The plan was that we could use the win2k/xp version of 'su' (whatever it is called, I don't remember) to do things that needed elevated privs. IT DIDN'T WORK. Some of the child processes, for example, of burning a CD would spawn as your unprivileged context - meaning you couldn't burn a damn CD. You had to log out, and log back in with your priv account for a simple task like burning a CD.
The problem is that it doesn't completely isolate processes from each other. Many windows programs perform a check at startup to see if they're already running, and if they are they pass the command line parameters to the already running instance. Your CD burning software was probably doing this (presumably as part of a 'quick load' feature). Other software that does this includes Windows Explorer, meaning that you can't bring up control panel (which is a COM object loaded in-process by Explorer when you activate it) as your second user.
I have successfully used this feature to perform CD burning from an account that doesn't usually have adequate permission (using various cd recording tools, including cdrdao, cdrecord and Nero).
I'm sure if one were to send the patent office the sudo info, MS would argue that they have an "already running admin. process" that then actively accepts requests from other user processes.
And, in fact, that _is_ what they do. Windows 2000 and above have a "RunAs service" which is a process that runs with "Local Computer" security priveleges, which accepts requests from user processes that include a username & password, and execute the command passed if the password is correct.
This isn't to say that this is a unique and original idea. This document describes a similar facility in Plan 9, where a process called 'factotum' grants access to a user's data when authenticated. Having never used plan 9 myself, I'm not sure how similar this is in practice, but the description makes it sound like a substantially more useful implementation of the same original idea.
So you don't want a windowing system that is flexible, because people might want to take advantage of that flexibility?
Actually, no. What I want, which is moving in the opposite direction to the one Gosling suggests, is a windowing system that's flexible enough that I don't have to put up with whatever UI widgets the applications I run decide is best.
The system would be composed of the following components:
* A basic windowing system that allows a single application window to be created using widgets provided by a variety of (probably server-side) shared libraries. * A set of API specifications defining how applications can access the features of standard widget types (e.g. buttons, menus, list boxes, edit boxes, etc.) * A basic implementation of each standard widget type * An application which allows me to define, for all applications running on the server, precisely which widget implementation should be used for each widget type.
I'm aware that this is how some of the 'skinnable' toolkits available work (although some of them only allow you to vary physical appearance of the controls, and not the way they behave) -- I just want to see a standardised interface that allows it to be done server side (so that I don't have to replicate the configuration on all the client systems I use) and for all applications that don't specifically override my choices.
This is one of the worst parts in M$ Windows - once an app hangs, there is no way of closing or minimizing a window or simply of getting it out of the way.
Huh? If an app hangs in MS windows, I find clicking on the window's close button results in an "Application not responding -- do you want to kill the process?" dialog box popping up. Whereas X tends to cope really badly with hung clients, generally requiring you to use an entirely different command (e.g. "kill window" rather than "close window", although the names vary according to your window manager) to get rid of such windows.
In the UK, the word "candy" has mostly gone out of use, and usually refers to brown sugar or alt least "old fashioned" sweets based on brown sugar.
That seems to suggest that the usage of candy in the UK has changed to exclude various groups of confectionary... this is not the case. In UK usage, candy has always referred specifically to crystallized sugar, as used in some kinds of boiled sweets. It is the US usage which has changed over time to refer to confectionary in general. Look up the etymological information about it in a good dictionary (e.g. the OED).
Nobody's perfect. Nobody lives forever. Nobody can be in two places at once. Nobody can be in travel faster than the speed of light. Nobody knows everything. Nobody can fool all the people all the time. Nobody is liked by everyone. Nobody can have it all.
I think you could make a lucrative business out of buying these cameras and parting them out on Ebay.
Eventually, this would probably force the market into a true renting model where you have to return the camera.
No, it would force the manufacturer to complain to Ebay, who, given past performance, would immediately cancel all outstanding auctions for any part that is used in the camera, whether it came from one or not.
My company had one of its accounts suspended briefly last year when one of our clueless clients hired a US company to send e-mails for them to "1 million opt-in UK addresses".
BTW: how gullible can you get? A single opt-in list with about 5% of the Internet-connected population on it? Wow.
What part of "I'm not trolling..." in a prefactly reasonably post did the mods miss.??
