he said one CSS PIXEL would become 2x2 real pixels.
presumablly fonts will simply be drawn at that new bigger size not drawn at the web designer specified size and scaled. Images will have to be scaled unless they would previously have been scaled down but not nessacerally by something as crude as just repeating the pixels.
basically the aim is to let you run your screen in a very high dpi mode without making your text unreadable. Right now on most machines the pixels are still the limiting factor in how small readable text can be made (at least for people with good eyesight) but this is starting to change and apple wan't thier software to be ready when tiny pixels take off in a big way.
works for hire are already done that way (though with a term thats far longer than I think it should be) but creations by individuals are not.
this effectively means that unless you know who created something you can NEVER be sure its in the public domain!
imo a couple of decades from the initial release to public domain entry would be perfectly sufficiant for companies to cover thier costs. It would also stop companies just sitting on classics for years and force them to innovate to stay in buisness.
well you obviously aren't pushing 30 gigabytes (your capital B did mean bytes right?) per second down gig-e links unless you are running several hundred in paralell.
I assumed the poster you're replying to was referring to the need to check the reference count on every write but you don't!
if there is only one reference you just tell the mmu to let the process write directly to the page.
afaict cow works by denying a process write access to shared pages then trapping the errors that creates. pages with a refcount of 1 can still be written directly.
they are written more like tutorials than hard specs in many cases and much crucial information (such as how unicode and non-unicode apis interact) is simply missing.
so what happens is that people find out how the api really works by trial and error and then use that behaviour in thier applications.
ok that may indeed be the case. malloc is a very general api and as such using it directly for performance sensitive code is stupid.
with java you have NO REAL CHOICE but to allocate every object (and by extention any records you wan't since java doesn't have a seperate record type) directly from javas heap.
tell me which will take longer to allocate. an array of 10000 stuctures in a traditional language where the stuctures sit in the array or an array of 10000 objects in java where the objects are only referenced by the array?
ofc there are direct bytebuffers but those have two issues, 1: horrible api to work with compared with a language provided array of structures. 2: if any code in the program decides to load the class for an indirect bytebuffer your performance will plumet (this is apparently something they plan to fix in mustang, time will tell).
simalarlly if i wan't to return multiple values from a function in a traditional language i can do so either by reference parameters or by using a structure as the result type. If i wan't to do the same in java then i have to use an object (depending on how the codes structured it may be possible to reuse that object but that brings complexity of its own).
and sun designed jni to give them maximum flexibility in vm implementation not maximum performance of calls to native code. This is understandable but it means that if you are moving work to native code for perfomance reasons it has to be done in big chunks.
do they order some percentage more than whats sold and just hope it balances out?
or do they still check the store every few months or so?
or what?
afaict one big problem with the pick your own system is that staff don't find out when stock runs out even though the computer thinks its in. Regular full stock checks are an option but expensive.
i thought that penalty was removed in 2.4 but i cba to look it up now.
btw you can have multiple swap partitions, multiple swapfiles and even a mixture of the two. You can also add an remove swap while the system is running!
nope, as long as you position your writes within the original files boundries its size won't change. The size of a file only changes if you write past the existing end.
It's the linspire of debian. wasn't lindows (as it was known at the time people were making a big deal about it) a debian branch too? (i dunno if the current linspire stuff still is or not)
unless you were prepared to stick with either stuff from the distro or stuff developed in house running stuff on woody was becoming a major pita due to the fact that opensource devs tend to develop on bleeding edge distros (not that i like this practice but its a fact of life in the linux world). Also woody lacked autoconfiguration of X making installing desktop systems with it difficult.
running sarge is fine atm but it remains to be seen if they will get etch out in a reasonable timeframe. If not yet more people will be driven to the second tier distros that prepare thier own releases from debian unstable (ubuntu is the big one right now largely because its free and essentially has a wealthy sponsor but there were others in the past).
is that apple codenames seem much better known than ms ones.
people seem to use ms codenames when discussing the betas but as soon as there is an actual release they become largely forgotten very quickly (does apple actually put its codenames on the boxes? ms certainly doesn't).
build always allowed sectors that couldn't be seen at once to overlap and later build games found a way to join sectors floor-ceiling (duke unfortunately didn't and while there have been attempts to add it getting it right is a lot harder than it looks)
And with the additon of the polymost renderer you have true 3D viewpoints too.
but yeah you still map in the 2.5D style even though you can create true 3D constructions in the later variants.
umm activex isn't a scripting technology its a way to run full native code with basically no restrictions.
but anyway if you have the power to edit the hosts file you probablly have the power to write an exe to the filesystem and mod win.ini to start it. Once you've done that the exe can change the dns settings to point to itself and proxy any requests.
once people who belive in actually creating an encyclopedia are in power then provided they are diligent enough they will drive away those who just wan't to use it as a free hosting site.
therefore more people who belive in actually creating an encyclopedia will get voted into positions of power and those who don't will be either driven away or outright banned.
the dedication of the early few may have mattered to start with and jimbo does occasionally still make decrees but mostly wikipedias quality is driven by the loads of regular users who care.
