Trivial features such as APM support, SMP and Unix sockets shouldn't require a full recompile to activate.
SMP and APM are not "trivial features".
SMP support affects everything in the kernel. There are locks everywhere which are only important when running SMP. They are an incredible waste of CPU cycles on a uniprocessor kernel. If you don't care about this waste, just run a SMP kernel all the time. SMP kernels will run on a uniprocessor system.
Likewise, I believe APM and ACPI affect many regions throughout the kernel. Lots of drivers need hooks for them. (ACPI in particular. It does configuration as well as power management.)
Finally, UNIX domain sockets are relatively "trivial" in that the code is concentrated in one place. Because of this, they are modular. From linux-2.4.9-13/net/Config.in:
tristate 'Unix domain sockets' CONFIG_UNIX
The "tristate" option means it can be compiled as a module.
LaTeX isn't an open alternative to Word documents, though. It's a completely different system. Arguably, it's superior for many (some would say all) tasks, but it would take a lot of getting used to for many people.
A description of TeX/LaTeX/LyX for those unfamiliar with them:
TeX is a formatting system by Donald Knuth. In TeX, you specify how the document is to be rendered in a source file. I believe it's actually a Turing-complete language. It runs through and performs a lot of calculations to lay out the text in the best way...balancing underful and overfull boxes, etc. It tries to minimize a "badness" value. It produces very good output. TeX is extremely popular especially in scientific/academic environments because it's basically the only package that lays out complex equations correctly. If you own a math textbook, chances are it was made with LaTeX. (TeX is also a very good example of solid software. It's open-source and he will give you a $2.56 check for finding an error in his software or documentation. People frame them.)
LaTeX is a popular TeX macro package. It provides abstraction so you can specify content organization instead of presentation. By changing the documentclass, you can change how it is presented.
LyX is a graphical LaTeX editor. It lets you do most of what raw LaTeX does relatively easily. It still isn't really WYSIWYG. It gives a decent approximation of the final output as you edit, but the way you enter stuff is pretty different from Word, because the underlying format is so different.
> > My RoadRunner upload is capped at 384Kb/sec (45KB) and the download is 2Mb/sec.
> Mine is also capped at 50KB/s upload and (approx., not quite sure if its officially capped or just physical medium restrictions) have a download cap of 500KB/s. Lucky you
Assuming your numbers are correct, your service is faster than his. Pay attention to the case of those letters. Big 'B's are bytes; little 'b's are bits.
His service: 45KB/s (384Kb/s) up.
Your service: 50KB/s (400Kb/s) up.
His service: 256KB/s (2048Kb/s) (2Mb/s) down.
Your service: 500KB/s (4000Kb/s) (~3.9Mb/s) down.
With that said, I don't think your numbers are correct. They just don't make sense...his are all "nice" base-2 numbers, which is makes me trust them a lot more. How did you measure? I bet his number for the upload cap is correct and the download is not capped, just dependant on area, connection to the other end, other cable modem users, etc.
Alot of people seem to want for Linux to gain a large user base. I am simply stating what is needed if Linux is to do such.
I think you are wrong. It's not necessary for each developer (or any developer) to have the goal of a larger Linux userbase for that to happen. I assert that Linux developers have never had that goal, yet obviously the Linux userbase continues to grow.
Each developer pushes for the features that are important to him or her and making them happen. Collectively, all those features result in a pretty good product. As a consequence, Linux gains a larger userbase. It doesn't have to be their primary goal for it to happen.
You will never get people to do the "most efficient contribution" or the one that "contributes to the overall usability of the OS/GUI/ETC" because you will never get people to agree on what that is.
It's not a bad system. If you really want something, you make it happen. If you don't, you whine on Slashdot. If someone else eventually finds that same feature useful, you can take advantage of it as well.
If Linux wishs to pull ahead, then a balance has to be found.
Linux is an operating system. It doesn't wish anything. And the open-source development community isn't the borg. Developers are individuals. If this feature is important to you, why don't you do something? Do the work yourself or, if you don't know how, pay someone else to do it.
Someone will surely get around to it eventually. But when that feaure becomes important to them. Once again, developers are individuals. Many are not striving for Linux dominance in any given market; they are simply making software they want or need.
Concentrate on the features that people really want. Or at least that people think that they really want.
Absolutely. The developers concentrate on the features they want. Oh, is that not what you meant? Well, too bad.
