Look pick one: either the global warming is the issue, or it isn't.
If so, then temperature is the appropriate measurement, but then the fact that it's not an unusual change makes the world-is-ending business a little questionable, and the anthropogenic hypothesis hard to support. (And there are these odd side things, like the fact that Mars seems to be experiencing global warming too. Two solar-powered golf carts aren't doing that.)
If not, then what is the issue we're trying to talk about?
I'm equally sure that the northern Europeans, whose civilization reached hitherto unknown heights in both climatic optima (ever hear of the Renaissance?), weren't disappointed at all.
Since we know that this has happened before industrialization and gigahuman populations, the argument that the warming is anthropogenic, and therefore preventable, is therefore weakened.
Since we know it has happened before without global catastrophe, the panicky "ten seconds from midnight" implications of the article are probably overwroght, and possibly disingenuous and politically based.
... and have been for 30 years, is all I'm saying.
(But, as far as "wery educated Indians" goes, look up who first got certified as CMM level 5. India isn't getting the work they get just because they work cheap.)
Depending on what you know and what you want to do, you can find that kind of job. If you want something that looks like job security, learn COBOL and mainframe programming, and/or AS/400. Those jobs will never die. Move to places like Columbus OH or Hartford CT, work for something like an insurance company. Exciting, no. But you can spend a lot of time writing code.
If you want to keep doing interesting things, you have to learn new skills. Java/JUnit/XP; C#/.Net; I see a fair lot of ads for Python. And you'll have to keep learning new skills, and looking for new things to do; that's just the way it is.
Get the hell out of Silicon Valley. It's too expensive, the market's too overloaded, and the traditional high-tech business is consolidating: you've got no reason to be confident that you can avoid layoffs (viz. Peoplesoft.) The Santa Clara Valley is used up. (Ergo, find someplace that isn't used up.)
I'm guessng you don't have a security clearance, so if you can find a way, get one. The higher the better. Those are big-time assets. AND you can tell all your friends you're a spy.
Am I missing something here and there really are reasons to not use a VCS?"
No, you're not. But I tell you what --- I've been consulting for, oh, close to 20 years, and I've seen probably in excess of 200 companies, and I'd hate to tell you how many of them had no version control. Hell, I'd hate to tell you how many of them had no code backup, and you'd be amazed how many companies --- big companies --- have web applications in particular that live on someone's desktop and couldn't be reconstructed if that person was run over by a truck without reimplementing.
I'd hate to tell you, but I'll say, if it's as high as 50 percent who have version contral, then that means it's about doubled in the last few years.
It's been a long time and God knows it was primarily memorable for things like the supersonic motorcycle, but I *think* they actually used the sonoluminescence hack as the explanation for making the fusion work.
If it's not American Airline's fault, then why did the requirement to provide this information [*] when an American Airline's employee discovered that he was an American Airline's Platinum frequent flyer?
Because American Airlines, like all the other airlines, get a special arrangement with TSA for their high-volume customers?
Which is why everyone's high-milage FF customers get to use that fast line at the security gate?
If you've actually been in an airport in the last four years, you might have inferred this all by yourself.
* (I'm kind of assuming that you intended a phrase like "go away" to go into that sentence somewhere. If not, could you re-phrase the sentence with all the appropriate subjects and objects and resumbmit?)
... and Cory's being more of a dolt than I would have expected if he really thinks it was.
That being said, I had that experience entering the US from Canada on a US passport in about 1996. Missed my flight in Pearson airport (Toronto) while I was going through the interminable questions ---
Q. Where are you going? A. North Carolina
Q. Why are you going there? A. I live there.
Q. What do you plan to do there? A. I'm a computer consultant.
Q. Do you have work when you arrive? A. Yes. That's why I live there.
Q. How long do you plan to be in the US? A. Until I leave again. I live there.
Q. Where do you plan to stay? A. At my home. The one where it says "Home Address." In Durhan NC.... after about 40 minutes of this, I insisted on seeing a supervisor, saying "Look, dammit, I'm an American citizen. I was born in the USA. My parents were born in the USA. Hell, I'm a quarter blood Choctaw Indian -- I'm a Native American native american!"
The demand that I speak with a supervisor broke the log jam; they let me through.
My grandfather, many years ago when I was eight or nine --- which is to say many years ago --- asked me this question: "Do you know why a dog will lie on a sunny porch licking his own balls?"
