In XP there is an option in display properties, pretty sure that is at least five clicks though (maybe more, i can't remember if the option is visible immediately you click on the correct tab or not).
Can you imagine transferring from one shuttle to another while in orbit? worst case: line up the shuttles, get the people to put on thier spacesuits and climb accross from one airlock to the other, i can't see how it would be much harder than any other spacewalk.
btw how exactly does the space shuttle dock with the ISS and could that mechanism be used for a shuttle to shuttle transfer avoiding the need to do it as a spacewalk.
But reading Sarge's release notes first, would allow you to do it the proper way (from top of my mind, Woody->Sarge transition had a problem with apt and some circular dependencies, so you first should upgrade apt and related, and only then try dist-upgrade. sorry i should have clarified, i did upgrade aptitude first and checked i didn't have doc-base installed.
the tools still didn't propose a sane upgrade.
so even if you have read the release things are not gauranteed to behave on every setup, make sure you read what apt plans to do and that you do the upgrade when you can afford some downtime and you consider what your backup plan is if a remote box becomes unreachable.
nice to know but if its not used by default few users will ever encounter it.
imo having network cards swap thier configuration is a major bug. there are at least 3 soloutions to the problem (udev,mac based mappings in/etc/network/interfaces and carefull control of module load order) but it seems the distros don't use any of them by default.
but seriously, I think the IE team just appreciates the competition yeah from what i've heared the IE team was rather neglected by MS until firefox started gaining prominance, firefox probablly made MS give the IE team the budget to actually start improving the browser again rather than being on minimal maintinance.
right so you are saying that your company made a minor upgrade on the NT line of windows and threw out the 9x line of windows (presumablly if you were running both then 9x and NT lines you already had most of your critical apps working on 2K). Some minor testing for breakages needed but as you say ultimately no big deal.
compare that to a migration from windows+office+IE to linux+openoffice+firefox where you have to 1: find every macro heavy excel spreadsheet that is serving some vital function and get that functionality reimplemented in a way that doesn't depend on VBScript macros. 2: ditto for access databases 3: make sure every website (internal or external) your employees need to use is compatible with firefox. This may involve negotiating with other companies. 4: test every windows app you wan't to keep under wine and if it fails either fix it, get someone else to fix it or replace it. This may involve negotiating with other companies. 5: come up with an acceptable document exchange policy with anyone you exchange documents with (is openoffices word import/export good enough? do you wan't to use pdf? do you plan to convince them to install openoffice too?)
Dapper is recommended for a casual user, Edgy is for a little more advanced users, who know what to do when something breaks. then why is edgy the first thing on the download page?
and it doesn't say anything about it being unstable either!
The step you forgot is you go to Debian's web site and you read the release notes *first*. Then, you go by its instructions which usually are centered around an apt-get update && apt-get dist-upgrade instance. actually debian reccomends aptitude nowadays but on one of the woody-sarge upgrades i did neither apt-get nor aptitude proposed a sane dist-upgrade (both wanted to remove sendmail even after i tried manually upgrading it first) and i ended up doing apt-get upgrade and sorting out the rest by hand.
That's why my system is sync'd with unstable more-or-less every few days. I'm a Debian user, but I suppose this would apply to Ubuntu too. ok for one or two systems but its a major time sink and if you follow unstable you need to expect major breakage every so often (for example this means you need to know how to get into your system if the login processing is broken)
Under some conditions it is safer to run unstable every day than to upgrade to a whole new release every 6 months. but when its one big upgrade it will (or should) have been well tested by many before and if you are sensible you will schedule it at a time when you can afford some downtime (if you can't afford downtime then you should have redundant boxes anyway which can be upgraded one at a time).
While working on my phd I stopped doing that for a few months and the when I dist-upgraded again I had to do some real magic to avoid massive problems (like the python transition that tried to uninstall most of my python software) and of course you would get very little support for that upgrade as noone else is likely to have done it, with stable releases at least there are a few known cases.