Re:Using "ClearType" with XFree86, GNOME, KDE
on
Looking Ahead at GNOME 2
·
· Score: 3, Informative
This is the difference between Linux and Windows. Microsoft makes their cleartype technology enablable with a big large red button shouting "PUSH ME DARN YOU PUSH ME!". Linux puts it in a text file.
The parent post was a bit deceiving
The code to do this on Linux very new. It's part of the Xrender extension to XFree86 which was introduced in 4.1. It requires toolkit support. Qt has it. Gtk does not have it (though 2.0 will).
Editing that text file won't do much for you if you aren't using Qt. And if you are using Qt, you probably have it enabled by default...there's a checkbox to use it, though the XftConfig settings also affect it. In the future, KDE will probably have a fancy graphical configurator that will do everything the text file has. These things take time. I'm sure in Microsoftland a lot of Control Panel options started out life as something you could only change with regedit. You're just seeing the process with Linux instead of getting the final product.
I have yet to see a real legal reason for broadband [] Are you going to run out and buy the latest greatest $80 copy of redhat when you can pop over to linuxiso.org and just burn a copy??? [] Then there's gaming
Umm, those are legal. You don't need to pay $80 for a RedHat CD unless you get the commercial version. If you want to download it...go ahead. Gaming is also legal, providing you have paid for the game.
It is not just on page load / unload.. some popups / popunders are timed. The user does nothing to cause the new window to be opened.. thus the pref says 'Open Windows by themselves'. I do agree though that popups / popunders should be included in parenthesis, but the original option is worded pretty well.
I didn't know that, but I still don't like the "by themselves". All of these are happening as a result of an event. If the distinction is user input events vs. other events, maybe that should be stated clearly: "Open windows without user request"
If these are toggling a whole bunch of event types, I think there should be text that states exactly what it does. If it's too long or too technical to put as the text right next to the checkbox, that's fine. That's what "What's this?", tooltips, and Help buttons are for. In this case, it could be something like "When this is not selected, windows may not be opened during page load/unload or timer events. This stops pop-up and pop-under ads. Windows may always be opened during events corresponding directly to user input, such as mouse clicks or key presses."
Scripts and Windows
x Enable Javascript
x Open Windows by themselves
x Move or resize existing windows
x Make windows flip over or under other windows
x Change status bar text
x Change Images
x Create or change cookies
x Read cookies
I propose instead:
Scripts and Windows
x Enable Javascript
Javascript code may:
x Open windows on page load/unload (pop-up and pop-under ads)
x Move and resize existing windows
x Change window ordering (pop-under ads)
x Change status bar text
x Change images (mouseover highlighting)
x Create and change cookies
x Read cookies
First, the other options definitely belong as a sub-item as the first one, disabled when it gets disabled. (If it's actually this way in the dialog, sorry, my Mozilla isn't quite new enough to have your feature. I'm going by the bug report.) There should be a little label to explain the wording of the subitems, since they are stated as what the code is doing ("open a window") rather than what you're doing ("allowing them to open a window").
Second, I really don't like the "by themselves". Obviously everything in Javascript happens because of some event firing. I think on page load/unload is more clear. (Or some other way of precisely stating what events you're talking about.)
Third, it has in parenthesis a common use of several features. This should give a better understanding of what you'll be breaking.
Re:Those bastards hacked the linux kernel too!
on
al Qaeda Hacks XP?
·
· Score: 1
It's definitely not standard. I knew my FreeBSD system supported it, but assumed it was because they had GNU find; guess not. HP-UX 11.00 definitely does not support it.
I pointed out the longer version because (A) find is much, much more flexible and (B) someone posted the incorrect find, which is dangerous. I didn't see the need to repost the original.
Where the grep -r works, cool, you've saved some typing. But if you want to do something a bit more specific like only checking the C source and not the makefiles/documentation, pure grep doesn't work out. find+xargs+grep does:
find linux -name '*.[ch]' -print0 | xargs -0 grep 'blah'
Re:Those bastards hacked the linux kernel too!
on
al Qaeda Hacks XP?
·
· Score: 4, Offtopic
Off-topic but important
find/usr/src/linux -type f | xargs grep -i "a.*l.*q.*a.*e.*d.*a"
Don't do that. You're right; it is much faster. But it's less safe: it breaks if filenames contain whitespace. Ideally, they wouldn't, but remember the iTunes 2 installer? It axed entire partitions for this very reason. Filenames on a lot of systems do have spaces. Code that breaks on them is bad.