. Terragen, a freeware program that converts the basic brightness data in aerial pictures into a topographical map, to generate the ground-level vista was used.
Posting to Slashdot did not know, Master Yoda was.
Look it up, Brainiac. That's how it's spelled. You know, the noun conjugation of condescend [m-w.com].
's', not 't'. 'Condescension'. Try your spelling at M-W and see what you get. And 'lagniappe' is idiomatic for "thrown in for free", like the 13th donut in a baker's dozen.
As any experienced C programmer knows, there were virtually no standards-compiliant C compilers before c89.
Look, you dolt, there were no standards-compliant C compilers before C89 because THERE WERE NO STANDARDS. Get it?
In the mean time, let's see: you've asserted that comments are removed by a C compiler's tokenizer, not the preprocessor, which is easily demonstrated wrong, and you've asserted that BY THE STANDARD text segments of C object files are merged into the same text segment at load time, but when challenged to cite the spec, you're complained about me pointing out your spelling mistakes.
Nice flaming with you, son, but there's a blow-out rule, and as we're running about 427-0 right now, I think maybe you should just change your name and hope to live it down.
I think part of the point is that including all the code through the preprocessor is an old, dumb way of doing it. Not a new trick.
By the way, another one of those "old fart" comments is coming, brace yourselves:... actually, you'll almost always find that modules in a 7 megaSLOC program are much the same size as in a 50KSLOC project, and if that means you have 7000 modules, then you do. A "module" has to be of a size to be comprehended by one person.
A really big forest doesn't generally have really big trees.
If we're talking about the business about the TEXT and DATA segments, since standards are your big deal here, can you cite anything in the ANSI standard that says the TEXT segments must be combined as you suggest?
And if not, wouldn't you think that it would be foolish to make assumptions about how it would behave?
Some years ago, I did some work for DARPA on trustworthy C compilation, and one of the things we discovered was that there are LOTS of places where you can't make assumptions about the standard just because it was what you were used to.
As far as "significant faults in standards compliance", look at the dates I'm talking about, son. C89 was a 1989 standard, and I'm talking about 1979. Those compilers didn't fail to comply with the standard in 1979-84 because it would be five years before the standard was promulgated.
UML: I have never known anyone to fight "Rational Rose" and actually come out a winner, but some pretty flow charts WITH AN ACCOMPANIYING LEDGENED can help express an idea.
Oh, absolutely. Martin Fowler's got a fine book on UML Distilled. If you stick to the notations that are in the endpapers of that book, you'll have a very useful basic notation.
I think you're having a little difficulty with a couple of real-world issues here (not to mention the one about "keeping a civil tongue", but this is slashdot.)
First of all, you're losing context: we're not talking about an actual C++ program compiling on a current-generation compiler, we're talking about someone's weird-ass C-like language compiled on some very peculiar implementation. We're not sure what the peculiarities are, other than knowing that startup time for the compiler must be awe-inspiringly bad.
Given that, and given the arcane nature of optimization in production compilers, the one thing we can bet on is that the optimization is not as smart as some. You're making an unwarranted assumption there, not to mention your assumptions about the implementation of the compiler, linker, etc.
Second, when someone tells me "that doesn't matter because no 'good' programmer would do it that way", I know I'm talking to someone who hasn't got a lot of actual real-world experience. If the specific semantics of the language allow something, some dfamn fool's gonna try it.
Sometimes the damn fool will be a real smart one: for example, Steve Bourne's version of sh was written with a collection of #define macros to make C's syntax look like Algol60. In other words, the code looked like
IF(x == NULL) THEN
y = 2* z;
ELSE
y = z * z;
FI
Steve is no dummy, and I'm sure it seemed like A Good Idea At The Time.
In this case, you're arguing that only amateur C programmers use file scope. That's a clear sign of someone who thinks C++ is C. You're completely right that there's no reason to use file scope in a C++ program: you've got everything from static const to namespaces to eliminate the need (and nearly eliminate the need for the preprocessor as well.)
IN straight C, however, file and static are one of the few things you have to manage the namespace in larger programs. Whether or not your particular religious version of "good coding practices" allows for them is unimportant, as there are undoubtedly legacies of code using that in the tens of billions of lines.
And as far as using file scope making me an "amateur", from your tone and your arguments, I suspect I'd been a professional programmer for 25 years the day you were born.
Son, when you've been doing this for 40 years, you'll have trouble keeping chronologies straight too.
