ok, make your case objectively?
on
VIM 6.0 is Out
·
· Score: 3, Insightful
Well, that's easy to say, but emacs has had ftp editing for at least 4 years (probably much longer). In my exeperience, it's been more powerful and more customizable than vi has been (and I learned vi first, yes).
Subjective issues aside, what does vi do that emacs doesn't? The only thing I've ever heard that I believe is that it loads faster.
I realize you're kidding, but it doesn't help to be alarmist about the DMCA. It protects only access to a copyrighted work, not anything having to do with "hacking" a proxy server to get out of your school's network.
The DMCA is a bad, scary law and should be overturned, but we won't win that battle by making it out to be something it's not. Educate, rather than knee-jerk.
I don't understand how usefulness is a measure of whether or not something is intellectual property.
It is polite to credit people for their work, but the idea of "owning" information is filled with troubles. This seems to be an especially luke-warm example, if the copying wasn't even verbatim.
I dunno man, that's not so good. Most key escrow schemes allow for the key to split among several agencies (and if they don't, you can use secret-sharing techniques to do it), which would mean that the information would be worthless unless the terrorists or organized criminals compromised *all* of the key stores. Sounds unlikely.
A better argument is that it just doesn't work; it is easy to use standard encryption on top of the escrow scheme, and the government won't be able to read your communication. The best solution I know of is to make sure that most people don't actually use encryption, which, well.. that's how it is now, isn't it?
One thing you could do is to run a web server not written in C (highly susceptible to buffer overflows). Like maybe one written in Java or SML.
Anyway, I actually think apache is pretty secure. It was not hacked because of holes in apache, but in Bugzilla, I believe. I haven't patched my apache in ages, nor seen any (non-alarmist) bugtraq tickets...
That's not true... you don't need a VM or Interpreter. Technologies like Proof Carrying Code and Typed Assembly Language let you verify security properties of real machine code.
All languages that I see in the "open source" community are clones of C designed for scripting. These make life easier for developers (sort of), but don't lead to more robust software.
I am not saying that Microsoft makes good software now or necessarily will soon, but I do know that they've got the plans and people.
Almost all of what most people consider "Linux" or "GNU/Linux" is written in C. And it's pretty bad C code, too; I've seen it...
Well, I guess you're mad, but I'll at least explain myself again if I wasn't clear.
#1, of course not.
#2, probably yes.
#3, I don't know about this -- but I do know that they have hired a lot of smart language designers, because they recognize that current software development practices don't scale. (Do you think they don't know their software is buggy and insecure?) The linux community is doing no such thing, as far as I know; in fact, they are really adamant about using C and Perl.
#4, Well, almost all software is written in C or C++ ("low level languages"), and almost all software sucks.
#5, The project is called "Vault" if you want to look it up.
#6, Efficiency matters, but not the kind of efficiency that C provides. Efficiency in general definitely matters *less* since hardwarwe is getting faster.
By the way, I have nothing to do with Microsoft. I want to see Free software succeed. I'm just worried that the C-programming and lots-of-eyeballs method won't scale. I don't expect you to believe me, but I hope some people learn the same lessons I have and then have the patience to convince people like you. =)
Well, it's clear you don't know what you're talking about, but I might as well point out that there's no reason such high-level safe languages need to be inefficient. In fact, a number of SML implementations are just as fast as C (and much faster than Java and friends). And yet, programs written in SML are 100% buffer-overflow free (other nice features of the language aside!)
Even with slower implementations, the speed of a package like Office is pretty immaterial given the speed at which hardware improves; we were running that on P-133s with 32mb just fine.
MSR has a project called "Vault" which is used to machine certify device drivers (written in a low-level language) for certain safety properties. This is the kind of thing I'm talking about; the linux folks seem to think that "lots" of eyeballs will do the trick... (if indeed we believe that "lots" of people work to improve Linux).
Well, Microsoft actually is working on language technologies which should make their stuff more secure. Expect to see that in a few generations, their stuff, written in C# or SML.NET or whatever, is totally buffer-overflow free.
It's very disappointing to me that the Linux crowd has settled on C, which, as far as I can tell, is the worst language for writing secure apps. (I suppose something with similar properties but less mature compilers might be worse, actually.) I think Linux and friends are going to be left behind as far as stability and security as programs get too big to be verified via the eyeball method.
I think RMS (like other folks with an agenda) will often lose favor among moderates by throwing in jabs like this.
But this remark was fair: It's tongue-in-cheek, of course, but it is reasonable to recall that Bush "won" by the most tiny possible margin imaginible. To give him complete say over the situation (with people urging us to "support him") is, well, a little scary.