Instead, do this:
find/usr/src/linux -type f -print0 | xargs -0 grep -i "a.*l.*q.*a.*e.*d.*a"
It requires GNU find and xargs. But it really is much, much better. (It uses a null character instead of a space/newline as the seperator. Filenames can't contain nulls, since the system calls expect C-style null-terminated strings.)
(GNU bashers: There is a reason people prefer their tools. They really are better in a lot of cases.)
It takes years of invested time and experience to become at all proficient at any comprehensive productivity application. No one wants to throw that investment away, just to move to Linux.
And that is, I think, at the very core of it all, a usability problem. If it isn't exactly like the original, it is less usable for many folk.
Then there's no point in ever creating anything different. I think a better goal is to make it so much more efficient/friendlier/whatever than the original that it's worth the initial loss in productivity.
These may be great for sales, but ultimately you will build crap. Garbage in, garbage out. I would rather design good software that was well designed and efficient than vomit up mounds of bloat that will ultimately topple under its own weight.
Everything he said was focused on achieving commercial success. He gave solid examples of times when companies did not do as he suggests and failed commercially. I can't think of too many examples of companies that have succeeded overwhelmingly by doing otherwise.
On the other hand, he did not talk about sculpting perfect software people will use for decades to come. I can think of software that succeeded in this respect, and it didn't do it by following his advice. (TeX comes to mind.)
For example, he talked about bloatware, saying it is a good thing. "Features make users' lives better if they use them, and don't usually hurt if they don't." I disagree with this when talking about "hurt" as "making the software more painful to use" instead of "cutting sales". Extra features introduce more bugs and take away from the time programmers could be fixing other bugs. They shouldn't be added until everything else is perfectly solid and possibly not even then.
Really, I think there is a time for to listen to what this guy has to say and a time to completely ignore it. If you're developing a commercial project, his advice has a certain merit. If you're doing something as a hobby, producing a good piece of software you're proud of is much more important than producing a product before your competitors.
Set up NTFS ACLs properly - this includes giving SYSTEM rights to what needs to have it, along with the Administrators group, etc. Users should only have read access.
I've tried this before; it doesn't work out well. The operating system has support for it but applications don't. Intervideo WinDVD, for example, doesn't work on my machine unless I'm at least a Power User. There are plenty of other applications that don't work. That's the problem with adding important security features while maintaining backward compatibilitythe programs available for the system don't get along well with those features.
But a fundamental difference on Unix type systems is that files aren't inherantly executable based simply on their extension, someone can't just save a file from their email and execute it, they need to know at least enough to "chmod u+x" the file which should at least make them think about it.
The execute bit isn't a security feature when you're the owner of the file. [1] It wasn't designed to make you think more about executing files, and it doesn't make you think more about executing files. There's plenty of software that will automatically set the executable bit for you. Ever seen a uuencoded file, for example? They start out with something like this:
begin 755 evilworm
M(R$@+V)I;B]S
M;G0@=F%R:6%B
The 755 is, of course, the file's access permissions. In symbolic form, u=rwx,og=rx. uudecode honors these. [2] Anyone can execute this file.
[1] Not exactly true. You actually need to be able to write to the directory to modify its access permissions. But generally that's true, and if the file is readable you can always just copy it somewhere else first. Presumably some directory on the filesystem is writable, like your home directory or/tmp.
If I don't have a soul, would you care to explain:
Answers follow. Not necessarily true, but certainly no less valid than the religious party line.
How I can have free will and be in control of my actions
Chemical signals in your brain...the same way "lesser animals" do it. (They don't have souls, right?)
How I can feel love for my Creator, my family, etc.
See above.
What happens to my state of being when I die (that is, how does it feel to die)
You're dead. You are no more. You cease to exist. There is nothing.
Why society should hold me responsible for my actions, if they are just caused by a bunch of chemical changes
By holding you responsible for your actions, society can control them for the good of the whole. Even if they are "just" caused by a bunch of chemical changes, the carrot and the stick still apply.
Near-death experiences
Psychosomatic.
The undeniable actions and sightings of spirits, saints, the Virgin Mary, and other acts of God
I deny them.
What makes us different from animals
A matter of degree, not of kind. We are more intelligent.
I simple cannot accept your blanket denial of the human soul unless you take the time to address these questions.
Alright, each user has an embedded key, and they use that key to place votes.