On a side note apt developers could try to make "dist-upgrade" more similar to a day-by-day upgrade than to a single massive "apt-get install", trying to keep track of what package affect what other with every new version and than try to use all the information to recognize an update path that could be longer than the "massive install" but safer. This is a bad idea upgrading every version would waste insane ammounts of bandwidth and time and may mean installing horriblly buggy/broken versions and probablly wouldn't solve much.
people want something to work that the distro doesn't ship or wan't the latest version of something (sometimes for good reason, sometimes because they just like being on the bleeding edge) so they either add unofficial repositries or worse (worse because it means that there is no obvious evidence they have done it or where they got the package from) install packages manually.
then it comes time to upgrade and things break, while i agree this is partly the users fault i also belive that update managers designed for lusers need to get more intelligent about this.
maybe a popup saying something like "versions of core packages x,y and z on your system have a higher version number than those in the distribution you are trying to upgrade to" and giving the user the options to continue anyway, downgrade the packages in question or abort the upgrade would be a reasonable soloution.
as it stands at the moment apt-get (and its various frontends) is a nice timesaver for most installation tasks, but you REALLY still need to know how to use dpkg manually and sometimes even how to delete crashing prerm scripts when things go wrong.
solaris was never a big player on anything other than (expensive) sun hardware and even there linux was creeping in.
sun java is the primary implementation of java. That is it is what everyone writes there code to work with and what you expect to find if you purchase java hosting.
as to the license terms iirc the CDDL is a mozilla like license, incompatible with the GPL (but then so is nearly every copyleft license other than the GPL itself). Opensourcing the real thing will remove most of the motivation to develop clones (afaict the main motivation for developing the clones has been to get java into linux distros etc).
If the same intermediate language is generated from the Java frontend and the C frontend, and the Java bytecode backend handles that full language, it would be possible to compile C to Java bytecode.
my understanding was it was more like
java-->java bytecode-->GCC internal-->native code
the trouble with java bytecode is that if you wan't it to run on suns vm and certainly if you wan't it to run in any kind of restricted environment it has to pass the bytecode verifier. Short of essentially having an emulated main ram with a C heap inside it (possible but almost certainly not good for performance) passing the bytecode verifier with something compiled from C would be pretty damn hard.
the complication of AC is that AC motors like to run at a relatively fixed speed relative to the generators (there are ways arround this but they add to the complexity significantly).
much of the point of using electrical transmission in a locomotive is to decouple the tourque speed curve of the engine from that presented to the wheels.
yeah, development of commerce and industry nearly always goes hand in hand with increased power usage, thats just the way things are.
cost wise big power stations are generally efficiant and/or run off cheap fuels, therefore you buy your power from the grid whenever you can and install generators to deal with the grid failing.
Wow, just wow. That's just an incredible amount of power to be putting in a datacenter. Is it even possible for these centers to run off the grid? 2MW at 33KV is only 60 amps, a 33KV 60A line is hardly going to be a challange to construct.
so linking them to the grid really shouldn't be a problem.
you'd be insane to try and run a datacenter off generators all the time. Electricity from deisel generators costs several times what grid electric costs even if you pay consumer prices for grid electricity.
say 4 kilowatt used in each rack (a 42U rack with say a switch a patch panel and 40 machines at 100W each), another kilowatt per rack for aircon (probablly more in reality). thats only 500 racks to get up to two megawatts.
what i guess unisys is hoping is that the mailroom guys/secratery will pass the execs fortune subscription to the exec directly since he presumablly asked for it and therefore they will slip thier message past the normal protection schemes.
windows not installed by default which means if you wan't to run windows and don't wan't to get involved with piracy or grey market OEM stuff your going to have to pay full retail for windows (afaict most corp/educational licenses are upgrade/downgrade only).
right, but making int 64 bits won't nessacerally increase performance and may actually reduce it (not sure what it does in the case of AMD64, could be interesting to benchmark it).
btw if you wan't your code to be truely portable you should really avoid using int as you have no idea just how small it could be.
