using anti personel mines in the crawlspaces would make working down thier rather risky, would probablly be illegal in most civilised countries and would do a lot of damage to your network infrastructure if an intruder or employee set them off.
using a short patch lead on show like that seems totally braindead to me when he could just take the incoming network cable straight to the mac
he used a bus powered hub for all the USB ports, frankly i'm surprised he made the dvd drive work on that at all and he himself admitted that it didn't work on the usb hub with the new led connected.
also he doesn't mention the power of the heater but i wonder if he has thought about the rating of the wall socket that he has connected everything to. some heaters basically use up the entire rating of a standard 13A socket.
in any case if the boosters or the external tank are damages its going to be pretty irrelevent to the astronauts after they dispatch, its the orbitor that has to make a return trip from orbit and protect the fairly delicate humans inside whilst doing it.
but people are going to do it anyway especially with software which doesn't degrade in the same way that physical equipment does.
a leap hour would be very very disruptive when it came. Essentially what this seems to be advocating is pushing the problem into the future and making it come as one big horrible lump rather than as little tweaks that we hardly notice.
distro kernels are usually pretty damn stable. Most distros stick with a single upstream release of the kernel or a pair with one from the current stable series (currently 2.6) and one from the previous stable series (currently 2.4) for the lifetime of a stable release of the distro.
The time you have to go to kernel.org kernels or kernels built for your distro but not in the stable releases of the distro itself is when you have new or obscure hardware that your distro kernel doesn't get along with. Unfortunately things often break from one 2.6.x release to another so finindg a kernel that works with all your hardware can sometimes be tricky.
the MS JVM was fucked some time ago by disputes between MS and SUN. If I remember correctly MS is no longer allowed to ship it at all let alone update it and using software as security critical as a JVM for browser applets without secuirty updates availible is a BAD idea.
whats really fun with winapi is if you have a mostly ansi app (e.g. delphi vcl) and wan't to use a bit of unicode stuff.
MS documentation gives hardly any detail on how things work when you use unicode and ansi apis in the same app. some things like text out is obvious but other things (involving input messages common controls etc) most deffinately are not.
you could use a real encryption scheme but its not really worth it. If your app has the key to decode it then any attacker with access to the apps data dirs probablly can get the key rendering the encryption moot.
any encryption scheme used would therefore be only for protecting against accidental viewing by admins it has no use against an attacker or against an admin that really wan'ts the password.
and in windows 95 and 98 (never used me but i presume its the same) a rouge win16 app could bring down the user interface. Sure some 32 bit stuff might keep running but its little use if the GUI is dead.
most of the world is either 60HZ at arround 115V or 50HZ at arround 230V. so making a universal voltage input without also allowing for both frequencies would be kind of stupid.
RMS is the standard for power voltage measurement PERIOD.
as for 110V vs 120V i belive the idea is that 120V is the nominal supply voltage and 110V is the nominal utilisation voltage with the remaining 10V being an allowance for volt drop in wiring. i don't know how this works out in practice though.
actually the universal voltage ones are almost certainly switched mode which tend to be more efficiant than traditional linear power supplies.
as for saying wall warts are ineffciant its probablly true in a lot of cases but you can't trust car adaptors to be all that efficiant either and the losses of distributing at 12V can really add up.
i heared that the big problem with the PPRO was that manufacturing wasn't really ready for on die level 2 cache with the result that yields were low and therefore prices high.
thing is while software like burning rom is nice for burning whole cds its not much good for using cds to quickly move files arround and since the switch of common oem supplied burning software to nero few systems seem to be set up for packet writing (nero always seemed to keep incd far more low key than adaptec did with directcd) also the two major systems for packet writing are afaict not compatible with each other.
yeah CDR is nice but for small files it always seems more of a pain than floppies.
i used to use CDRW quite a bit. but then burner makers started shipping nero instead of adaptec and all my directcd discs became unwritable (i guess i could have bought a retail non locked adaptec easy cd creator but it was fairly pricey). I did consider switching to INCD but it all seemed to much hassle. for a while i used CDRWs like CDRs for tests etc but i gave up on that because it was so much hassle and write once cds became so cheap.
well its actually flash not rom in a modern board but anyway the rom/flash chips are fairly slow so its copied to ram on boot. thats the meaning of "system bios shadowed" and "video bios shadowed" in the boot process on some systems.
Slashdot Mirror
afaict gentoo is a bleeding edge distro that doesn't have stable releases.
i hope that last line was a joke
using anti personel mines in the crawlspaces would make working down thier rather risky, would probablly be illegal in most civilised countries and would do a lot of damage to your network infrastructure if an intruder or employee set them off.
if you are database driven (which he is) with a fairly slow server but a fast link the former is probablly worse.
if bandwidth is the constraining issue then splitting images between pages makes sense.
using a short patch lead on show like that seems totally braindead to me when he could just take the incoming network cable straight to the mac
he used a bus powered hub for all the USB ports, frankly i'm surprised he made the dvd drive work on that at all and he himself admitted that it didn't work on the usb hub with the new led connected.
also he doesn't mention the power of the heater but i wonder if he has thought about the rating of the wall socket that he has connected everything to. some heaters basically use up the entire rating of a standard 13A socket.