How do you make sure each user gets exactly one key? As I said in my first post, a per-user key doesn't really help unless you have a method of ensuring that. And if you can ensure that, why can't you directly ensure each person only votes once? It just adds complexity without solving the problem.
This works both ways. Sure, you can find a route that avoids security cameras. But if you're the "bad guys" you now know where you need to install more security cameras.
Umm. No. If you were one of the "bad guys", you'd know where the cameras were already because you put them there. The site isn't giving them any information they don't already have, except possibly where people are clicking if they are monitoring the traffic. And that knowledge is of limited utility, considering that most of the visitors aren't even from New York and are clicking randomly.
Has [sic] ever appeared before in a Slashdot article? That amazed me. Granted, it was put there by the submitter, not an editor, but still that's pretty amazing.
With a google-written client communicating with their server, they should be able to come close (or at least make it very difficult to vote twice). There are lots of techniques that could work...dynamically generated keys, encryption, etc.
"If you think encryption will solve your problem, you don't understand encryption and you don't understand your problem." Bruce Schneier
Anything that could be done by your hypothetical client could also be done by a person who has used a debugger on it. It's just not theoretically possible to prevent something like that with an authentication key embedded in the software. Or per-client keys...remember, they have to get the key somehow. How do you restrict it to one key per person? That's back to the original problem.
And what about the firewall code? I don't know about you guys, but I just love spending a large chunk of a weekend learning the new firewall every time a new kernel series arrives. It just wouldn't be the same if it didn't fsck up my firewall scripts...
What are you talking about? You can still run 2.0.x ipfwadm stuff on the latest 2.4 kernel. Likewise 2.2.x ipchains stuff. iptables has backwards compatibility. If you want to use the newer interface, you can. No one's forcing you to.
From linux-2.4.14/Documentation/Configure.help:
pchains (2.2-style) support
CONFIG_IP_NF_COMPAT_IPCHAINS
This option places ipchains (with masquerading and redirection
support) back into the kernel, using the new netfilter
infrastructure. It is not recommended for new installations (see
`Packet filtering'). With this enabled, you should be able to use
the ipchains tool exactly as in 2.2 kernels.
If you want to compile it as a module, say M here and read
Documentation/modules.txt. If unsure, say `N'.
ipfwadm (2.0-style) support
CONFIG_IP_NF_COMPAT_IPFWADM
This option places ipfwadm (with masquerading and redirection
support) back into the kernel, using the new netfilter
infrastructure. It is not recommended for new installations (see
`Packet filtering'). With this enabled, you should be able to use
the ipfwadm tool exactly as in 2.0 kernels.
If you want to compile it as a module, say M here and read
Documentation/modules.txt. If unsure, say `N'.
Sensors in the units measure the finger movements and artificial intelligence and a language processor determine appropriate keystrokes or mouse movements.
Artificial intelligence? A language processor? "Appropriate" keystrokes or mouse movements? This sounds like a fancy way of saying "it guesses a lot".
Slashdot Mirror
SMP and APM are not "trivial features".
SMP support affects everything in the kernel. There are locks everywhere which are only important when running SMP. They are an incredible waste of CPU cycles on a uniprocessor kernel. If you don't care about this waste, just run a SMP kernel all the time. SMP kernels will run on a uniprocessor system.
Likewise, I believe APM and ACPI affect many regions throughout the kernel. Lots of drivers need hooks for them. (ACPI in particular. It does configuration as well as power management.)
Finally, UNIX domain sockets are relatively "trivial" in that the code is concentrated in one place. Because of this, they are modular. From linux-2.4.9-13/net/Config.in:
The "tristate" option means it can be compiled as a module.
A description of TeX/LaTeX/LyX for those unfamiliar with them:
TeX is a formatting system by Donald Knuth. In TeX, you specify how the document is to be rendered in a source file. I believe it's actually a Turing-complete language. It runs through and performs a lot of calculations to lay out the text in the best way...balancing underful and overfull boxes, etc. It tries to minimize a "badness" value. It produces very good output. TeX is extremely popular especially in scientific/academic environments because it's basically the only package that lays out complex equations correctly. If you own a math textbook, chances are it was made with LaTeX. (TeX is also a very good example of solid software. It's open-source and he will give you a $2.56 check for finding an error in his software or documentation. People frame them.)
LaTeX is a popular TeX macro package. It provides abstraction so you can specify content organization instead of presentation. By changing the documentclass, you can change how it is presented.
