i'd think on a campus you'd have a lot of computers that with appropriate software setup could be repurposed at night. (you have roomfulls of pcs for students to use don't you? are ALL those rooms open 24/7 i doubt it so just set it up to reboot after closing time into your clustering install and either shut down or reboot back into the standard install afterwards).
yeah the price is probablly a bit on the high side. though you didn't account for the fact that your own machines are unlikely to be at full load doing usefull work 24/7.
how much do places like specialist render farms charge by comparison
i'll belive they have really addressed the running as administrator by default when i see it. (and no power user doesn't count when you consider it easilly has sufficiant power to make a system takeover).
local telcos need many staff to maintain the lines to customers and the breakout systems that connect to them (when a line breaks or someone orders a new line or someone orders ISDN or ADSL on thier existing line someone has to go and physically do the needed wiring).
voip to the end user relies on an existing physical line to the end user generally provided by either a traditional telco or a cable tv co.
the point of A6 is to allow nets to be re-numbered and/or dual numbered (to allow a net to be multi homed without having to have an independantly advertised ip prefix) with minimal work
price/performance sucks as a way of measuring desktop systems. What matters far more is price/usable life (where usable life depends on what you plan to run) or cost/benifit
price/performance matters if you are buying say a render farm full of machines but in this case you need to take account of the whole cost of a machine (including things like the cost of the rackspace to mount it it) not of just one component.
e.g. if a CPU twice the speed costs 4 times as much it may still be worthwhile if it means you can cut the number of complete systems in half.
even with apache identifying itself as apache you STILL see requests that are targeting IIS security holes. It seems its easier for attackers to just fire at as many hosts as possible rather than checking them first.
and (until we get true virtual reality) the question is always how much damage (to the environment, the people involved, the area being trained in etc) is acceptable in the name of making the training experiance more realistic.
1: ratings bodies. This one is basically not optional for the producers, if industry led ratings schemes fail lots of countries would probablly replace them with government control and the industry really don't wan't that.
2: judges for major awards series (e.g. oscars). Here its a case of not wanting to get left out because the competition had thier films in the judges hands earlier.
the faq entry is not binding in any way its just an interpretation
and the bit from the license would not seem to prohibit dynamic linking provided you made sure you put no part of the gpled source (including headers) into your apps code.
until/unless this comes up in court its unlikely that the exact meaning will really be clarified,
Just like calling everything "Ethernet", even though it's now completely different in every way from the original.
not sure i'd agree with that comment about ethernet.
ethernet still uses 48 bit mac addresses to address hosts, still has a MTU of 1500, still relies on the idea of using broadcasts to find a host with a particular higher level identifier (machine name for netbeui, ip address for tcp/ip etc).
you can still hook up an old or propietry machine (where it is not possible to simply insert a modern pci or isa card) to a modern ethernet lan without too much difficulty. If it has an AUI port you can buy a 10baseT tranciver and if it has 10base2 you can get hold of hubs with a 10base2 port as well as the 10baseT ports. Once you've connected over to 10baseT you can just plug that straight into your modern network with no trouble.
iirc they use a static database for identifying which words to link and only actually fetch any text from when a user hovers over a link causing the start of the wp article to pop up in a tooltip like box.
actually clicking on the link brings up a menu with a few options including wikipedia and some kind of web search (not sure which one off hand).
i'll admit i dunno why google hosting is going nowhere atm, some offers are already being used (yahoo knams and lost oasis).
however the way the wikimedia server setup is currently structured all the high CPU work is done in one place. The other clusters are just caches. Maybe this will change but apparently mysql replication accross networks other than lans is not all that stable.
by basic C abi i mean choices like sizeof(int) sizeof(long) sizeof(long long) calling conventions etc that really have to be the same if you wan't any chance of linking with a lib. Yes there are alternative libcs availible which probablly have different prototypes etc for the stanrd functions and indeed you can install them in paralell if you so wish.
i was more responding to the comment that we should just use distro names i don't really have a strong position either way on the linux or GNU/linux issue (though i do think stallman is a prick for never letting it drop).
yes you can run linux binaries on freebsd and they mostly run its the little differences between two different implementations that get you (and generally in forms that are almost impossible to debug like your event loop hanging with 100% cpu after many days of running under load with no obvious cause). This is especailly bad with apps that use older syscalls (such as any app built with freepascal 1.0.x).