He'd observed that milkmaids and other people who regularly came into contact with cows and cowpox seemed to be immune to smallpox
and based on that he developed the theory that cowpox infection built an immunity to smallpox.
and he did an experiment to test his theory. If his theory had been wrong said experiment would have stood a good chance of killing his test subject and putting him in jail for a long time.
while modern trials (at least in the west) are far less reckless than that there is still a chance of stuff going wrong during trials and either killing or seriously injuring the trial subjects and there comes the difficult descision of how much risk to test subjects is acceptable in the hunt for new cures.
lost sales aren't the only reason for a manufacturer to be concerned, lost reputation is too.
if gear with your name on it starts failing a lot more than normal that is bad for your reputation whether you authorised the relase of that gear or not and gear that hurts somebody or starts fires is worse still.
if a product is made specifically to be a knockoff its hardly going to be made using good quality components or given good QA. And if a product is a reject from an official manufacturing run, well it was probablly rejected for a reason.
as for redundancy, planning the ammount of redundancy you need for a given availibility level depends on knowing how often kit is likely to fail and that depends on the quality control standards of the companies you buy from.
parent post appears to be a joke, british english firefox most certainly still uses the name cookie.
personally i think the mozillazine block entry is very US centric in saying "directly links to an.exe for one [correct for one country, but mostly-wrong for everyone else] locale", in an international sense british english is just as valid as american.
carefull though, none of those names are on the reserved list in rfc2606 and.cd is actually a valid top level domain already.
ab.cd is currently claims to be for sale!.tld while not official is so widely used in examples that its probablly safe but still best to stick to the officially reserved ones.
Slashdot Mirror
In XP there is an option in display properties, pretty sure that is at least five clicks though (maybe more, i can't remember if the option is visible immediately you click on the correct tab or not).
Can you imagine transferring from one shuttle to another while in orbit?
worst case: line up the shuttles, get the people to put on thier spacesuits and climb accross from one airlock to the other, i can't see how it would be much harder than any other spacewalk.
btw how exactly does the space shuttle dock with the ISS and could that mechanism be used for a shuttle to shuttle transfer avoiding the need to do it as a spacewalk.
But reading Sarge's release notes first, would allow you to do it the proper way (from top of my mind, Woody->Sarge transition had a problem with apt and some circular dependencies, so you first should upgrade apt and related, and only then try dist-upgrade.
sorry i should have clarified, i did upgrade aptitude first and checked i didn't have doc-base installed.
the tools still didn't propose a sane upgrade.
so even if you have read the release things are not gauranteed to behave on every setup, make sure you read what apt plans to do and that you do the upgrade when you can afford some downtime and you consider what your backup plan is if a remote box becomes unreachable.
nice to know but if its not used by default few users will ever encounter it.
/etc/network/interfaces and carefull control of module load order) but it seems the distros don't use any of them by default.
imo having network cards swap thier configuration is a major bug. there are at least 3 soloutions to the problem (udev,mac based mappings in
but seriously, I think the IE team just appreciates the competition
yeah from what i've heared the IE team was rather neglected by MS until firefox started gaining prominance, firefox probablly made MS give the IE team the budget to actually start improving the browser again rather than being on minimal maintinance.
right so you are saying that your company made a minor upgrade on the NT line of windows and threw out the 9x line of windows (presumablly if you were running both then 9x and NT lines you already had most of your critical apps working on 2K). Some minor testing for breakages needed but as you say ultimately no big deal.
compare that to a migration from windows+office+IE to linux+openoffice+firefox where you have to
1: find every macro heavy excel spreadsheet that is serving some vital function and get that functionality reimplemented in a way that doesn't depend on VBScript macros.