LyX is a graphical LaTeX editor. It lets you do most of what raw LaTeX does relatively easily. It still isn't really WYSIWYG. It gives a decent approximation of the final output as you edit, but the way you enter stuff is pretty different from Word, because the underlying format is so different.
> Mine is also capped at 50KB/s upload and (approx., not quite sure if its officially capped or just physical medium restrictions) have a download cap of 500KB/s. Lucky you
Assuming your numbers are correct, your service is faster than his. Pay attention to the case of those letters. Big 'B's are bytes; little 'b's are bits.
His service: 45KB/s (384Kb/s) up.
Your service: 50KB/s (400Kb/s) up.
His service: 256KB/s (2048Kb/s) (2Mb/s) down.
Your service: 500KB/s (4000Kb/s) (~3.9Mb/s) down.
With that said, I don't think your numbers are correct. They just don't make sense...his are all "nice" base-2 numbers, which is makes me trust them a lot more. How did you measure? I bet his number for the upload cap is correct and the download is not capped, just dependant on area, connection to the other end, other cable modem users, etc.
I think you are wrong. It's not necessary for each developer (or any developer) to have the goal of a larger Linux userbase for that to happen. I assert that Linux developers have never had that goal, yet obviously the Linux userbase continues to grow.
Each developer pushes for the features that are important to him or her and making them happen. Collectively, all those features result in a pretty good product. As a consequence, Linux gains a larger userbase. It doesn't have to be their primary goal for it to happen.
You will never get people to do the "most efficient contribution" or the one that "contributes to the overall usability of the OS/GUI/ETC" because you will never get people to agree on what that is.
It's not a bad system. If you really want something, you make it happen. If you don't, you whine on Slashdot. If someone else eventually finds that same feature useful, you can take advantage of it as well.
Linux is an operating system. It doesn't wish anything. And the open-source development community isn't the borg. Developers are individuals. If this feature is important to you, why don't you do something? Do the work yourself or, if you don't know how, pay someone else to do it.
Someone will surely get around to it eventually. But when that feaure becomes important to them. Once again, developers are individuals. Many are not striving for Linux dominance in any given market; they are simply making software they want or need.
Concentrate on the features that people really want. Or at least that people think that they really want.
Absolutely. The developers concentrate on the features they want. Oh, is that not what you meant? Well, too bad.
This is the difference between Linux and Windows. Microsoft makes their cleartype technology enablable with a big large red button shouting "PUSH ME DARN YOU PUSH ME!". Linux puts it in a text file.
The parent post was a bit deceiving
The code to do this on Linux very new. It's part of the Xrender extension to XFree86 which was introduced in 4.1. It requires toolkit support. Qt has it. Gtk does not have it (though 2.0 will).
Editing that text file won't do much for you if you aren't using Qt. And if you are using Qt, you probably have it enabled by default...there's a checkbox to use it, though the XftConfig settings also affect it. In the future, KDE will probably have a fancy graphical configurator that will do everything the text file has. These things take time. I'm sure in Microsoftland a lot of Control Panel options started out life as something you could only change with regedit. You're just seeing the process with Linux instead of getting the final product.
I have yet to see a real legal reason for broadband [] Are you going to run out and buy the latest greatest $80 copy of redhat when you can pop over to linuxiso.org and just burn a copy??? [] Then there's gaming
Umm, those are legal. You don't need to pay $80 for a RedHat CD unless you get the commercial version. If you want to download it...go ahead. Gaming is also legal, providing you have paid for the game.
Soyou disproved your own point. Good work.
It is not just on page load / unload.. some popups / popunders are timed. The user does nothing to cause the new window to be opened.. thus the pref says 'Open Windows by themselves'. I do agree though that popups / popunders should be included in parenthesis, but the original option is worded pretty well.
I didn't know that, but I still don't like the "by themselves". All of these are happening as a result of an event. If the distinction is user input events vs. other events, maybe that should be stated clearly: "Open windows without user request"
If these are toggling a whole bunch of event types, I think there should be text that states exactly what it does. If it's too long or too technical to put as the text right next to the checkbox, that's fine. That's what "What's this?", tooltips, and Help buttons are for. In this case, it could be something like "When this is not selected, windows may not be opened during page load/unload or timer events. This stops pop-up and pop-under ads. Windows may always be opened during events corresponding directly to user input, such as mouse clicks or key presses."