As for glibc compatibility accross versions yes it suffers from the same problem as other linux libs: coders who only care about compatibility in one direction but theese problems can be worked arround either by using special tools or by building your binaries on the oldest distro you expect your users to use.
Where am I going with this? Well, I think a Linux system should actually be refered to by the name of the distribution, i.e., you have a Red Hat OS, a Debian system, etc. There is a lot more that goes into putting together a distribution than just throwing together a kernel and a bunch of utilities. The only thing that should actually be called a GNU or GNU/Linux system is a distribution assembled and published by the GNU folks themselves.
thing is we need a way to talk about linux systems in general as they share a lot of attributes. The basic C abi is the same for all linux systems on a particular CPU architecture as is the syscall interface into the kernel used by staticlly linked applications and the standard libraries come from the same upstream sources (though they may be different versions). Ok there are some compatibility problems caused by library devs who only care about compatibility in one direction (e.g. if i build a GTK app on a system with GTK 2.4 it probablly won't run on a system with GTK 2.2 even if the app doesn't use any of the new features from GTK 2.2)
while there is ucLinux, standard linux is dependent on a processor with a mmu. Were there any other such processors easilly availible when linux started.
Also sometimes its easier to get stuff working first with as little complications as possible and refactor it later once you understand whats going on and where you need abstraction.
hmm you do have some level of conditionals in config.sys in ms-dos (you can create a menu that choses a section and then that section can use an include command to run other sections).
i'm pretty sure there are ready made scripts out there to build a msi from the latest firefox release if thats your preffered method of deployment.
here at uni they deploy firefox on demand through zenworks like pretty much every other app they have avilible. I think they build thier own packages for that though.
Slashdot Mirror
i'd think on a campus you'd have a lot of computers that with appropriate software setup could be repurposed at night. (you have roomfulls of pcs for students to use don't you? are ALL those rooms open 24/7 i doubt it so just set it up to reboot after closing time into your clustering install and either shut down or reboot back into the standard install afterwards).
yeah the price is probablly a bit on the high side. though you didn't account for the fact that your own machines are unlikely to be at full load doing usefull work 24/7.
how much do places like specialist render farms charge by comparison
i'll belive they have really addressed the running as administrator by default when i see it. (and no power user doesn't count when you consider it easilly has sufficiant power to make a system takeover).
afaict if you have mixed sized drives in a raid array the larger ones are just left with some unused space at the end hardly a disaster.
doesn't someone have to check that the person matches the passport picture?
local telcos need many staff to maintain the lines to customers and the breakout systems that connect to them (when a line breaks or someone orders a new line or someone orders ISDN or ADSL on thier existing line someone has to go and physically do the needed wiring).
voip to the end user relies on an existing physical line to the end user generally provided by either a traditional telco or a cable tv co.
what if the only pstn interconnects are outside the usa?
the point of A6 is to allow nets to be re-numbered and/or dual numbered (to allow a net to be multi homed without having to have an independantly advertised ip prefix) with minimal work
price/performance sucks as a way of measuring desktop systems. What matters far more is price/usable life (where usable life depends on what you plan to run) or cost/benifit
price/performance matters if you are buying say a render farm full of machines but in this case you need to take account of the whole cost of a machine (including things like the cost of the rackspace to mount it it) not of just one component.
e.g. if a CPU twice the speed costs 4 times as much it may still be worthwhile if it means you can cut the number of complete systems in half.
ofc if you are the milatery you probablly have the equipment to put them through 8 hours of actually whirling at high speed ;)
even with apache identifying itself as apache you STILL see requests that are targeting IIS security holes. It seems its easier for attackers to just fire at as many hosts as possible rather than checking them first.
am i right in thinking that skype uses a central login server and if you had control of that you could probablly mitm any conversation you wanted
and (until we get true virtual reality) the question is always how much damage (to the environment, the people involved, the area being trained in etc) is acceptable in the name of making the training experiance more realistic.
i was under the impression that the bfbc was largely industry led. am i wrong?
afaict screeners go to two main places.
1: ratings bodies. This one is basically not optional for the producers, if industry led ratings schemes fail lots of countries would probablly replace them with government control and the industry really don't wan't that.
2: judges for major awards series (e.g. oscars). Here its a case of not wanting to get left out because the competition had thier films in the judges hands earlier.
the faq entry is not binding in any way its just an interpretation
and the bit from the license would not seem to prohibit dynamic linking provided you made sure you put no part of the gpled source (including headers) into your apps code.
until/unless this comes up in court its unlikely that the exact meaning will really be clarified,
Just like calling everything "Ethernet", even though it's now completely different in every way from the original.
