yeah are crucial prepared to gaurantee shiments of X units per day for a few months at a fixed price though.
and if so will that fixed price be higher or lower than todays retail? (i would guess higher)
the thing is with retail if you can't get stock of a product (in this case ram) easilly then you can just jack up your prices to match and many customers will hold off on purchasing that particular product until the price comes down.
if on the other hand you are making pcs to standard specs you need a gauranteed supply of sufficiant ram to keep your production line going or you loose big time
there is the option of holding lots of stock (say a couple of months supply) and buying when you can get it cheap but that has issues of its own (depriciation storage capital)
the other significanct difference is in home is that they disable the network secuity features and force you to use "simple file sharing" mode.
which essentially means open access to everyone on your network read only access to everyone on your network of no access at all to your network.
to put it bluntly the network features in XP home are far far worse than those in 9X which it replaced and basically mean you either have to totally trust your network or not use it at all
iirc there are cpu count support differences too (though theese won't effect most users)
i seem to remember reading that raises in the prices of totally unbundled local loops forced them to buy all new kit to use shared local loops but the local telco has been awkward regarding theese as well.
with any service over exiting phone line only the regulation of the telco stands between the current state of play and being unable to run your service at all.
and i belive it would be very difficult (in terms of permisison) and expensive to run entirely new infrastructure.
remember the "power currupts power failure currupts absoloutely" incident
lukilly there was one slave db server being used for offline queries that wasn't actively replicating at the time of the power failure and it was possible to recover this database server and bring it up to date from the binlogs. From here the data could be restored fairly quickly to the other db servers (it still took more than a few hours) and service restored.
if that non replicating slave hadn't been it may have been nessacery to restore from a dump and replay a huge number of binlogs which could have taken a day or so.
it is belive that the problem was caused by hard drives claiming data was written to disc when in fact it was still in buffer.
unstable is meant to provide a mostly usable system that automated scripts can feed into testing. (IE unstable should only be unstable in terms of the fact that is can change at any time sometimes with breakage not unstable as in unusable)
experimental is for trying out REALLY radical stuff like new x servers developement versions of apt etc
mmm can't say i've used usenet myself but it seems to me you either run your own newserver (which afaict requires users to add accounts to thier client etc) or you try and get isp newservers to cary you group.
afaict mailinglists became popular because they worked for anyone with any mail client (including webmail setups) to view list mail in the same interface they use for normal mail.
a mailinglists structure standard that worked at the mailserver level would go a long way to solving theese problems (so the user doesn't have to do anything particularlly specail just the mailserver admin has to run some extra software and add some kind of dns record for mailinglists to use).
ie I run a mailing list, someone tries to subscribe. the mail provider has a record indicating they support this system so the process of opt-in is delegated to them and thier server requests copies of the mail which it distributes to any of thier users who are on the list (and lets them unsubscribe without having to go back to the original site).
this wouldn't exactly be difficult to code its just a matter of getting mailserver operators to accept it.
unfortunately you can't really have the libre without the gratis
if you had to pay per unit licensing for every little bit of code you used then collaborative programming would go nowhere as projects would quickly become far too expensive to use
people coding as a hobby will always be destructive to certain segments of the software development profession
software thats freely availible will also be destructive to some segments (companies that sell boxed software)
but at the end of the day there will always be software that companies need that noone wants to develop for fun and someone is going to have to pay to have that software developed because they need to use it.
if you know something is all uppercase then imo you may as well convert it to all lowercase since you aren't going to lose any information by doing so.
email was designed to be as reliable as a push service on the internet can be.
the thing with push services like email is that the recipiant doesn't know if there is a communication issue or if there is simply no one trying to push to them
well if its a replacement for init then it sure as hell is a daemon;)
basically it seems to me like they decided to replace init for a bloody good reason (reducing startup time significantly) and decided whilst they were doing so they may as well give scheduled tasks to the same daemon that handles boot and shutdown tasks (which makes quite a bit of sense why shouldn't i schedule a periodic task using the same tool i use to schedule a task on bootup or shutdown).
the thing with write caching is people expect when they hit save that thier document hits nonvoltile storage pretty damn quick
this means that if anything is saved on a system with traditional cache systems (without elaborate systems like flash or battery backed ram) the system has to spin up the drive soon after you hit save
on the other hand if the data can be stored in a non-volatile solid state manner by the drive there is no need to spin up the drive at all for small writes.
lots of people are getting deeply into java (not something i personally advise)
if sun ends up as a desperate public company in crisis who can say what will happen to java. Remember java is NOT free software dispite the fact its source is availible and if sun decide to tighten the screws you could find yourself with no updates to java unless you pay and no way to legally make them yourself.