And actually, the existing interface in 0.9.7 is:
[...] Allow scripts to do the following: [...]
Ahh, cool. I'm happy with that wording.
Well, if you have any suggestions, do share them.
I don't like the options stated here:
Scripts and Windows
x Enable Javascript
x Open Windows by themselves
x Move or resize existing windows
x Make windows flip over or under other windows
x Change status bar text
x Change Images
x Create or change cookies
x Read cookies
I propose instead:
Scripts and Windows
x Enable Javascript
Javascript code may:
x Open windows on page load/unload (pop-up and pop-under ads)
x Move and resize existing windows
x Change window ordering (pop-under ads)
x Change status bar text
x Change images (mouseover highlighting)
x Create and change cookies
x Read cookies
First, the other options definitely belong as a sub-item as the first one, disabled when it gets disabled. (If it's actually this way in the dialog, sorry, my Mozilla isn't quite new enough to have your feature. I'm going by the bug report.) There should be a little label to explain the wording of the subitems, since they are stated as what the code is doing ("open a window") rather than what you're doing ("allowing them to open a window").
Second, I really don't like the "by themselves". Obviously everything in Javascript happens because of some event firing. I think on page load/unload is more clear. (Or some other way of precisely stating what events you're talking about.)
Third, it has in parenthesis a common use of several features. This should give a better understanding of what you'll be breaking.
It's definitely not standard. I knew my FreeBSD system supported it, but assumed it was because they had GNU find; guess not. HP-UX 11.00 definitely does not support it.
I pointed out the longer version because (A) find is much, much more flexible and (B) someone posted the incorrect find, which is dangerous. I didn't see the need to repost the original.
Where the grep -r works, cool, you've saved some typing. But if you want to do something a bit more specific like only checking the C source and not the makefiles/documentation, pure grep doesn't work out. find+xargs+grep does:
Off-topic but important
Don't do that. You're right; it is much faster. But it's less safe: it breaks if filenames contain whitespace. Ideally, they wouldn't, but remember the iTunes 2 installer? It axed entire partitions for this very reason. Filenames on a lot of systems do have spaces. Code that breaks on them is bad.
Instead, do this:
It requires GNU find and xargs. But it really is much, much better. (It uses a null character instead of a space/newline as the seperator. Filenames can't contain nulls, since the system calls expect C-style null-terminated strings.)
(GNU bashers: There is a reason people prefer their tools. They really are better in a lot of cases.)
Then there's no point in ever creating anything different. I think a better goal is to make it so much more efficient/friendlier/whatever than the original that it's worth the initial loss in productivity.
These may be great for sales, but ultimately you will build crap. Garbage in, garbage out. I would rather design good software that was well designed and efficient than vomit up mounds of bloat that will ultimately topple under its own weight.
Everything he said was focused on achieving commercial success. He gave solid examples of times when companies did not do as he suggests and failed commercially. I can't think of too many examples of companies that have succeeded overwhelmingly by doing otherwise.
On the other hand, he did not talk about sculpting perfect software people will use for decades to come. I can think of software that succeeded in this respect, and it didn't do it by following his advice. (TeX comes to mind.)
For example, he talked about bloatware, saying it is a good thing. "Features make users' lives better if they use them, and don't usually hurt if they don't." I disagree with this when talking about "hurt" as "making the software more painful to use" instead of "cutting sales". Extra features introduce more bugs and take away from the time programmers could be fixing other bugs. They shouldn't be added until everything else is perfectly solid and possibly not even then.
Really, I think there is a time for to listen to what this guy has to say and a time to completely ignore it. If you're developing a commercial project, his advice has a certain merit. If you're doing something as a hobby, producing a good piece of software you're proud of is much more important than producing a product before your competitors.
Set up NTFS ACLs properly - this includes giving SYSTEM rights to what needs to have it, along with the Administrators group, etc. Users should only have read access.
I've tried this before; it doesn't work out well. The operating system has support for it but applications don't. Intervideo WinDVD, for example, doesn't work on my machine unless I'm at least a Power User. There are plenty of other applications that don't work. That's the problem with adding important security features while maintaining backward compatibilitythe programs available for the system don't get along well with those features.
But a fundamental difference on Unix type systems is that files aren't inherantly executable based simply on their extension, someone can't just save a file from their email and execute it, they need to know at least enough to "chmod u+x" the file which should at least make them think about it.