2: ditto for access databases
3: make sure every website (internal or external) your employees need to use is compatible with firefox. This may involve negotiating with other companies.
4: test every windows app you wan't to keep under wine and if it fails either fix it, get someone else to fix it or replace it. This may involve negotiating with other companies.
5: come up with an acceptable document exchange policy with anyone you exchange documents with (is openoffices word import/export good enough? do you wan't to use pdf? do you plan to convince them to install openoffice too?)
this can all take some time.
UL Lafayette and MIT
i'm assuming that by schools the parent poster meant the things they force kids to go to not degree level education.
Dapper is recommended for a casual user, Edgy is for a little more advanced users, who know what to do when something breaks.
then why is edgy the first thing on the download page?
and it doesn't say anything about it being unstable either!
The step you forgot is you go to Debian's web site and you read the release notes *first*. Then, you go by its instructions which usually are centered around an apt-get update && apt-get dist-upgrade instance.
actually debian reccomends aptitude nowadays but on one of the woody-sarge upgrades i did neither apt-get nor aptitude proposed a sane dist-upgrade (both wanted to remove sendmail even after i tried manually upgrading it first) and i ended up doing apt-get upgrade and sorting out the rest by hand.
the question then becomes why isn't this setup used by default?
That's why my system is sync'd with unstable more-or-less every few days. I'm a Debian user, but I suppose this would apply to Ubuntu too.
ok for one or two systems but its a major time sink and if you follow unstable you need to expect major breakage every so often (for example this means you need to know how to get into your system if the login processing is broken)
Under some conditions it is safer to run unstable every day than to upgrade to a whole new release every 6 months.
but when its one big upgrade it will (or should) have been well tested by many before and if you are sensible you will schedule it at a time when you can afford some downtime (if you can't afford downtime then you should have redundant boxes anyway which can be upgraded one at a time).
While working on my phd I stopped doing that for a few months and the when I dist-upgraded again I had to do some real magic to avoid massive problems (like the python transition that tried to uninstall most of my python software)
and of course you would get very little support for that upgrade as noone else is likely to have done it, with stable releases at least there are a few known cases.
On a side note apt developers could try to make "dist-upgrade" more similar to a day-by-day upgrade than to a single massive "apt-get install", trying to keep track of what package affect what other with every new version and than try to use all the information to recognize an update path that could be longer than the "massive install" but safer.
This is a bad idea upgrading every version would waste insane ammounts of bandwidth and time and may mean installing horriblly buggy/broken versions and probablly wouldn't solve much.
people want something to work that the distro doesn't ship or wan't the latest version of something (sometimes for good reason, sometimes because they just like being on the bleeding edge) so they either add unofficial repositries or worse (worse because it means that there is no obvious evidence they have done it or where they got the package from) install packages manually.
then it comes time to upgrade and things break, while i agree this is partly the users fault i also belive that update managers designed for lusers need to get more intelligent about this.
maybe a popup saying something like "versions of core packages x,y and z on your system have a higher version number than those in the distribution you are trying to upgrade to" and giving the user the options to continue anyway, downgrade the packages in question or abort the upgrade would be a reasonable soloution.
as it stands at the moment apt-get (and its various frontends) is a nice timesaver for most installation tasks, but you REALLY still need to know how to use dpkg manually and sometimes even how to delete crashing prerm scripts when things go wrong.
there is a big difference
solaris was never a big player on anything other than (expensive) sun hardware and even there linux was creeping in.
sun java is the primary implementation of java. That is it is what everyone writes there code to work with and what you expect to find if you purchase java hosting.
as to the license terms iirc the CDDL is a mozilla like license, incompatible with the GPL (but then so is nearly every copyleft license other than the GPL itself). Opensourcing the real thing will remove most of the motivation to develop clones (afaict the main motivation for developing the clones has been to get java into linux distros etc).
