Ever wonder why hex FF never gets a printable character?
Hex 7F (all holes punched in a 7 bit system) doesn't get a printable character, at least not in ASCII. But FF is a printable character in a lot of character sets - Latin 1 has ÿ in that spot.
Did you need to RTFM in order to get out of a car, ever?
My friend has an old Beatle, that you have to hold the handle when closing the door or it will unlock. He was surprised when I got the first time; it took some of his friends a half an hour to figure it out.
...making the point that companies who don't want to deal with the BSA [suing them for pirating software] can always use Open Source software [Comments in brackets his]
The issue is always not dealing with a BSA suit. What about dealing with BSA threats? What if they want to investiage you because of rumors or lies? What happens if you own 12 copies of Windows and 12 machines, but due to admin carelessness, two machines use the same serial number?
You could also simply pay for the proprietary software that you need to use rather than stealing it.
That won't stop their threats. That won't nessecarily stop them from investigating you. Even if all your software is legal, they can still waste your time going through your records on the matter.
Any comment that starts off by saying "Ha, damn those creationist bastards, they're all stupid and don't believe in science."...
Creationists don't believe in science. They may find it interesting and accurate in some places, but to believe in creationist is believe that science often relys on erroneous, politically biased information to form incorrect conclusions even after extended periods of time. That's they don't believe science reliably works - they don't believe in science.
Yes, it's unfair to describe creationists as stupid. Mostly they're working from postulates that are alien to scientists.
How about common shared principles of well designed software? That's a _start_
One of the main principles of well-designed _anything_ is that humans are fallible. No matter how bright, they will screw up every so often, and some of them screw up frequently. Hence, if and when possible, you must try to reduce the negative effects of erroneous human behavior. Devices should not have exposed dangerous surfaces. Microwaves do not function while open. Incorrect behavior that is common must not caused serious injury, if at all possible.
In light of that, a programming language that permits and encourages fixed length buffers (no strings of unbounded length, for example) and turns a common mistake given fixed length buffers (buffere overflow) into a potential remote exploit of the computer is not well-designed.
For the discourse at hand, it'd be nice to consider performance/scalability figures.
Because performance and scalability is always the first and last issue in any argument. The fact that the Java X server is portable to any system that has Java is irrelevant. Whether it can display 2,400 windows or whether it can create a window for Mozilla in 2.4 microseconds (which Mozilla will take 12.7 seconds to draw in) is.
No, assuming equally competent implementations, the Lisp and Java implementations won't be as fast as the C implementation. A Fortran implemenation will be faster, though, as will an assembly implemenation. All hail Fortran and assmebly.
OK, go on and implement X in Lisp/Java/MSIL/whatever.
Why should I waste my time? The current X works well enough, and nobody's going to jump to my version. There's no reason why it couldn't be done in Lisp or Java, and I believe there's been at least one implementation in each. See http://www.jcraft.com/weirdx/ for the Java implementation.
Or maybe you prefer the GNU C library (which, if you didn't know, powers Lisp/Java/whatever)?
Really? Because I thought, you know, that Lisp dated back to the 1950s and that the GNU C library only runs on Linux and Hurd systems and only dated back to the late 1980s.
The only reason Lisp and Java implementations on Unix tend to depend on the C libraries, is the simple fact that low level wrapping of the kernel and system is painful and unportable work. Hence since a C library is always present on Unix systems, most implementors use the C library as a hardware-wrapping laywer for simplcity. Both Lisp machines and VMS are examples of systems that don't depend on C.
I'll respond to those pinnacles of perfect software one at a time.... I believe that I am as good or better at designing software than all the people you mentioned.
I take it you're having yourself cloned so you can service every programming job in the world? Until then, these people and people like them are going to be coding the software that the world uses, so a programming language should be designed for them.
Why is that programmers believe that all of C and C++'s problems are due to bad programmers?
Because we've worked with them.