mirrordot sucks because they only get the first page and network mirror seems to have failed to get this story
however the original site seems to be alive albiet a bit intermittent atm
doesn't the ET burn up before it lands
in any case if the boosters or the external tank are damages its going to be pretty irrelevent to the astronauts after they dispatch, its the orbitor that has to make a return trip from orbit and protect the fairly delicate humans inside whilst doing it.
but people are going to do it anyway especially with software which doesn't degrade in the same way that physical equipment does.
a leap hour would be very very disruptive when it came. Essentially what this seems to be advocating is pushing the problem into the future and making it come as one big horrible lump rather than as little tweaks that we hardly notice.
yeah it just goes up and down. not only does it not go as high as the shuttle it also has very little horizontal movement.
the shuttle actually goes into orbit which requires it to achive a huge horizontal velocity as well as the problem of getting up there.
distro kernels are usually pretty damn stable. Most distros stick with a single upstream release of the kernel or a pair with one from the current stable series (currently 2.6) and one from the previous stable series (currently 2.4) for the lifetime of a stable release of the distro.
The time you have to go to kernel.org kernels or kernels built for your distro but not in the stable releases of the distro itself is when you have new or obscure hardware that your distro kernel doesn't get along with. Unfortunately things often break from one 2.6.x release to another so finindg a kernel that works with all your hardware can sometimes be tricky.
the MS JVM was fucked some time ago by disputes between MS and SUN. If I remember correctly MS is no longer allowed to ship it at all let alone update it and using software as security critical as a JVM for browser applets without secuirty updates availible is a BAD idea.
bad example modern engines are fules injected so they don't even have a carburator.
whats really fun with winapi is if you have a mostly ansi app (e.g. delphi vcl) and wan't to use a bit of unicode stuff.
MS documentation gives hardly any detail on how things work when you use unicode and ansi apis in the same app. some things like text out is obvious but other things (involving input messages common controls etc) most deffinately are not.
you could use a real encryption scheme but its not really worth it. If your app has the key to decode it then any attacker with access to the apps data dirs probablly can get the key rendering the encryption moot.
any encryption scheme used would therefore be only for protecting against accidental viewing by admins it has no use against an attacker or against an admin that really wan'ts the password.
and in windows 95 and 98 (never used me but i presume its the same) a rouge win16 app could bring down the user interface. Sure some 32 bit stuff might keep running but its little use if the GUI is dead.
not to mention slash can't seem to make up its mind if its outputting html 3.2 or 4.0 as the validator shows
g .nyud.net%3A8090%2Farticle.pl%3Fsid%3D05%2F07%2F15 %2F1212241%26threshold%3D-1%26tid%3D169%26tid%3D8
a shdot.org.nyud.net%3A8090%2Farticle.pl%3Fsid%3D05% 2F07%2F15%2F1212241%26threshold%3D-1%26tid%3D169%2 6tid%3D8&charset=(detect+automatically)&doctype=HT ML+4.01+Transitional
http://validator.w3.org/check?uri=www.slashdot.or
http://validator.w3.org/check?uri=http%3A%2F%2Fsl
notice that very few things are flagges as errors in both.
most of the world is either 60HZ at arround 115V or 50HZ at arround 230V. so making a universal voltage input without also allowing for both frequencies would be kind of stupid.
RMS is the standard for power voltage measurement PERIOD.
as for 110V vs 120V i belive the idea is that 120V is the nominal supply voltage and 110V is the nominal utilisation voltage with the remaining 10V being an allowance for volt drop in wiring. i don't know how this works out in practice though.
actually the universal voltage ones are almost certainly switched mode which tend to be more efficiant than traditional linear power supplies.
as for saying wall warts are ineffciant its probablly true in a lot of cases but you can't trust car adaptors to be all that efficiant either and the losses of distributing at 12V can really add up.
afaict amd got together with ms and managed to persuade them to go with amd64
once that was done intel had little choice but to implement amd64
i heared that the big problem with the PPRO was that manufacturing wasn't really ready for on die level 2 cache with the result that yields were low and therefore prices high.
it may be an idea to ocr it anyway purely for search perposes but i agree OCR is nowhere near reliable enough to jusitify throwing away scans.
what variant of TIFF exactly do they reccomend and why do they reccomend it over options like png?
thing is while software like burning rom is nice for burning whole cds its not much good for using cds to quickly move files arround and since the switch of common oem supplied burning software to nero few systems seem to be set up for packet writing (nero always seemed to keep incd far more low key than adaptec did with directcd) also the two major systems for packet writing are afaict not compatible with each other.
yeah CDR is nice but for small files it always seems more of a pain than floppies.
i used to use CDRW quite a bit. but then burner makers started shipping nero instead of adaptec and all my directcd discs became unwritable (i guess i could have bought a retail non locked adaptec easy cd creator but it was fairly pricey). I did consider switching to INCD but it all seemed to much hassle. for a while i used CDRWs like CDRs for tests etc but i gave up on that because it was so much hassle and write once cds became so cheap.
well its actually flash not rom in a modern board but anyway the rom/flash chips are fairly slow so its copied to ram on boot. thats the meaning of "system bios shadowed" and "video bios shadowed" in the boot process on some systems.