Not to mention i belive that some binary distributions of java (freebsd springs to mind) are made based on revocable licenses so those could also dissapear if sun wants to tighten the screws on java users.
the trouble with gateways is they can be (and often are) ip blocked. With support in the client they either have to force update thier client (which isn't exactly great pr) or find a way to detect the difference between thier existing client and yours.
also i think the im nets consider the likes of gaim less threatening than the likes of jabber with gateways because gaim simply tries to make all nets equal whilst jabber with gateways allows you to contact the existing nets whilst pushing you towards using jabber itself.
whilst albert equated mass and energy the fact is that in normal life the vast majority of mass has no chance whatsoever of becoming energy.
the fact is that getting into orbit (i say orbit and not space because i'm not counting efforts such as rutans that just scrape space only those that can stay there) is right on the edge of what chemical rockets can do. We only manage it by such techniques as discarding parts (staged rockets shuttle boosters shuttle external tanks etc) and by reducing the non-fuel mass of the craft to an absoloute minimum.
so we are trying to control a lot (by normal human standards anyway) of energy in one place whilst keeping the mass of the systems that control it to an absoloute minium. sooner or later something is going to go wrong and blow with such systems.
then you have to get back and you have to do the trip back using stuff you carried into orbit at great expense so most breaking is done using the atnosphere. This however creates massive heat that needs special shielding which must deal with insane amounts of heat whilst being as lightwight as possible. That can fail too.
basically the overriding problem is that weight is at a premium on chemical rockets going to orbit and when weight it at a premium you simply can't make everything multiple times redundant.
it depends WHERE you are putting the wireless acess point
if its close enough to an existing socket that the LV flex from the wall wart to the access point then i agree that power over ethernet is pointless
on the other hand if you are installing a wireless access point in a location that has no existing services at all then POE would save you the hassle of getting mains nearby (i dunno about in the US but here in the UK you aren't allowed to put mains and ethernet in the same conduit/trunking comparment and i think many network installers are not qualified electricians).
dc rocks for long distance transmission especially if that transmission has to be underground or undersea (inductance and capacitance issues basically go away if you use DC) but converting between AC and DC and between different voltages of DC is relatively inefficiant (DC voltage conversion was pretty much impossible before the introduction of the SMPSU).
Slashdot Mirror
yeah are crucial prepared to gaurantee shiments of X units per day for a few months at a fixed price though.
and if so will that fixed price be higher or lower than todays retail? (i would guess higher)
the thing is with retail if you can't get stock of a product (in this case ram) easilly then you can just jack up your prices to match and many customers will hold off on purchasing that particular product until the price comes down.
if on the other hand you are making pcs to standard specs you need a gauranteed supply of sufficiant ram to keep your production line going or you loose big time
there is the option of holding lots of stock (say a couple of months supply) and buying when you can get it cheap but that has issues of its own (depriciation storage capital)
the other significanct difference is in home is that they disable the network secuity features and force you to use "simple file sharing" mode.
which essentially means open access to everyone on your network read only access to everyone on your network of no access at all to your network.
to put it bluntly the network features in XP home are far far worse than those in 9X which it replaced and basically mean you either have to totally trust your network or not use it at all
iirc there are cpu count support differences too (though theese won't effect most users)
my guess is that it will move into sid sometime not too long after sarge release and into the testing branch (etch) after that
when etch will release noone knows
i seem to remember reading that raises in the prices of totally unbundled local loops forced them to buy all new kit to use shared local loops but the local telco has been awkward regarding theese as well.
with any service over exiting phone line only the regulation of the telco stands between the current state of play and being unable to run your service at all.
and i belive it would be very difficult (in terms of permisison) and expensive to run entirely new infrastructure.
does wikipedia meet your definition of huge site?
remember the "power currupts power failure currupts absoloutely" incident
lukilly there was one slave db server being used for offline queries that wasn't actively replicating at the time of the power failure and it was possible to recover this database server and bring it up to date from the binlogs. From here the data could be restored fairly quickly to the other db servers (it still took more than a few hours) and service restored.
if that non replicating slave hadn't been it may have been nessacery to restore from a dump and replay a huge number of binlogs which could have taken a day or so.
it is belive that the problem was caused by hard drives claiming data was written to disc when in fact it was still in buffer.
unstable is meant to provide a mostly usable system that automated scripts can feed into testing. (IE unstable should only be unstable in terms of the fact that is can change at any time sometimes with breakage not unstable as in unusable)
experimental is for trying out REALLY radical stuff like new x servers developement versions of apt etc
mmm can't say i've used usenet myself but it seems to me you either run your own newserver (which afaict requires users to add accounts to thier client etc) or you try and get isp newservers to cary you group.
afaict mailinglists became popular because they worked for anyone with any mail client (including webmail setups) to view list mail in the same interface they use for normal mail.
a mailinglists structure standard that worked at the mailserver level would go a long way to solving theese problems (so the user doesn't have to do anything particularlly specail just the mailserver admin has to run some extra software and add some kind of dns record for mailinglists to use).
ie I run a mailing list, someone tries to subscribe. the mail provider has a record indicating they support this system so the process of opt-in is delegated to them and thier server requests copies of the mail which it distributes to any of thier users who are on the list (and lets them unsubscribe without having to go back to the original site).
this wouldn't exactly be difficult to code its just a matter of getting mailserver operators to accept it.