5 &SearchType=1&CategorySelectedId=11030&SearchTerms =Transceiver&PageMode=3&SearchKey=All&SearchMode=A ll&NavigationKey=110301 J&CategorySelectedId=11176&PageMode=1&NavigationKe y=11176,4294958628
not sure i'd agree with that comment about ethernet.
ethernet still uses 48 bit mac addresses to address hosts, still has a MTU of 1500, still relies on the idea of using broadcasts to find a host with a particular higher level identifier (machine name for netbeui, ip address for tcp/ip etc).
you can still hook up an old or propietry machine (where it is not possible to simply insert a modern pci or isa card) to a modern ethernet lan without too much difficulty. If it has an AUI port you can buy a 10baseT tranciver and if it has 10base2 you can get hold of hubs with a 10base2 port as well as the 10baseT ports. Once you've connected over to 10baseT you can just plug that straight into your modern network with no trouble.
http://www.dabs.com/productview.aspx?Quicklinx=6W
http://www.dabs.com/productview.aspx?Quicklinx=3P
iirc they use a static database for identifying which words to link and only actually fetch any text from when a user hovers over a link causing the start of the wp article to pop up in a tooltip like box.
actually clicking on the link brings up a menu with a few options including wikipedia and some kind of web search (not sure which one off hand).
i'll admit i dunno why google hosting is going nowhere atm, some offers are already being used (yahoo knams and lost oasis).
however the way the wikimedia server setup is currently structured all the high CPU work is done in one place. The other clusters are just caches. Maybe this will change but apparently mysql replication accross networks other than lans is not all that stable.
by basic C abi i mean choices like sizeof(int) sizeof(long) sizeof(long long) calling conventions etc that really have to be the same if you wan't any chance of linking with a lib. Yes there are alternative libcs availible which probablly have different prototypes etc for the stanrd functions and indeed you can install them in paralell if you so wish.
i was more responding to the comment that we should just use distro names i don't really have a strong position either way on the linux or GNU/linux issue (though i do think stallman is a prick for never letting it drop).
yes you can run linux binaries on freebsd and they mostly run its the little differences between two different implementations that get you (and generally in forms that are almost impossible to debug like your event loop hanging with 100% cpu after many days of running under load with no obvious cause). This is especailly bad with apps that use older syscalls (such as any app built with freepascal 1.0.x).
As for glibc compatibility accross versions yes it suffers from the same problem as other linux libs: coders who only care about compatibility in one direction but theese problems can be worked arround either by using special tools or by building your binaries on the oldest distro you expect your users to use.
iirc the reason ff moved from a zip to an installer was some registry key that was needed to make it work properly with a newer version of suns java.
r eleases/1.0.7/update/win32/
they do however distribute xpis which are apparently just zips with some scripts inside maybe those would be a good starting point for autobuilding custom packages without re-building firefox itself from source. ftp://ftp.mozilla.org/Public/mozilla.org/firefox/
Where am I going with this? Well, I think a Linux system should actually be refered to by the name of the distribution, i.e., you have a Red Hat OS, a Debian system, etc. There is a lot more that goes into putting together a distribution than just throwing together a kernel and a bunch of utilities. The only thing that should actually be called a GNU or GNU/Linux system is a distribution assembled and published by the GNU folks themselves.
thing is we need a way to talk about linux systems in general as they share a lot of attributes. The basic C abi is the same for all linux systems on a particular CPU architecture as is the syscall interface into the kernel used by staticlly linked applications and the standard libraries come from the same upstream sources (though they may be different versions). Ok there are some compatibility problems caused by library devs who only care about compatibility in one direction (e.g. if i build a GTK app on a system with GTK 2.4 it probablly won't run on a system with GTK 2.2 even if the app doesn't use any of the new features from GTK 2.2)
while there is ucLinux, standard linux is dependent on a processor with a mmu. Were there any other such processors easilly availible when linux started.
Also sometimes its easier to get stuff working first with as little complications as possible and refactor it later once you understand whats going on and where you need abstraction.
hmm you do have some level of conditionals in config.sys in ms-dos (you can create a menu that choses a section and then that section can use an include command to run other sections).
i'm pretty sure there are ready made scripts out there to build a msi from the latest firefox release if thats your preffered method of deployment.
here at uni they deploy firefox on demand through zenworks like pretty much every other app they have avilible. I think they build thier own packages for that though.