They're probably reacting to the fact that you didn't read the story which is about a privately funded network before posting a suggestion on how the government should spend its money.
Re:Call me ignorant perhaps..
on
Semper WiFi
·
· Score: 3, Informative
But that isn't what they're talking about. What is being discussed is a situation where, for example, an article is talking about caffeine containing drinks, and you'll suddenly find a random link... perhaps they'd be talking about coffee and then when you click on the link you find it isn't more information as you'd expect, but is rather somewhere trying to _sell_ coffee.
I recommend the Ethiopian Yirgacheffe, by the way.;)
I've heard that Intel made a lot of mistakes with the P4 architecture. Optimisation guides for it recommend avoiding several techniques that were very good on previous designs... for instance use of the Load Effective Address instruction, which apparently is substantially slower on a P4 than it is on a P3.
One document said the mistake they made was in removing the ability for the processor to perform shift operations efficiently, presumably because a substantial number of transisters were previously committed to this and they wanted to reduce complexity.
And what's so bad about little endian? It allows for more efficient casting of pointers to integers, which can be handy. I don't see any benefits to big endian, other than it being the standard used for communication in most Internet related protocols.
In Three words: Variable Length Instructions
Without variable length instructions, it would have been impossible to extend the x86 as far as it has been extended without breaking backward compatibility. I understand the performance problems associated with them, and the fact that something like half of the power that goes into the chip is compensating for it, but this doesn't seem to have held up Intel or AMD from producing very fast chips using this architecture.
The extra energy and cooling problems are the price we pay for backward compatibility. Unfortunately, backward compatibility has proved itself to be absolutely essential.
I'm sure I remember hearing about a similar experiment about 10 years ago. They'd connected a 5x5 array of electrodes to a patient's optical nerves, and he could see vaguely defined objects. So this is exactly how much progress...?
I hate disc protection as much as the next guy, but if it's really such hard work to put a disc in your CD drive then maybe you need to lose some weight and take some exercise because you are clearly a lazy bastard.
My problem with this is that I'm usually using that CD drive for something else -- playing music.
Other things forcing people away from PC games include the game designers' attitude that only a cutting edge PC will do. Some of us are happy with our 4-5 year old machines and would rather keep them for a few more years, thanks.
When I write software, I make sure it is usable on a wide range of hardware. I've written stuff that's graphics intensive, and I make sure that you can run it without all the fancy features and in a low res mode. A 700MHz processor with plenty of RAM and a last generation 3D graphics card _should_ be able to run just about anything, but almost everything released today will just laugh at you if you try. And give you obscure error messages that only a DirectX programmer could hope to understand, if you're not very lucky.
I actually sometimes will buy the game then download the crack because I'm tired of dealing with shitty copy protection./rant
Didn't you know? That's what the cracks are for. Of course crack producers don't condone piracy. You should _always_ buy the game. They say so in the.nfo files.
It probably has something to do with the fact that most people using the words don't actually have a clue what they mean -- i.e. an intentionally weak argument or hypothesis presented with the explicit goal of finding and subsequently fixing the problems with it, whereas most such people use it just to mean a weak argument or hypothesis.
There should also be a "developer education" programme, whereby Windows developers are instructed in no uncertain terms that all software (except system administration software and "dangerous" software) should be runnable as an ordinary user, especially games.
I agree. Unfortunately, many legacy games (i.e. those written before about 2001) won't run on NT type systems without administrator access, so this won't solve the problem for a while. The percentage of any kind of application which can be _installed_ without administrator access is somewhere in the low single digits. The problem is so widespread that MS placed a kludge in the windows 2000 UI to get around it -- whenever you try to run a program called 'setup.exe', it gives you the option to run it as administrator (by supplying the administrator password, of course). Part of this is bad design by application programmers. But it is partially caused by problems with Windows itself (see below).
(If need be (for system-wide high score tables), games should be setgid (or the Windows equivalent of setgid) games,
I'm actually not aware of any equivalent mechanism in Windows to setuid or setgid, unfortunately. Probably the only way is to install a service that runs as local system and passes handles with adequate security to authenticated processes (this is how Windows itself implements what would be done through setuid on Unix systems - see the RunAs service in W2K and above). Again, you'd need the setup program to be run as administrator for that to work.
the equivalent of rm -rf ~/* ~/.* (I'm not sure what the equivalent incantation is in Windows land).