here in the uk i don't think i have EVER seen a urinal in the home
shops collages universitis offices tourist attractions etc sure they have urinals but never in the home.
unfortunately you can't really have the libre without the gratis
if you had to pay per unit licensing for every little bit of code you used then collaborative programming would go nowhere as projects would quickly become far too expensive to use
people coding as a hobby will always be destructive to certain segments of the software development profession
software thats freely availible will also be destructive to some segments (companies that sell boxed software)
but at the end of the day there will always be software that companies need that noone wants to develop for fun and someone is going to have to pay to have that software developed because they need to use it.
hmm i'm sure the losses i've seen quoted are much lower than that at least for modern generating plants
got a url backing up those claims of incrediblly low efficiancy.
i do agrree gas is a better choice for heating than resistive heaters though (heat pumps may be another issue)
my guess is its techically required but they only check the requirement if you get infected/cause network problems and you want to be reconnected
you mean http://www.acm.org/classics/sep95/ ?
if you know something is all uppercase then imo you may as well convert it to all lowercase since you aren't going to lose any information by doing so.
email was designed to be as reliable as a push service on the internet can be.
the thing with push services like email is that the recipiant doesn't know if there is a communication issue or if there is simply no one trying to push to them
this is funny but it looks like something that could be serious
can you confirm if its real info or a joke?
well if its a replacement for init then it sure as hell is a daemon ;)
basically it seems to me like they decided to replace init for a bloody good reason (reducing startup time significantly) and decided whilst they were doing so they may as well give scheduled tasks to the same daemon that handles boot and shutdown tasks (which makes quite a bit of sense why shouldn't i schedule a periodic task using the same tool i use to schedule a task on bootup or shutdown).
the thing with write caching is people expect when they hit save that thier document hits nonvoltile storage pretty damn quick
this means that if anything is saved on a system with traditional cache systems (without elaborate systems like flash or battery backed ram) the system has to spin up the drive soon after you hit save
on the other hand if the data can be stored in a non-volatile solid state manner by the drive there is no need to spin up the drive at all for small writes.
calling that a security hole is braindead
it may make it slightly easiler for a currupt sysadmin to get his users passwords but he can get them anyway by replacing tools.
iirc it was watchmaker not plumber
don't have an exact copy of the quote handy though
lots of people are getting deeply into java (not something i personally advise)
if sun ends up as a desperate public company in crisis who can say what will happen to java. Remember java is NOT free software dispite the fact its source is availible and if sun decide to tighten the screws you could find yourself with no updates to java unless you pay and no way to legally make them yourself.
Not to mention i belive that some binary distributions of java (freebsd springs to mind) are made based on revocable licenses so those could also dissapear if sun wants to tighten the screws on java users.
the trouble with gateways is they can be (and often are) ip blocked. With support in the client they either have to force update thier client (which isn't exactly great pr) or find a way to detect the difference between thier existing client and yours.
also i think the im nets consider the likes of gaim less threatening than the likes of jabber with gateways because gaim simply tries to make all nets equal whilst jabber with gateways allows you to contact the existing nets whilst pushing you towards using jabber itself.
sure but what is im software doing screwing with those things in the first place.
or putting them back automatically when someone tries to remove them.
if i remove something from the startup items or a toolbar from my browser it shouldn't go putting them back without permission.
seems to be a troll but i'll reply anyway
whilst albert equated mass and energy the fact is that in normal life the vast majority of mass has no chance whatsoever of becoming energy.
the fact is that getting into orbit (i say orbit and not space because i'm not counting efforts such as rutans that just scrape space only those that can stay there) is right on the edge of what chemical rockets can do. We only manage it by such techniques as discarding parts (staged rockets shuttle boosters shuttle external tanks etc) and by reducing the non-fuel mass of the craft to an absoloute minimum.
so we are trying to control a lot (by normal human standards anyway) of energy in one place whilst keeping the mass of the systems that control it to an absoloute minium. sooner or later something is going to go wrong and blow with such systems.
then you have to get back and you have to do the trip back using stuff you carried into orbit at great expense so most breaking is done using the atnosphere. This however creates massive heat that needs special shielding which must deal with insane amounts of heat whilst being as lightwight as possible. That can fail too.
basically the overriding problem is that weight is at a premium on chemical rockets going to orbit and when weight it at a premium you simply can't make everything multiple times redundant.
yes and no
it depends WHERE you are putting the wireless acess point
if its close enough to an existing socket that the LV flex from the wall wart to the access point then i agree that power over ethernet is pointless
on the other hand if you are installing a wireless access point in a location that has no existing services at all then POE would save you the hassle of getting mains nearby (i dunno about in the US but here in the UK you aren't allowed to put mains and ethernet in the same conduit/trunking comparment and i think many network installers are not qualified electricians).
the real problem is voltage conversion
dc rocks for long distance transmission especially if that transmission has to be underground or undersea (inductance and capacitance issues basically go away if you use DC) but converting between AC and DC and between different voltages of DC is relatively inefficiant (DC voltage conversion was pretty much impossible before the introduction of the SMPSU).