So you worked with who wrote the original finger, that the Internet worm exploited? Or the people who wrote the GNU C library, which had a buffer overflow recently? Or Eric S. Raymond, who wrote Fetchmail, which had a buffer overflow in version 5.33? Or the authors of X? Or the authors of MySQL? And they're all bad programmers?
Maybe you're such a god that you've never done that, but if ESR and the glibc authors have such problems, maybe it's more than just bad programmers.
Running into those problems frequently is only a sign of poor design and poor coding, not a language flaw.
Except for the fact that Lisp programers never have memory leaks, have memory corruption only in the rarest of situations and never have to reimplement the linked list.
Why is that programmers believe that all of C and C++'s problems are due to bad programmers?
Comparing an employee::employer relationship
in a bureaucratic organization, even a public corporation, to the relationship of a volunteer
developer to the contact person for a project, just isn't fair.
Why? You work on someone's project and want to continue working on that project, you need to respond to the leader. Why does that change? If I started translating part of GNOME I was responsible for into Ada, and Miguel demanded why, and I didn't answer, likely I would have CVS access revoked and basically be kicked out of the project. Same thing.
Our goal is to stop terrorist attacks on our soil, and stop the resulting deaths and injuries, which we try to follow with minimal civilian deaths. Their goal regards killing civilians as a good thing, and doesn't care how many on their side die.
I mean, even I can see that hitting one air force base isn't going to do anything. They only had those planes -- they had to make them count, and no air force base "counts" that much.
They made them count, but in a way that would piss off the rest of the world. Had they attacked an Air Force base, it would have been a sign that they weren't killing unrelated civilians. A show of strength (and everyone would have known what they could have hit), without the mass death.
If it is intent that matters, then in this case the specific target does not. You, I, and the military leaders of the US, knew that civilians casualties would result from our attacks. Yet we performed them anyway.
But if we had an easy option, we wouldn't. Our goal is not civilian death.
You can hang on the fact that bin Laden attacked a civilian target, but that's a result of the difference in capability, not intent.
Bull. You think that they couldn't have hit a military target with those planes? Nellis Air Force Base could have been hit without much problem. There are many other military installations that are easy targets. Yet they chose a purely civilain target. That's intent.
Our goals are more important than their deaths.
Our goal is to save lives from further civilian attacks. Thousands more people die, versus a few hundred.
Maybe I've misunderstood something, but I don't believe that Miguel works for Stallman, uses any of Stallman's intellecutal or physical assets, or has any real obligation to him.
If you're the project manager for Microsoft John, then you're going to have to listen to Bill Gates. If you're the project manager for GNU Network Object Model Environment, you're going to have to listen to Richard Stallman. It's as simple as that.
How is this situation any different from free software projects using Sun's Java technologies? Isn't this just two sides of the same coin?
Yes and no. Yes, and RMS has never encouraged the use of proprietary Java technology. No, because Apache and Jakarta aren't GNU, where as Gnome is; also nobody has suggested rewriting a major part of the GNU system so that it depends on Java, which is what this suggestion is. Also, Microsoft has proven itself openly hostile to the open source world, where as Sun hasn't.
Something being based on open standards doesn't always mean much. Take a look at the BASIC and Pascal code on the net, and see how much of it runs on ANSI Basic and ISO Pascal. No one forces people or Microsoft to use open standards when they don't want to.
Why? English's sounds aren't too hostile - the dental frictaves and affricates don't compare to the complexity of a tonal language, nor does it make distinctions on unusual properties, besides voiced/unvoiced. It's an Indo-European language, meaning more children learn a language in that language family than any other, and many people in many places already know it, a definite plus. On the downside would be the unusual spelling system, but it's survivable. Grammatically English is fairly normal, at least for an IE language. Where's the crappiness?
I dont know about where you live, but in the unvierse in which I reside, the halting problem has no solution.
Right. Read what I said again.
A completely optimizating compiler solves the halting problem as a side effect. Hence, any real compiler has places where it can optimize better.