The specific example of a system in which I think this was a problem was Bell Labs UNIX running on a PDP 11/45 in about 1979. But for all I know, it could have been anything from VMS 3 VAX C (probably not, actually, since the memory architecture didn't work that way) to Tiny C on a Cromemco S100 bus system.
Oh, God, RTTI and templates are such a pain in the ass. I'm not quite clear on what your argument is, but it is possible to do templates without doing textual replication and substitution, even in the presence of RTTI. This one's even pretty easy, with a hidden "class" object class method.
"Thanks, half-pint. You've just saved me a lot of investigative work."
Yeah, that's a good point. I knew there was something, but I couldn't think of an example.
Re:The immortal advice of Rocket J Squirrel
on
Abandoning Header Files?
·
· Score: 2, Interesting
Uh, which loader and which linker, and for that matter which operating system etc? I suspect you're right about linux, and it'd be lovely to believe that there aren't any systems (other than in hobbyists' basements) that don't do it that way any more, but with the information we've got, we can't exclude the possibility.
No, it's a limitation of the C language that was imposed by the old compilers' environments'.
Remember that in The Old Days, the C compiler had to run each phase in less than 64K 16-bit words, text and data. Separate compilation allowed the separate segments to be compiled within those limits, and then the.o files could be linked as a last step. This also allowed for such niceties as overlaying loaders etc.
#include was a mechanism to eliminate code duplication; instead of recoding the interfaces in each module, you could use the preprocessor to let you bring in the definitions once.
Languages like Pascal, that depended on compiling everything all in one lump had serious memory limitations that limited the size of the programs badly.
Slashdot Mirror
Look pick one: either the global warming is the issue, or it isn't.
If so, then temperature is the appropriate measurement, but then the fact that it's not an unusual change makes the world-is-ending business a little questionable, and the anthropogenic hypothesis hard to support. (And there are these odd side things, like the fact that Mars seems to be experiencing global warming too. Two solar-powered golf carts aren't doing that.)
If not, then what is the issue we're trying to talk about?
I'm equally sure that the northern Europeans, whose civilization reached hitherto unknown heights in both climatic optima (ever hear of the Renaissance?), weren't disappointed at all.
Since we know that this has happened before industrialization and gigahuman populations, the argument that the warming is anthropogenic, and therefore preventable, is therefore weakened.
Since we know it has happened before without global catastrophe, the panicky "ten seconds from midnight" implications of the article are probably overwroght, and possibly disingenuous and politically based.
the world ended in the 1100s and the 1500s when the temperatures were that high before, so this is just academic.
... and have been for 30 years, is all I'm saying.
(But, as far as "wery educated Indians" goes, look up who first got certified as CMM level 5. India isn't getting the work they get just because they work cheap.)
Deathmarches, poor project management and PHBs are no more the norm now than they were 30 years ago.
Which is to say "completely so" and it's really kind of depressing, you know?
Depending on what you know and what you want to do, you can find that kind of job. If you want something that looks like job security, learn COBOL and mainframe programming, and/or AS/400. Those jobs will never die. Move to places like Columbus OH or Hartford CT, work for something like an insurance company. Exciting, no. But you can spend a lot of time writing code.
If you want to keep doing interesting things, you have to learn new skills. Java/JUnit/XP; C#/.Net; I see a fair lot of ads for Python. And you'll have to keep learning new skills, and looking for new things to do; that's just the way it is.
Get the hell out of Silicon Valley. It's too expensive, the market's too overloaded, and the traditional high-tech business is consolidating: you've got no reason to be confident that you can avoid layoffs (viz. Peoplesoft.) The Santa Clara Valley is used up. (Ergo, find someplace that isn't used up.)
I'm guessng you don't have a security clearance, so if you can find a way, get one. The higher the better. Those are big-time assets. AND you can tell all your friends you're a spy.
Am I missing something here and there really are reasons to not use a VCS?"
No, you're not. But I tell you what --- I've been consulting for, oh, close to 20 years, and I've seen probably in excess of 200 companies, and I'd hate to tell you how many of them had no version control. Hell, I'd hate to tell you how many of them had no code backup, and you'd be amazed how many companies --- big companies --- have web applications in particular that live on someone's desktop and couldn't be reconstructed if that person was run over by a truck without reimplementing.
I'd hate to tell you, but I'll say, if it's as high as 50 percent who have version contral, then that means it's about doubled in the last few years.