You do realise that the device you used to write this on uses "component spectra". Also that the human eye works in just this way... the thing is while two lights may look the same under direct viewing that doesn't mean that surfaces will respond to them in the same way.
monitors are direct viewing so there is no reason to care about CRI
that the powercos hate having to expand capacity (high upfront cost that takes a LONG time to pay off) and they are often limited in how much they can push up prices to try and reduce demand by government regulation.
at least here in the uk some powercos seem to be actively pushing cfls
mmaping is often both simpler and lower overhead than reading from the file into a buffer only to pass the data straight back to the OS to send down a socket or whatever.
however on a 32 bit architecture you can't mmap anything above a few gigabytes and thats assuming your only handling the one large file in your process at a time.
thats true but most projects do provide a significant ammount of support for the users.
with smaller projects this comes from the developers themselves with larger projects it tends to come from other users (which can lead to blind leading the blind type problems) but virtually any sucessfull OSS project will have some sort of freely availible support.
red light district is tricky, i couldn't do it until someone showed me.
afaict there is no way to complete that level without passing through a secret (an airduct to be precise) to get behind the curtain and open it (a friend and i have checked quite hard in build and we can't seem to find any other switches that open it).
yeah while builds overuse of the keyboard (its the only app i've seen that used the two control keys seperately and used the numpad for totally different things to the corresponding keys on the main keyboard) took some getting used to The principle was very simple, draw a 2D map and then fill in the heights.
True 3D editors are much harder to learn as our main means of interacting with computers are all 2D. I tried once with unrealed but never got very far (the fact it was slow as molasses on my pc probablly didn't help much either).
Slashdot Mirror
he said one CSS PIXEL would become 2x2 real pixels.
presumablly fonts will simply be drawn at that new bigger size not drawn at the web designer specified size and scaled. Images will have to be scaled unless they would previously have been scaled down but not nessacerally by something as crude as just repeating the pixels.
basically the aim is to let you run your screen in a very high dpi mode without making your text unreadable. Right now on most machines the pixels are still the limiting factor in how small readable text can be made (at least for people with good eyesight) but this is starting to change and apple wan't thier software to be ready when tiny pixels take off in a big way.
works for hire are already done that way (though with a term thats far longer than I think it should be) but creations by individuals are not.
this effectively means that unless you know who created something you can NEVER be sure its in the public domain!
imo a couple of decades from the initial release to public domain entry would be perfectly sufficiant for companies to cover thier costs. It would also stop companies just sitting on classics for years and force them to innovate to stay in buisness.
well you obviously aren't pushing 30 gigabytes (your capital B did mean bytes right?) per second down gig-e links unless you are running several hundred in paralell.
I assumed the poster you're replying to was referring to the need to check the reference count on every write
but you don't!
if there is only one reference you just tell the mmu to let the process write directly to the page.
afaict cow works by denying a process write access to shared pages then trapping the errors that creates. pages with a refcount of 1 can still be written directly.
the windows api docs SUCK
they are written more like tutorials than hard specs in many cases and much crucial information (such as how unicode and non-unicode apis interact) is simply missing.
so what happens is that people find out how the api really works by trial and error and then use that behaviour in thier applications.
ok that may indeed be the case. malloc is a very general api and as such using it directly for performance sensitive code is stupid.
with java you have NO REAL CHOICE but to allocate every object (and by extention any records you wan't since java doesn't have a seperate record type) directly from javas heap.
tell me which will take longer to allocate. an array of 10000 stuctures in a traditional language where the stuctures sit in the array or an array of 10000 objects in java where the objects are only referenced by the array?
ofc there are direct bytebuffers but those have two issues,
1: horrible api to work with compared with a language provided array of structures.
2: if any code in the program decides to load the class for an indirect bytebuffer your performance will plumet (this is apparently something they plan to fix in mustang, time will tell).
simalarlly if i wan't to return multiple values from a function in a traditional language i can do so either by reference parameters or by using a structure as the result type. If i wan't to do the same in java then i have to use an object (depending on how the codes structured it may be possible to reuse that object but that brings complexity of its own).
and sun designed jni to give them maximum flexibility in vm implementation not maximum performance of calls to native code. This is understandable but it means that if you are moving work to native code for perfomance reasons it has to be done in big chunks.
on *nix it at least was standard practice to run something by using fork and then exec. i don't know if there are newer systems now or not.