Or, more formally, a completely optimizing compiler implies a solution to the halting problem. Since there does not exist a solution to the halting problem, there does not exist a completely optimizing compiler.
But on a dual-boot machine will it kill my other OS?
I've always installed Debian dual-boot and never had a problem. Even on the install, Debian is not a automatically-screw-everything-up system. It drops you right into cfdisk and you can even edit the lilo.conf before lilo runs, IIRC.
Context: GCJ compiles Java source better than bytecode.
All you say is true, however it is merely a flaw in the current GCJ implementation.
A completely optimizating compiler solves the halting problem as a side effect. Hence, any real compiler has places where it can optimize better.
There are only so many compiler writers working on GCJ and many improvements that can made. IMO, the normal mode of GCJ is feeding your own source to it, so their time could be better spent improving the libraries and overall optimization rather then something that's more flashy than useful.
because they make us all realize that our government cannot be trusted not to lie.
If Watergate and Iran-Contra and all the stuff about Social Security haven't taught them that yet, why do you think some website with no intent to defraud or ability to harm will?
While GCJ accepts either Java source or bytecode, it's not clear from the documentation I've read whether or not it first translates source to bytecode or goes straight from source to native.
It goes straight from source to native; for one thing, the source format exposes stuff that allows it to be more heavily optimized by GCC's optimizer than the bytecode format does.
Slashdot Mirror
Ever wonder why hex FF never gets a printable character?
Hex 7F (all holes punched in a 7 bit system) doesn't get a printable character, at least not in ASCII. But FF is a printable character in a lot of character sets - Latin 1 has ÿ in that spot.
Did you need to RTFM in order to get out of a car, ever?
My friend has an old Beatle, that you have to hold the handle when closing the door or it will unlock. He was surprised when I got the first time; it took some of his friends a half an hour to figure it out.
...making the point that companies who don't want to deal with the BSA [suing them for pirating software] can always use Open Source software [Comments in brackets his]
The issue is always not dealing with a BSA suit. What about dealing with BSA threats? What if they want to investiage you because of rumors or lies? What happens if you own 12 copies of Windows and 12 machines, but due to admin carelessness, two machines use the same serial number?
You could also simply pay for the proprietary software that you need to use rather than stealing it.
That won't stop their threats. That won't nessecarily stop them from investigating you. Even if all your software is legal, they can still waste your time going through your records on the matter.
Any comment that starts off by saying "Ha, damn those creationist bastards, they're all stupid and don't believe in science."...
Creationists don't believe in science. They may find it interesting and accurate in some places, but to believe in creationist is believe that science often relys on erroneous, politically biased information to form incorrect conclusions even after extended periods of time. That's they don't believe science reliably works - they don't believe in science.
Yes, it's unfair to describe creationists as stupid. Mostly they're working from postulates that are alien to scientists.
Because performance and scalability is always the first and last issue in any argument.
That's because we, the practicing programmers, live and die by it.
Thanks, "Mel". We must always provide the fastest system for script kiddies to DDOS other people's computers from it; that's only common courtesy.
Frankly, most customers are going to be more interested in security than the last ounce of speed. But it's not about the customers, is it?
this makes your entire point laughable.
You name the system and the languages, and my point is laughable because you don't like my perfectly valid example?
What's relevant is whether it doesn't take forever to open an xterm.
Because it will kill you to wait a couple seconds to open up an xterm. How about a little patience, man?
Use the tool that fits the job, not some high-horse languages because they rock your world
Use the tool that fits the job, not some low-level language because it rocks your world.
How about common shared principles of well designed software? That's a _start_
One of the main principles of well-designed _anything_ is that humans are fallible. No matter how bright, they will screw up every so often, and some of them screw up frequently. Hence, if and when possible, you must try to reduce the negative effects of erroneous human behavior. Devices should not have exposed dangerous surfaces. Microwaves do not function while open. Incorrect behavior that is common must not caused serious injury, if at all possible.