You'll be thinking of "echo y | del/sf %USERPROFILE%\*.*"
Incidentally, does Windows not even have su or sudo? I find that hard to believe (as such a tool helps conscientous users to use root only when essential, although it also makes it easier to overuse root).
The RunAs service I mentioned earlier is vaguely equivalent to su, and can be used for many things. Unfortunately, the system user interface is screwed up and cannot adequately support two users at a time -- for instance, only one copy of 'explorer.exe' can be running at a time. There are workarounds for most of the problems this causes but it can be very inconvenient.
If the proper APIs are used the only time a Windows box really *needs* to be restarted is after youve downloaded a kernel security update.
Its been a while since I looked at the relevant APIs, but surely you need to reboot if you've upgraded a DLL that was in use by an application at the time your install program ran... or have they fixed this problem?
That reminds me of the autonomously motion-guided poison dart in the movie, Dune.
Except of course (a) this isn't autonomous and motion guided, (b) it doesn't float motionlessly in mid-air until it spots its prey, (c) isn't lethal without serious modding and (d) just ain't that cool.
I think they mean "Microsoft-funded Study Cinches 10yr Deal". Although it could be "Microsoft funded study, cinches 10yr deal" implying (correctly, as it happens) that these are two things MS did: they funded a study, then cinched a 10 year deal (presumably related to the study, otherwise specifying them in the same sentence would be bad style).
Slashdot Mirror
The plan was that we could use the win2k/xp version of 'su' (whatever it is called, I don't remember) to do things that needed elevated privs. IT DIDN'T WORK. Some of the child processes, for example, of burning a CD would spawn as your unprivileged context - meaning you couldn't burn a damn CD. You had to log out, and log back in with your priv account for a simple task like burning a CD.
The problem is that it doesn't completely isolate processes from each other. Many windows programs perform a check at startup to see if they're already running, and if they are they pass the command line parameters to the already running instance. Your CD burning software was probably doing this (presumably as part of a 'quick load' feature). Other software that does this includes Windows Explorer, meaning that you can't bring up control panel (which is a COM object loaded in-process by Explorer when you activate it) as your second user.
I have successfully used this feature to perform CD burning from an account that doesn't usually have adequate permission (using various cd recording tools, including cdrdao, cdrecord and Nero).
I'm sure if one were to send the patent office the sudo info, MS would argue that they have an "already running admin. process" that then actively accepts requests from other user processes.
And, in fact, that _is_ what they do. Windows 2000 and above have a "RunAs service" which is a process that runs with "Local Computer" security priveleges, which accepts requests from user processes that include a username & password, and execute the command passed if the password is correct.
This isn't to say that this is a unique and original idea. This document describes a similar facility in Plan 9, where a process called 'factotum' grants access to a user's data when authenticated. Having never used plan 9 myself, I'm not sure how similar this is in practice, but the description makes it sound like a substantially more useful implementation of the same original idea.
So you don't want a windowing system that is flexible, because people might want to take advantage of that flexibility?
Actually, no. What I want, which is moving in the opposite direction to the one Gosling suggests, is a windowing system that's flexible enough that I don't have to put up with whatever UI widgets the applications I run decide is best.
The system would be composed of the following components:
* A basic windowing system that allows a single application window to be created using widgets provided by a variety of (probably server-side) shared libraries.
* A set of API specifications defining how applications can access the features of standard widget types (e.g. buttons, menus, list boxes, edit boxes, etc.)
* A basic implementation of each standard widget type
* An application which allows me to define, for all applications running on the server, precisely which widget implementation should be used for each widget type.
I'm aware that this is how some of the 'skinnable' toolkits available work (although some of them only allow you to vary physical appearance of the controls, and not the way they behave) -- I just want to see a standardised interface that allows it to be done server side (so that I don't have to replicate the configuration on all the client systems I use) and for all applications that don't specifically override my choices.
This is one of the worst parts in M$ Windows - once an app hangs, there is no way of closing or minimizing a window or simply of getting it out of the way.
