes with the best available drivers for the card and doesn't require you to go and download additional ones...
As for hardware which comes out _AFTER_ a particular os is packaged and distributed, well you cant expect an os to support a piece of hardware which doesn't yet exist, however linux being updated more regularly gives it an advantage here, the current consumer version of windows came out what... 2-3 years ago?
Gravis UltraSound, only supported by DOS.. couldn't get one of those working with win95 let alone anything later.. Linux handles it just fine, even the latest versions.
You can also use Xmove for graphical apps
http://www.osnews.com/story.php?news_id=6136&page= 1 is a site showing benchmarks with different kinds of optimizations, yes i know it's for sparc but many things apply to other architectures aswell.
OpenSSL is used by MANY users, in the web browser example for instance, openssl is often the back end used when accessing https sites..
As for word processor, openoffice seems noticeably more responsive when self-compiled than the downloadable precompiled binaries, mozilla is noticeably quicker too.. ESPECIALLY when the precompiled binaries are compiled using older compilers. As for small barely noticeable performance gains, i have 2 identical (in hardware terms) dell laptops here, one running redhat and one running gentoo. Identical versions of kde, mozilla, openoffice and many other tools run quicker on the gentoo box when the two are run side by side (this involves typing a command on both and hitting enter simultaneously on both, not very scientific but it works)
Small performance gains, or slight memory usage reudction, in various different libraries and programs soon add up.. imagine how many library calls a large application like a word processor or web browser makes..
As for john the ripper, download the precompiled version and benchmark it on your duron, then download the source and compile it with mmx support and benchmark it again.. mmx extensions benefit john the ripper HUGELY.. but distributing a precompiled binary that used mmx would render it useless to any user without an mmx-supporting cpu..
you may say everyone has mmx nowadays, but consider 3dnow (amd only) sse/2/3 etc, all of which can provide performance improvements in some cases and all of which are unsupported in various currently-sold processors
Why is compiling from source a waste of time? many cpu intensive apps benefit greatly from being compiled from source, openssl for instance, and john the ripper for one.. precompiled binaries have to be as generic as possible, and thus can't make use of features present in newer cpus.. the alternative is multiple binaries, which serves to confuse users aswell as increase the program size.. Also many features can be disabled at compiletime, features which may create dependencies, increase binary size, slow down program operation.. a precompiled binary has to include everything enabled just incase someone needs it...
The key is, to make compiling programs easier.. portage and ports already go a very long way towards this, and as computers become faster so compilation times will decrease
Giving away development tools would encourage people to learn to program and write their own programs.. These programs could and often would be better than MS tools, thus competing with them.. Why do you think windows has never included development tools when pretty much everything else does?
Remember the days when every computer shipped with development tools and manuals?
You could always hook the scanner up to a cheap computer operating as a server, and control it remotely over existing wireless lan protocols..
Per CPU Hour? Well then theres an easy way to make money here... Just buy lots of machines with very slow cpus, or buy some ancient multiprocessor machines that contain lots of slow cpus..
Or perhaps run vmware, and create 500 virtual ultra-slow cpus on your one machine
Actually an xbox targetted game would work better if it hit the hardware directly, this would improve performance by not needing the extra overhead of the directx api
This is because of a bug in the ultra-1 chips that allows a userspace program to deadlock the cpus when running in 64bit mode..
Well, i create subdomains under my domain for everything i sign up to, and if one of them gets spammed i will remove the dns for that subdomain..
The problem with the tags is that a spammer can remove or change the tags
If windows never existed, then Apple, perhaps even Commodore/Amiga or Atari would have a place in business alongside sun, ibm, dec etc... x86 machines would still be around and mostly running os/2..
A diversity of systems coupled with the internet would FORCE vendors to adopt cross-platform standards, such standards did not exist in the past because there was little demand for data portability between systems, microsoft is still living in the old days when each vendor created their own systems which weren`t compatible with anything else. There are no standards, the sitation with microsoft is no different than if everyone used macs or everyone used amigas etc, the only sense of standardization comes from dominance..
The world would be a much better place if all the players had been forced to adopt standards to compete in the modern interconnected world.. Unix did this to some extent atleast, but not far enough in some ways.. That way we could choose the best tool for the job and be guaranteed compatibility with standards, rather than being forced to use the only tool for the job.
Look at the auto industry, how would you like being forced to drive a yugo because no other cars are allowed on the road you have to drive to work?
There are also a large number of pirated windows installations out there, especially in poorer countries
Well Motorola hoped the PPC would be the successor to the M68k, a very successfull processor that was very widely used, easy to program for and very good for learning assembly on.
Vmware is partly emulation and partly virtualization, the x86 architecture isn`t really designed to virtualize like this.. hence the performance hit from vmware as it has to trap and emulate priveleged instructions..
the x86 architecture can virtualise as an original 8086 tho, but thats fairly useless nowadays