In light of that, a programming language that permits and encourages fixed length buffers (no strings of unbounded length, for example) and turns a common mistake given fixed length buffers (buffere overflow) into a potential remote exploit of the computer is not well-designed.
For the discourse at hand, it'd be nice to consider performance/scalability figures.
Because performance and scalability is always the first and last issue in any argument. The fact that the Java X server is portable to any system that has Java is irrelevant. Whether it can display 2,400 windows or whether it can create a window for Mozilla in 2.4 microseconds (which Mozilla will take 12.7 seconds to draw in) is.
No, assuming equally competent implementations, the Lisp and Java implementations won't be as fast as the C implementation. A Fortran implemenation will be faster, though, as will an assembly implemenation. All hail Fortran and assmebly.
OK, go on and implement X in Lisp/Java/MSIL/whatever.
Why should I waste my time? The current X works well enough, and nobody's going to jump to my version. There's no reason why it couldn't be done in Lisp or Java, and I believe there's been at least one implementation in each. See http://www.jcraft.com/weirdx/ for the Java implementation.
Or maybe you prefer the GNU C library (which, if you didn't know, powers Lisp/Java/whatever)?
Really? Because I thought, you know, that Lisp dated back to the 1950s and that the GNU C library only runs on Linux and Hurd systems and only dated back to the late 1980s.
The only reason Lisp and Java implementations on Unix tend to depend on the C libraries, is the simple fact that low level wrapping of the kernel and system is painful and unportable work. Hence since a C library is always present on Unix systems, most implementors use the C library as a hardware-wrapping laywer for simplcity. Both Lisp machines and VMS are examples of systems that don't depend on C.
I'll respond to those pinnacles of perfect software one at a time. ... I believe that I am as good or better at designing software than all the people you mentioned.
I take it you're having yourself cloned so you can service every programming job in the world? Until then, these people and people like them are going to be coding the software that the world uses, so a programming language should be designed for them.
Why is that programmers believe that all of C and C++'s problems are due to bad programmers?
Because we've worked with them.
So you worked with who wrote the original finger, that the Internet worm exploited? Or the people who wrote the GNU C library, which had a buffer overflow recently? Or Eric S. Raymond, who wrote Fetchmail, which had a buffer overflow in version 5.33? Or the authors of X? Or the authors of MySQL? And they're all bad programmers?
Maybe you're such a god that you've never done that, but if ESR and the glibc authors have such problems, maybe it's more than just bad programmers.
"Programming OCR for Dummies"
It may be narrow, but I've never seen an entry level book to writing an OCR program.
Running into those problems frequently is only a sign of poor design and poor coding, not a language flaw.
Except for the fact that Lisp programers never have memory leaks, have memory corruption only in the rarest of situations and never have to reimplement the linked list.
Why is that programmers believe that all of C and C++'s problems are due to bad programmers?
Comparing an employee::employer relationship
in a bureaucratic organization, even a public corporation, to the relationship of a volunteer
developer to the contact person for a project, just isn't fair.
Why? You work on someone's project and want to continue working on that project, you need to respond to the leader. Why does that change? If I started translating part of GNOME I was responsible for into Ada, and Miguel demanded why, and I didn't answer, likely I would have CVS access revoked and basically be kicked out of the project. Same thing.
Their "goal" isn't to kill civilians, either.
Our goal is to stop terrorist attacks on our soil, and stop the resulting deaths and injuries, which we try to follow with minimal civilian deaths. Their goal regards killing civilians as a good thing, and doesn't care how many on their side die.
I mean, even I can see that hitting one air force base isn't going to do anything. They only had those planes -- they had to make them count, and no air force base "counts" that much.
They made them count, but in a way that would piss off the rest of the world. Had they attacked an Air Force base, it would have been a sign that they weren't killing unrelated civilians. A show of strength (and everyone would have known what they could have hit), without the mass death.