Huh? If an app hangs in MS windows, I find clicking on the window's close button results in an "Application not responding -- do you want to kill the process?" dialog box popping up. Whereas X tends to cope really badly with hung clients, generally requiring you to use an entirely different command (e.g. "kill window" rather than "close window", although the names vary according to your window manager) to get rid of such windows.
In the UK, the word "candy" has mostly gone out of use, and usually refers to brown sugar or alt least "old fashioned" sweets based on brown sugar.
That seems to suggest that the usage of candy in the UK has changed to exclude various groups of confectionary... this is not the case. In UK usage, candy has always referred specifically to crystallized sugar, as used in some kinds of boiled sweets. It is the US usage which has changed over time to refer to confectionary in general. Look up the etymological information about it in a good dictionary (e.g. the OED).
Nobody's perfect.
Nobody lives forever.
Nobody can be in two places at once.
Nobody can be in travel faster than the speed of light.
Nobody knows everything.
Nobody can fool all the people all the time.
Nobody is liked by everyone.
Nobody can have it all.
Nobody gets the girl, too, apparently.
portman gnaa new-here kur0shin duplicate fp penisbird overlord base belong russia grits lameness evil bit beowulf goatsex
I'm thinking that you could rearrange these words into the definitve slashdot comment.
I've made a start, it goes like this:
In soviet russia, all your evil natalie portman duplicate goatse grits belong to base. (fp!)
Obviously this needs some work, but I can see it going places...
I think you could make a lucrative business out of buying these cameras and parting them out on Ebay.
Eventually, this would probably force the market into a true renting model where you have to return the camera.
No, it would force the manufacturer to complain to Ebay, who, given past performance, would immediately cancel all outstanding auctions for any part that is used in the camera, whether it came from one or not.
Okay, that acronymn is pretty crappy. What can YOU come up with?
ReVDoS - Reversed Vigalante Denial of Service?
My company had one of its accounts suspended briefly last year when one of our clueless clients hired a US company to send e-mails for them to "1 million opt-in UK addresses".
BTW: how gullible can you get? A single opt-in list with about 5% of the Internet-connected population on it? Wow.
What part of "I'm not trolling..." in a prefactly reasonably post did the mods miss.??
They're probably reacting to the fact that you didn't read the story which is about a privately funded network before posting a suggestion on how the government should spend its money.
It's a pun on this
But that isn't what they're talking about. What is being discussed is a situation where, for example, an article is talking about caffeine containing drinks, and you'll suddenly find a random link... perhaps they'd be talking about coffee and then when you click on the link you find it isn't more information as you'd expect, but is rather somewhere trying to _sell_ coffee.
;)
I recommend the Ethiopian Yirgacheffe, by the way.
I've heard that Intel made a lot of mistakes with the P4 architecture. Optimisation guides for it recommend avoiding several techniques that were very good on previous designs... for instance use of the Load Effective Address instruction, which apparently is substantially slower on a P4 than it is on a P3.
One document said the mistake they made was in removing the ability for the processor to perform shift operations efficiently, presumably because a substantial number of transisters were previously committed to this and they wanted to reduce complexity.
In Two words: Little Endian
And what's so bad about little endian? It allows for more efficient casting of pointers to integers, which can be handy. I don't see any benefits to big endian, other than it being the standard used for communication in most Internet related protocols.
In Three words: Variable Length Instructions
Without variable length instructions, it would have been impossible to extend the x86 as far as it has been extended without breaking backward compatibility. I understand the performance problems associated with them, and the fact that something like half of the power that goes into the chip is compensating for it, but this doesn't seem to have held up Intel or AMD from producing very fast chips using this architecture.
The extra energy and cooling problems are the price we pay for backward compatibility. Unfortunately, backward compatibility has proved itself to be absolutely essential.
I'm sure I remember hearing about a similar experiment about 10 years ago. They'd connected a 5x5 array of electrodes to a patient's optical nerves, and he could see vaguely defined objects. So this is exactly how much progress...?
I hate disc protection as much as the next guy, but if it's really such hard work to put a disc in your CD drive then maybe you need to lose some weight and take some exercise because you are clearly a lazy bastard.
My problem with this is that I'm usually using that CD drive for something else -- playing music.