That's pretty offensive -- obviously, you don't understand that many people have differing opinions. Personally, I don't support a war because I feel it will escalate the conflict. I would support a firm but civilized extraction of the perpetrators and bringing them to justice (which does not mean execution). Does this mean I don't care about lost friends and family? Of course not. I am just trying to not let emotion make us careless. Though we'd win a contest of might, it'd come at a substantial cost.
Examples of things which would be more like acts of war, since you're stumped:
Invading our country with troops.
Declaring war on us.
A missile strike with someone claiming responsibility.
There have been suggestions that secondary devices were used to actually level the buildings after the crash. If that's the case, then what they're doing makes sense, right?
Yes, indeed it is a crisis. But I hardly think that testifying against unjust legislation is self-centered.
Among other terrible consequences, the government is likely to use this event to leverage more scary laws which limit our freedom, this time for the sake of intelligence groups. It will also not be self-centered to be a voice of reason in these issues as well, so don't give up!
That's crazy. C absolutely can meet or exceed asm speeds, and high level languages often do the same. The problem is that it's nearly impossible to manage optimized assembly code, so there are some kinds of optimizations (ie, software pipelining) which can only realistically be done by a compiler. Even if speed was the primary concern (and this is rather dubious to begin with), it's not clear that asm would always win.
That's true of the 99 competition, and maybe to a lesser degree this last competition. But they ray-tracer really was not geared towards functional languages at all. The language was totally trivial to parse and easy to evaluate; the real test was to see how much you could code without making mistakes in just the one weekend.
Almost all of the C/C++ entries were disqualified for crashing or being buggy!
At least for last year's task, I believe this was a fair language comparison in terms of developer productivity.
Java is inefficient, but not that bad. Emulation is a task particularly well-suited to C, but I think a well-written java program running in a native-code-translating JVM should run well within 2x C code. Graphics depends on the JVM. Java is a memory hog, yes. (That might explain poor performance on old computers.)
Anyway, Java is just the language I threw out to give an idea of the class of languages that I'm talking about; I know it is not the most efficient of the bunch. Check out O'Caml in these benchmarks, for instance:
Slashdot Mirror
Well, that's easy to say, but emacs has had ftp editing for at least 4 years (probably much longer). In my exeperience, it's been more powerful and more customizable than vi has been (and I learned vi first, yes).
Subjective issues aside, what does vi do that emacs doesn't? The only thing I've ever heard that I believe is that it loads faster.
I realize you're kidding, but it doesn't help to be alarmist about the DMCA. It protects only access to a copyrighted work, not anything having to do with "hacking" a proxy server to get out of your school's network.
The DMCA is a bad, scary law and should be overturned, but we won't win that battle by making it out to be something it's not. Educate, rather than knee-jerk.
I don't understand how usefulness is a measure of whether or not something is intellectual property.
It is polite to credit people for their work, but the idea of "owning" information is filled with troubles. This seems to be an especially luke-warm example, if the copying wasn't even verbatim.
It's easy enough to put non-backdoored crypto inside the backdoor one. So, even that argument doesn't work.
I dunno man, that's not so good. Most key escrow schemes allow for the key to split among several agencies (and if they don't, you can use secret-sharing techniques to do it), which would mean that the information would be worthless unless the terrorists or organized criminals compromised *all* of the key stores. Sounds unlikely.
A better argument is that it just doesn't work; it is easy to use standard encryption on top of the escrow scheme, and the government won't be able to read your communication. The best solution I know of is to make sure that most people don't actually use encryption, which, well.. that's how it is now, isn't it?
One thing you could do is to run a web server not written in C (highly susceptible to buffer overflows). Like maybe one written in Java or SML.
Anyway, I actually think apache is pretty secure. It was not hacked because of holes in apache, but in Bugzilla, I believe. I haven't patched my apache in ages, nor seen any (non-alarmist) bugtraq tickets...
Python is totally ad-hoc, too, it was just designed by less insane people.
Actually, Perl and Python are pretty crummy languages; they just happen to be useful.
That's not true... you don't need a VM or Interpreter. Technologies like Proof Carrying Code and Typed Assembly Language let you verify security properties of real machine code.
All languages that I see in the "open source" community are clones of C designed for scripting. These make life easier for developers (sort of), but don't lead to more robust software.
I am not saying that Microsoft makes good software now or necessarily will soon, but I do know that they've got the plans and people.
Almost all of what most people consider "Linux" or "GNU/Linux" is written in C. And it's pretty bad C code, too; I've seen it...
Well, I guess you're mad, but I'll at least explain myself again if I wasn't clear.
#1, of course not.
#2, probably yes.
#3, I don't know about this -- but I do know that they have hired a lot of smart language designers, because they recognize that current software development practices don't scale. (Do you think they don't know their software is buggy and insecure?) The linux community is doing no such thing, as far as I know; in fact, they are really adamant about using C and Perl.