If it is intent that matters, then in this case the specific target does not. You, I, and the military leaders of the US, knew that civilians casualties would result from our attacks. Yet we performed them anyway.
But if we had an easy option, we wouldn't. Our goal is not civilian death.
You can hang on the fact that bin Laden attacked a civilian target, but that's a result of the difference in capability, not intent.
Bull. You think that they couldn't have hit a military target with those planes? Nellis Air Force Base could have been hit without much problem. There are many other military installations that are easy targets. Yet they chose a purely civilain target. That's intent.
Our goals are more important than their deaths.
Our goal is to save lives from further civilian attacks. Thousands more people die, versus a few hundred.
Maybe I've misunderstood something, but I don't believe that Miguel works for Stallman, uses any of Stallman's intellecutal or physical assets, or has any real obligation to him.
If you're the project manager for Microsoft John, then you're going to have to listen to Bill Gates. If you're the project manager for GNU Network Object Model Environment, you're going to have to listen to Richard Stallman. It's as simple as that.
How is this situation any different from free software projects using Sun's Java technologies? Isn't this just two sides of the same coin?
Yes and no. Yes, and RMS has never encouraged the use of proprietary Java technology. No, because Apache and Jakarta aren't GNU, where as Gnome is; also nobody has suggested rewriting a major part of the GNU system so that it depends on Java, which is what this suggestion is. Also, Microsoft has proven itself openly hostile to the open source world, where as Sun hasn't.
Something being based on open standards doesn't always mean much. Take a look at the BASIC and Pascal code on the net, and see how much of it runs on ANSI Basic and ISO Pascal. No one forces people or Microsoft to use open standards when they don't want to.
Why? English's sounds aren't too hostile - the dental frictaves and affricates don't compare to the complexity of a tonal language, nor does it make distinctions on unusual properties, besides voiced/unvoiced. It's an Indo-European language, meaning more children learn a language in that language family than any other, and many people in many places already know it, a definite plus. On the downside would be the unusual spelling system, but it's survivable. Grammatically English is fairly normal, at least for an IE language. Where's the crappiness?
Look at LSB mailing list, and see how many times Alan Cox has posted in just the last month. I count 23, myself.
I dont know about where you live, but in the unvierse in which I reside, the halting problem has no solution.
Right. Read what I said again.
A completely optimizating compiler solves the halting problem as a side effect. Hence, any real compiler has places where it can optimize better.
Or, more formally, a completely optimizing compiler implies a solution to the halting problem. Since there does not exist a solution to the halting problem, there does not exist a completely optimizing compiler.
But on a dual-boot machine will it kill my other OS?
I've always installed Debian dual-boot and never had a problem. Even on the install, Debian is not a automatically-screw-everything-up system. It drops you right into cfdisk and you can even edit the lilo.conf before lilo runs, IIRC.
Context: GCJ compiles Java source better than bytecode.
All you say is true, however it is merely a flaw in the current GCJ implementation.
A completely optimizating compiler solves the halting problem as a side effect. Hence, any real compiler has places where it can optimize better.
There are only so many compiler writers working on GCJ and many improvements that can made. IMO, the normal mode of GCJ is feeding your own source to it, so their time could be better spent improving the libraries and overall optimization rather then something that's more flashy than useful.
Eiffel and Ada are garbage collected out of the box.
Eiffel, yes. Ada is designed so it can easily be garbage collected, but no one supports it in practice unless you're compiling to the Java VM.
because they make us all realize that our government cannot be trusted not to lie.
If Watergate and Iran-Contra and all the stuff about Social Security haven't taught them that yet, why do you think some website with no intent to defraud or ability to harm will?
While GCJ accepts either Java source or bytecode, it's not clear from the documentation I've read whether or not it first translates source to bytecode or goes straight from source to native.
It goes straight from source to native; for one thing, the source format exposes stuff that allows it to be more heavily optimized by GCC's optimizer than the bytecode format does.