The execute bit isn't a security feature when you're the owner of the file. [1] It wasn't designed to make you think more about executing files, and it doesn't make you think more about executing files. There's plenty of software that will automatically set the executable bit for you. Ever seen a uuencoded file, for example? They start out with something like this:
The 755 is, of course, the file's access permissions. In symbolic form, u=rwx,og=rx. uudecode honors these. [2] Anyone can execute this file.
[1] Not exactly true. You actually need to be able to write to the directory to modify its access permissions. But generally that's true, and if the file is readable you can always just copy it somewhere else first. Presumably some directory on the filesystem is writable, like your home directory or /tmp.
[2] except where overridden by law or umask(2).
If I don't have a soul, would you care to explain:
Answers follow. Not necessarily true, but certainly no less valid than the religious party line.
Chemical signals in your brain...the same way "lesser animals" do it. (They don't have souls, right?)
See above.
You're dead. You are no more. You cease to exist. There is nothing.
By holding you responsible for your actions, society can control them for the good of the whole. Even if they are "just" caused by a bunch of chemical changes, the carrot and the stick still apply.
Psychosomatic.
I deny them.
A matter of degree, not of kind. We are more intelligent.
I simple cannot accept your blanket denial of the human soul unless you take the time to address these questions.
They've been considered. Accept it.
Alright, each user has an embedded key, and they use that key to place votes.
How do you make sure each user gets exactly one key? As I said in my first post, a per-user key doesn't really help unless you have a method of ensuring that. And if you can ensure that, why can't you directly ensure each person only votes once? It just adds complexity without solving the problem.
This works both ways. Sure, you can find a route that avoids security cameras. But if you're the "bad guys" you now know where you need to install more security cameras.
Umm. No. If you were one of the "bad guys", you'd know where the cameras were already because you put them there. The site isn't giving them any information they don't already have, except possibly where people are clicking if they are monitoring the traffic. And that knowledge is of limited utility, considering that most of the visitors aren't even from New York and are clicking randomly.
where the giant computor [sic]
Has [sic] ever appeared before in a Slashdot article? That amazed me. Granted, it was put there by the submitter, not an editor, but still that's pretty amazing.
With a google-written client communicating with their server, they should be able to come close (or at least make it very difficult to vote twice). There are lots of techniques that could work...dynamically generated keys, encryption, etc.
"If you think encryption will solve your problem, you don't understand encryption and you don't understand your problem." Bruce Schneier
Anything that could be done by your hypothetical client could also be done by a person who has used a debugger on it. It's just not theoretically possible to prevent something like that with an authentication key embedded in the software. Or per-client keys...remember, they have to get the key somehow. How do you restrict it to one key per person? That's back to the original problem.
And what about the firewall code? I don't know about you guys, but I just love spending a large chunk of a weekend learning the new firewall every time a new kernel series arrives. It just wouldn't be the same if it didn't fsck up my firewall scripts...
What are you talking about? You can still run 2.0.x ipfwadm stuff on the latest 2.4 kernel. Likewise 2.2.x ipchains stuff. iptables has backwards compatibility. If you want to use the newer interface, you can. No one's forcing you to.
From linux-2.4.14/Documentation/Configure.help:
pchains (2.2-style) support
CONFIG_IP_NF_COMPAT_IPCHAINS
This option places ipchains (with masquerading and redirection
support) back into the kernel, using the new netfilter
infrastructure. It is not recommended for new installations (see
`Packet filtering'). With this enabled, you should be able to use
the ipchains tool exactly as in 2.2 kernels.
If you want to compile it as a module, say M here and read
Documentation/modules.txt. If unsure, say `N'.
ipfwadm (2.0-style) support
CONFIG_IP_NF_COMPAT_IPFWADM
This option places ipfwadm (with masquerading and redirection
support) back into the kernel, using the new netfilter
infrastructure. It is not recommended for new installations (see
`Packet filtering'). With this enabled, you should be able to use
the ipfwadm tool exactly as in 2.0 kernels.
If you want to compile it as a module, say M here and read
Documentation/modules.txt. If unsure, say `N'.
fs/fs.o: In function `dput':
fs/fs.o(.text+0x118a8): undefined reference to `atomic_dec_and_lock'
make: *** [vmlinux] Error 1
Heh. Deja vu.
Artificial intelligence? A language processor? "Appropriate" keystrokes or mouse movements? This sounds like a fancy way of saying "it guesses a lot".