#4, Well, almost all software is written in C or C++ ("low level languages"), and almost all software sucks.
#5, The project is called "Vault" if you want to look it up.
#6, Efficiency matters, but not the kind of efficiency that C provides. Efficiency in general definitely matters *less* since hardwarwe is getting faster.
By the way, I have nothing to do with Microsoft. I want to see Free software succeed. I'm just worried that the C-programming and lots-of-eyeballs method won't scale. I don't expect you to believe me, but I hope some people learn the same lessons I have and then have the patience to convince people like you. =)
Well, it's clear you don't know what you're talking about, but I might as well point out that there's no reason such high-level safe languages need to be inefficient. In fact, a number of SML implementations are just as fast as C (and much faster than Java and friends). And yet, programs written in SML are 100% buffer-overflow free (other nice features of the language aside!)
Even with slower implementations, the speed of a package like Office is pretty immaterial given the speed at which hardware improves; we were running that on P-133s with 32mb just fine.
MSR has a project called "Vault" which is used to machine certify device drivers (written in a low-level language) for certain safety properties. This is the kind of thing I'm talking about; the linux folks seem to think that "lots" of eyeballs will do the trick... (if indeed we believe that "lots" of people work to improve Linux).
Well, Microsoft actually is working on language technologies which should make their stuff more secure. Expect to see that in a few generations, their stuff, written in C# or SML.NET or whatever, is totally buffer-overflow free.
It's very disappointing to me that the Linux crowd has settled on C, which, as far as I can tell, is the worst language for writing secure apps. (I suppose something with similar properties but less mature compilers might be worse, actually.) I think Linux and friends are going to be left behind as far as stability and security as programs get too big to be verified via the eyeball method.
I like mozilla a lot, actually.
I think RMS (like other folks with an agenda) will often lose favor among moderates by throwing in jabs like this.
But this remark was fair: It's tongue-in-cheek, of course, but it is reasonable to recall that Bush "won" by the most tiny possible margin imaginible. To give him complete say over the situation (with people urging us to "support him") is, well, a little scary.
Voltron!
That's pretty offensive -- obviously, you don't understand that many people have differing opinions. Personally, I don't support a war because I feel it will escalate the conflict. I would support a firm but civilized extraction of the perpetrators and bringing them to justice (which does not mean execution). Does this mean I don't care about lost friends and family? Of course not. I am just trying to not let emotion make us careless. Though we'd win a contest of might, it'd come at a substantial cost.
Examples of things which would be more like acts of war, since you're stumped:
Invading our country with troops.
Declaring war on us.
A missile strike with someone claiming responsibility.
There have been suggestions that secondary devices were used to actually level the buildings after the crash. If that's the case, then what they're doing makes sense, right?
What's up with this 'support the president' stuff? Why does he need our support?
I do not support going to war.
Yes, indeed it is a crisis. But I hardly think that testifying against unjust legislation is self-centered.
Among other terrible consequences, the government is likely to use this event to leverage more scary laws which limit our freedom, this time for the sake of intelligence groups. It will also not be self-centered to be a voice of reason in these issues as well, so don't give up!
That's crazy. C absolutely can meet or exceed asm speeds, and high level languages often do the same. The problem is that it's nearly impossible to manage optimized assembly code, so there are some kinds of optimizations (ie, software pipelining) which can only realistically be done by a compiler. Even if speed was the primary concern (and this is rather dubious to begin with), it's not clear that asm would always win.
9th place baby!!!
SML rulez!!!!!@!!!
I know it didn't get best results in all of the tests; our program was the best for one of them. =)
Maybe it was disqualified for crashing or producing incorrect results?
That's true of the 99 competition, and maybe to a lesser degree this last competition. But they ray-tracer really was not geared towards functional languages at all. The language was totally trivial to parse and easy to evaluate; the real test was to see how much you could code without making mistakes in just the one weekend.
Almost all of the C/C++ entries were disqualified for crashing or being buggy!
At least for last year's task, I believe this was a fair language comparison in terms of developer productivity.
(Of course, I am biased. 9th place, woo hoo!)
Java is inefficient, but not that bad. Emulation is a task particularly well-suited to C, but I think a well-written java program running in a native-code-translating JVM should run well within 2x C code. Graphics depends on the JVM. Java is a memory hog, yes. (That might explain poor performance on old computers.)
l
Anyway, Java is just the language I threw out to give an idea of the class of languages that I'm talking about; I know it is not the most efficient of the bunch. Check out O'Caml in these benchmarks, for instance:
http://www.bagley.org/~doug/shootout/craps.shtm