It's been a long time and God knows it was primarily memorable for things like the supersonic motorcycle, but I *think* they actually used the sonoluminescence hack as the explanation for making the fusion work.
... you know, the movie where he outruns the shock wave on a motorcycle?
Val Kilmer built a giant death-ray laser.
If it's not American Airline's fault, then why did the requirement to provide this information [*] when an American Airline's employee discovered that he was an American Airline's Platinum frequent flyer?
Because American Airlines, like all the other airlines, get a special arrangement with TSA for their high-volume customers?
Which is why everyone's high-milage FF customers get to use that fast line at the security gate?
If you've actually been in an airport in the last four years, you might have inferred this all by yourself.
* (I'm kind of assuming that you intended a phrase like "go away" to go into that sentence somewhere. If not, could you re-phrase the sentence with all the appropriate subjects and objects and resumbmit?)
... and Cory's being more of a dolt than I would have expected if he really thinks it was.
... after about 40 minutes of this, I insisted on seeing a supervisor, saying "Look, dammit, I'm an American citizen. I was born in the USA. My parents were born in the USA. Hell, I'm a quarter blood Choctaw Indian -- I'm a Native American native american!"
That being said, I had that experience entering the US from Canada on a US passport in about 1996. Missed my flight in Pearson airport (Toronto) while I was going through the interminable questions ---
Q. Where are you going?
A. North Carolina
Q. Why are you going there?
A. I live there.
Q. What do you plan to do there?
A. I'm a computer consultant.
Q. Do you have work when you arrive?
A. Yes. That's why I live there.
Q. How long do you plan to be in the US?
A. Until I leave again. I live there.
Q. Where do you plan to stay?
A. At my home. The one where it says "Home Address." In Durhan NC.
The demand that I speak with a supervisor broke the log jam; they let me through.
My grandfather, many years ago when I was eight or nine --- which is to say many years ago --- asked me this question: "Do you know why a dog will lie on a sunny porch licking his own balls?"
The answer, of course, is "because he can."
. Terragen, a freeware program that converts the basic brightness data in aerial pictures into a topographical map, to generate the ground-level vista was used.
Posting to Slashdot did not know, Master Yoda was.
condescention[sic]
Look it up, Brainiac. That's how it's spelled. You know, the noun conjugation of condescend [m-w.com].
's', not 't'. 'Condescension'. Try your spelling at M-W and see what you get. And 'lagniappe' is idiomatic for "thrown in for free", like the 13th donut in a baker's dozen.
As any experienced C programmer knows, there were virtually no standards-compiliant C compilers before c89.
Look, you dolt, there were no standards-compliant C compilers before C89 because THERE WERE NO STANDARDS. Get it?
In the mean time, let's see: you've asserted that comments are removed by a C compiler's tokenizer, not the preprocessor, which is easily demonstrated wrong, and you've asserted that BY THE STANDARD text segments of C object files are merged into the same text segment at load time, but when challenged to cite the spec, you're complained about me pointing out your spelling mistakes.
Nice flaming with you, son, but there's a blow-out rule, and as we're running about 427-0 right now, I think maybe you should just change your name and hope to live it down.
I think part of the point is that including all the code through the preprocessor is an old, dumb way of doing it. Not a new trick.
... actually, you'll almost always find that modules in a 7 megaSLOC program are much the same size as in a 50KSLOC project, and if that means you have 7000 modules, then you do. A "module" has to be of a size to be comprehended by one person.
By the way, another one of those "old fart" comments is coming, brace yourselves:
A really big forest doesn't generally have really big trees.
That's okay, the condescention[sic] was lagniappe.
/* foo.c -- a comment */
/* another comment */ ;
;
So which behavior is it we're talking about? The business about the CPP phase stripping comments is certainly correct, viz:
1000 $ cat foo.c
#define ACONST 42
static int foo(){
return ACONST
}
int bar(){
return foo();
}
1001 $ gcc -E foo.c
# 1 "foo.c"
# 1 ""
# 1 ""
# 1 "foo.c"
static int foo(){
return 42
}
int bar(){
return foo();
}
1002 $
If we're talking about the business about the TEXT and DATA segments, since standards are your big deal here, can you cite anything in the ANSI standard that says the TEXT segments must be combined as you suggest?
And if not, wouldn't you think that it would be foolish to make assumptions about how it would behave?
Some years ago, I did some work for DARPA on trustworthy C compilation, and one of the things we discovered was that there are LOTS of places where you can't make assumptions about the standard just because it was what you were used to.