In this environment cow made a lot of sense, the most likely thing to happen after a fork was that one of the processes would be replaced very soon.
also btw its every write to a new page that triggers a copy not every single write
do they order some percentage more than whats sold and just hope it balances out?
or do they still check the store every few months or so?
or what?
afaict one big problem with the pick your own system is that staff don't find out when stock runs out even though the computer thinks its in. Regular full stock checks are an option but expensive.
i thought that penalty was removed in 2.4 but i cba to look it up now.
btw you can have multiple swap partitions, multiple swapfiles and even a mixture of the two. You can also add an remove swap while the system is running!
far far more flexible than windows.
nope, as long as you position your writes within the original files boundries its size won't change. The size of a file only changes if you write past the existing end.
It's the linspire of debian.
wasn't lindows (as it was known at the time people were making a big deal about it) a debian branch too? (i dunno if the current linspire stuff still is or not)
unless you were prepared to stick with either stuff from the distro or stuff developed in house running stuff on woody was becoming a major pita due to the fact that opensource devs tend to develop on bleeding edge distros (not that i like this practice but its a fact of life in the linux world). Also woody lacked autoconfiguration of X making installing desktop systems with it difficult.
running sarge is fine atm but it remains to be seen if they will get etch out in a reasonable timeframe. If not yet more people will be driven to the second tier distros that prepare thier own releases from debian unstable (ubuntu is the big one right now largely because its free and essentially has a wealthy sponsor but there were others in the past).
is that apple codenames seem much better known than ms ones.
people seem to use ms codenames when discussing the betas but as soon as there is an actual release they become largely forgotten very quickly (does apple actually put its codenames on the boxes? ms certainly doesn't).
write to files it doesn't need to change the size of?
that should be enough to install linux in an image file on a NTFS partition if the base file was created first in windows.
build always allowed sectors that couldn't be seen at once to overlap and later build games found a way to join sectors floor-ceiling (duke unfortunately didn't and while there have been attempts to add it getting it right is a lot harder than it looks)
And with the additon of the polymost renderer you have true 3D viewpoints too.
but yeah you still map in the 2.5D style even though you can create true 3D constructions in the later variants.
umm activex isn't a scripting technology its a way to run full native code with basically no restrictions.
but anyway if you have the power to edit the hosts file you probablly have the power to write an exe to the filesystem and mod win.ini to start it. Once you've done that the exe can change the dns settings to point to itself and proxy any requests.
you are describing holywood holocaust not red light district.
I think its kind of self perpetuating
once people who belive in actually creating an encyclopedia are in power then provided they are diligent enough they will drive away those who just wan't to use it as a free hosting site.
therefore more people who belive in actually creating an encyclopedia will get voted into positions of power and those who don't will be either driven away or outright banned.
the dedication of the early few may have mattered to start with and jimbo does occasionally still make decrees but mostly wikipedias quality is driven by the loads of regular users who care.
You do realise that the device you used to write this on uses "component spectra". Also that the human eye works in just this way...
the thing is while two lights may look the same under direct viewing that doesn't mean that surfaces will respond to them in the same way.
monitors are direct viewing so there is no reason to care about CRI
that the powercos hate having to expand capacity (high upfront cost that takes a LONG time to pay off) and they are often limited in how much they can push up prices to try and reduce demand by government regulation.
at least here in the uk some powercos seem to be actively pushing cfls
is it allows large files to be mmaped.
mmaping is often both simpler and lower overhead than reading from the file into a buffer only to pass the data straight back to the OS to send down a socket or whatever.
however on a 32 bit architecture you can't mmap anything above a few gigabytes and thats assuming your only handling the one large file in your process at a time.
thats true but most projects do provide a significant ammount of support for the users.
with smaller projects this comes from the developers themselves with larger projects it tends to come from other users (which can lead to blind leading the blind type problems) but virtually any sucessfull OSS project will have some sort of freely availible support.
red light district is tricky, i couldn't do it until someone showed me.
afaict there is no way to complete that level without passing through a secret (an airduct to be precise) to get behind the curtain and open it (a friend and i have checked quite hard in build and we can't seem to find any other switches that open it).
yeah while builds overuse of the keyboard (its the only app i've seen that used the two control keys seperately and used the numpad for totally different things to the corresponding keys on the main keyboard) took some getting used to The principle was very simple, draw a 2D map and then fill in the heights.
True 3D editors are much harder to learn as our main means of interacting with computers are all 2D. I tried once with unrealed but never got very far (the fact it was slow as molasses on my pc probablly didn't help much either).
that with use of iphlpapi its trivial to bypass the windows dns code (and therefore the hosts file) altogether right?
any anti-malware vendor could do this very easilly (writing a dns client isn't hard if you are a reasonable coder).
come on you are exaggerating a bit there.
if a tank is in an accident with virtually any other vehircle including a SUV its going to be the other vehicle that is reduced to twisted metal.