Other things forcing people away from PC games include the game designers' attitude that only a cutting edge PC will do. Some of us are happy with our 4-5 year old machines and would rather keep them for a few more years, thanks.
When I write software, I make sure it is usable on a wide range of hardware. I've written stuff that's graphics intensive, and I make sure that you can run it without all the fancy features and in a low res mode. A 700MHz processor with plenty of RAM and a last generation 3D graphics card _should_ be able to run just about anything, but almost everything released today will just laugh at you if you try. And give you obscure error messages that only a DirectX programmer could hope to understand, if you're not very lucky.
I've been hearing it since my ZX Spectrum days, so that means ooohhhh twenty-four years?
I think you have your dates wrong. The Spectrum was launched in '82.
I actually sometimes will buy the game then download the crack because I'm tired of dealing with shitty copy protection. /rant
.nfo files.
Didn't you know? That's what the cracks are for. Of course crack producers don't condone piracy. You should _always_ buy the game. They say so in the
It probably has something to do with the fact that most people using the words don't actually have a clue what they mean -- i.e. an intentionally weak argument or hypothesis presented with the explicit goal of finding and subsequently fixing the problems with it, whereas most such people use it just to mean a weak argument or hypothesis.
There should also be a "developer education" programme, whereby Windows developers are instructed in no uncertain terms that all software (except system administration software and "dangerous" software) should be runnable as an ordinary user, especially games.
/sf %USERPROFILE%\*.*"
I agree. Unfortunately, many legacy games (i.e. those written before about 2001) won't run on NT type systems without administrator access, so this won't solve the problem for a while. The percentage of any kind of application which can be _installed_ without administrator access is somewhere in the low single digits. The problem is so widespread that MS placed a kludge in the windows 2000 UI to get around it -- whenever you try to run a program called 'setup.exe', it gives you the option to run it as administrator (by supplying the administrator password, of course). Part of this is bad design by application programmers. But it is partially caused by problems with Windows itself (see below).
(If need be (for system-wide high score tables), games should be setgid (or the Windows equivalent of setgid) games,
I'm actually not aware of any equivalent mechanism in Windows to setuid or setgid, unfortunately. Probably the only way is to install a service that runs as local system and passes handles with adequate security to authenticated processes (this is how Windows itself implements what would be done through setuid on Unix systems - see the RunAs service in W2K and above). Again, you'd need the setup program to be run as administrator for that to work.
the equivalent of rm -rf ~/* ~/.* (I'm not sure what the equivalent incantation is in Windows land).
You'll be thinking of "echo y | del
Incidentally, does Windows not even have su or sudo? I find that hard to believe (as such a tool helps conscientous users to use root only when essential, although it also makes it easier to overuse root).
The RunAs service I mentioned earlier is vaguely equivalent to su, and can be used for many things. Unfortunately, the system user interface is screwed up and cannot adequately support two users at a time -- for instance, only one copy of 'explorer.exe' can be running at a time. There are workarounds for most of the problems this causes but it can be very inconvenient.
If the proper APIs are used the only time a Windows box really *needs* to be restarted is after youve downloaded a kernel security update.
Its been a while since I looked at the relevant APIs, but surely you need to reboot if you've upgraded a DLL that was in use by an application at the time your install program ran... or have they fixed this problem?
That reminds me of the autonomously motion-guided poison dart in the movie, Dune.
Except of course (a) this isn't autonomous and motion guided, (b) it doesn't float motionlessly in mid-air until it spots its prey, (c) isn't lethal without serious modding and (d) just ain't that cool.
As for the original headline, I think it was just too fractured and unclear: "MS Funded Study Deciding Factor in 10yr Deal"
The current one isn't much better. There's a major punctuation fault in both, is the reason. Take the current title:
Microsoft Funded Study Cinches 10yr Deal.
Let's analyse that. "Microsoft Funded Study"... subject verb object, a complete phrase... "cinches"... verb???
I think they mean "Microsoft-funded Study Cinches 10yr Deal". Although it could be "Microsoft funded study, cinches 10yr deal" implying (correctly, as it happens) that these are two things MS did: they funded a study, then cinched a 10 year deal (presumably related to the study, otherwise specifying them in the same sentence would be bad style).