As far as "significant faults in standards compliance", look at the dates I'm talking about, son. C89 was a 1989 standard, and I'm talking about 1979. Those compilers didn't fail to comply with the standard in 1979-84 because it would be five years before the standard was promulgated.
In this business, 1979 << 1989.
UML: I have never known anyone to fight "Rational Rose" and actually come out a winner, but some pretty flow charts WITH AN ACCOMPANIYING LEDGENED can help express an idea.
Oh, absolutely. Martin Fowler's got a fine book on UML Distilled. If you stick to the notations that are in the endpapers of that book, you'll have a very useful basic notation.
It's the arrow lawyers who spoil it for everyone.
I think you're having a little difficulty with a couple of real-world issues here (not to mention the one about "keeping a civil tongue", but this is slashdot.)
;
First of all, you're losing context: we're not talking about an actual C++ program compiling on a current-generation compiler, we're talking about someone's weird-ass C-like language compiled on some very peculiar implementation. We're not sure what the peculiarities are, other than knowing that startup time for the compiler must be awe-inspiringly bad.
Given that, and given the arcane nature of optimization in production compilers, the one thing we can bet on is that the optimization is not as smart as some. You're making an unwarranted assumption there, not to mention your assumptions about the implementation of the compiler, linker, etc.
Second, when someone tells me "that doesn't matter because no 'good' programmer would do it that way", I know I'm talking to someone who hasn't got a lot of actual real-world experience. If the specific semantics of the language allow something, some dfamn fool's gonna try it.
Sometimes the damn fool will be a real smart one: for example, Steve Bourne's version of sh was written with a collection of #define macros to make C's syntax look like Algol60. In other words, the code looked like
IF(x == NULL) THEN
y = 2* z;
ELSE
y = z * z
FI
Steve is no dummy, and I'm sure it seemed like A Good Idea At The Time.
In this case, you're arguing that only amateur C programmers use file scope. That's a clear sign of someone who thinks C++ is C. You're completely right that there's no reason to use file scope in a C++ program: you've got everything from static const to namespaces to eliminate the need (and nearly eliminate the need for the preprocessor as well.)
IN straight C, however, file and static are one of the few things you have to manage the namespace in larger programs. Whether or not your particular religious version of "good coding practices" allows for them is unimportant, as there are undoubtedly legacies of code using that in the tens of billions of lines.
And as far as using file scope making me an "amateur", from your tone and your arguments, I suspect I'd been a professional programmer for 25 years the day you were born.
OKay, this one made me laugh. Virtual Mod FUNNY.
Remember that Bjarne has said from the first that a lot of C++ was designed to get away from the damn preprocessor.
Son, when you've been doing this for 40 years, you'll have trouble keeping chronologies straight too.
The specific example of a system in which I think this was a problem was Bell Labs UNIX running on a PDP 11/45 in about 1979. But for all I know, it could have been anything from VMS 3 VAX C (probably not, actually, since the memory architecture didn't work that way) to Tiny C on a Cromemco S100 bus system.
Implementations differ, as usual, but I promise you that many C preprocessors do this. Um, I know for sure that it did in 11/70 UNIX C.
But what do you want for stone knives and bearskins.
Oh, God, RTTI and templates are such a pain in the ass. I'm not quite clear on what your argument is, but it is possible to do templates without doing textual replication and substitution, even in the presence of RTTI. This one's even pretty easy, with a hidden "class" object class method.
But this would be one reason why RTTI is evil.
"Thanks, half-pint. You've just saved me a lot of investigative work."
Yeah, that's a good point. I knew there was something, but I couldn't think of an example.
Uh, which loader and which linker, and for that matter which operating system etc? I suspect you're right about linux, and it'd be lovely to believe that there aren't any systems (other than in hobbyists' basements) that don't do it that way any more, but with the information we've got, we can't exclude the possibility.
Remember that in The Old Days, the C compiler had to run each phase in less than 64K 16-bit words, text and data. Separate compilation allowed the separate segments to be compiled within those limits, and then the .o files could be linked as a last step. This also allowed for such niceties as overlaying loaders etc.
#include was a mechanism to eliminate code duplication; instead of recoding the interfaces in each module, you could use the preprocessor to let you bring in the definitions once.
Languages like Pascal, that depended on compiling everything all in one lump had serious memory limitations that limited the size of the programs badly.