If the same intermediate language is generated from the Java frontend and the C frontend, and the Java bytecode backend handles that full language, it would be possible to compile C to Java bytecode.
my understanding was it was more like
java-->java bytecode-->GCC internal-->native code
the trouble with java bytecode is that if you wan't it to run on suns vm and certainly if you wan't it to run in any kind of restricted environment it has to pass the bytecode verifier. Short of essentially having an emulated main ram with a C heap inside it (possible but almost certainly not good for performance) passing the bytecode verifier with something compiled from C would be pretty damn hard.
the complication of AC is that AC motors like to run at a relatively fixed speed relative to the generators (there are ways arround this but they add to the complexity significantly).
much of the point of using electrical transmission in a locomotive is to decouple the tourque speed curve of the engine from that presented to the wheels.
yeah, development of commerce and industry nearly always goes hand in hand with increased power usage, thats just the way things are.
cost wise big power stations are generally efficiant and/or run off cheap fuels, therefore you buy your power from the grid whenever you can and install generators to deal with the grid failing.
Wow, just wow. That's just an incredible amount of power to be putting in a datacenter. Is it even possible for these centers to run off the grid?
2MW at 33KV is only 60 amps, a 33KV 60A line is hardly going to be a challange to construct.
so linking them to the grid really shouldn't be a problem.
you'd be insane to try and run a datacenter off generators all the time. Electricity from deisel generators costs several times what grid electric costs even if you pay consumer prices for grid electricity.
say 4 kilowatt used in each rack (a 42U rack with say a switch a patch panel and 40 machines at 100W each), another kilowatt per rack for aircon (probablly more in reality). thats only 500 racks to get up to two megawatts.
what i guess unisys is hoping is that the mailroom guys/secratery will pass the execs fortune subscription to the exec directly since he presumablly asked for it and therefore they will slip thier message past the normal protection schemes.
windows not installed by default
which means if you wan't to run windows and don't wan't to get involved with piracy or grey market OEM stuff your going to have to pay full retail for windows (afaict most corp/educational licenses are upgrade/downgrade only).
right, but making int 64 bits won't nessacerally increase performance and may actually reduce it (not sure what it does in the case of AMD64, could be interesting to benchmark it).
btw if you wan't your code to be truely portable you should really avoid using int as you have no idea just how small it could be.
He'd observed that milkmaids and other people who regularly came into contact with cows and cowpox seemed to be immune to smallpox
and based on that he developed the theory that cowpox infection built an immunity to smallpox.
and he did an experiment to test his theory. If his theory had been wrong said experiment would have stood a good chance of killing his test subject and putting him in jail for a long time.
while modern trials (at least in the west) are far less reckless than that there is still a chance of stuff going wrong during trials and either killing or seriously injuring the trial subjects and there comes the difficult descision of how much risk to test subjects is acceptable in the hunt for new cures.
lost sales aren't the only reason for a manufacturer to be concerned, lost reputation is too.
if gear with your name on it starts failing a lot more than normal that is bad for your reputation whether you authorised the relase of that gear or not and gear that hurts somebody or starts fires is worse still.
if a product is made specifically to be a knockoff its hardly going to be made using good quality components or given good QA. And if a product is a reject from an official manufacturing run, well it was probablly rejected for a reason.
as for redundancy, planning the ammount of redundancy you need for a given availibility level depends on knowing how often kit is likely to fail and that depends on the quality control standards of the companies you buy from.
parent post appears to be a joke, british english firefox most certainly still uses the name cookie.
.exe for one [correct for one country, but mostly-wrong for everyone else] locale", in an international sense british english is just as valid as american.
personally i think the mozillazine block entry is very US centric in saying "directly links to an
the registry is still in the USA whereever the registrar is and the registry certainly has the power to block a domain being added.
carefull though, none of those names are on the reserved list in rfc2606 and .cd is actually a valid top level domain already.
.tld while not official is so widely used in examples that its probablly safe but still best to stick to the officially reserved ones.
ab.cd is currently claims to be for sale!