Ever hear of arguing a possibility and not a chosen solution?
People like to assume that int pointers are always the same size, thus mixing them with void pointers. So you get people doing stupid things like cross assigning them since they fit into each other.
Under languages that don't provide the use of pointers or allow mix-matched assignment doesn't suffer this.
That's a dump arguement. Java has the same problem of having to detect pointer size and the presence of certain function calls, except its one level down, inside the libraries.
Er? You dont' have explicit pointers in Java. You don't have to worry about pointer sizes. The implementator of the java vm on choice machine has to worry about that.
C99 compliancy is still a problem for all but one (Commercial and expensive) compiler. C89 and ANSI C++ is standardised and supported by all of them. However, C++ compiler vendors have a static, well defined target to aim for. GCC is aiming for it and will achieve full C99 compliancy within the next few versions, for example.
Problem is, now you have split versions of C. You dont' have that problem with java. Foreshadowing your 1.3->1.4 reference, yeah, they can change the spec, but when you write a java 1.3 app, it's consistent across all java 1.3 jvms. Now we have C89, 99 K&R etc.. cross compilation becomes harder since every compiler does something different regardless of what it implements.
Which version of Java are you are on at the moment? 1.4? Sun can and do change the spec of Java at any time, which creates a moving target for any idependent Java compiler vendor.
I rather my vendor follow a standard than me having to sort it out. It makes things less difficult. It also makes things easier to go to any box running jdk 1.4 and run 1.4 code. Or i can run 1.3 code on that 1.4 box. Heck, run 1.0 code on it too.
No, it is C++ with vendor extensions and you have shot yourself in the foot. The Linux kernel for example is written in GNU C, which until recently could only be compiled with GCC (Intel introduced enough GCC compatability that ICC 7 can compile it too)
Nono.. i mean what if I write a c++ compiler and leave out...say.. template pointers. Everything else works. Can I say it's C++? What version?
Not since 1989 when C was standardised! If you write in ANSI C89 then your code is portable and can be compiled on any C89 (Or C99) compiler. If you use platform specific functions then no, you're code will not be portable. The same problem affects Java code that uses JNI.
Even AFTER 89, there was still that problem.
Just because you do not know what they're for, does not make them crap. Why does Java have the XML based Ant?
I meant crap as in stuff.. junk. Not as in pieces of shit:) . You need ant/automake/autoconf just to detect (autoconf specifically) what type of c compatability you have. Easy one. pointer sizes. On a 64 bit machine or 32 bit machine, they are different.. and if you write c in an unsafe method, you can easily screw up 64->32 bit assignments.
Yet FreeBSD has a C89 compiler. Which just proves my point that Java portability is largly an illusion which can only be fulfiled in a small subset of cases. Oddly enough this is also the case for other languages such as C, which gives Java little edge over these languages.
But while the core very core things may be portable in some parts, we have so many extensions to obfuscate the language. You can't really extend java past JNI, or perl past XS. And even then, you aren't extending the language itself further. You've just added new functions/methods.
Heh, gcc doesn't govern portability. For instance... take networking. Bsd, Linux, windows.. they all implement it differently. Maybe recently it changed, but at least 4 years ago, they were different.
When you need things like autoconf to detect the size of a pointer (32bit vs 64) or if flock works, it's not portable. sorry charlie.
What are you babling on about? Seriously, did you even stop to think about what you were saying before you wrote that? If ANSI has no clout, why does every compiler vendor strive for ANSI C/C++ compliance? Why is that usually one of the first things any vendor lists as a marketing point? It's because vendors and coders alike see ANSI compliance as important.
Heh, what are YOU babling about. gcc isn't ansi c++ compliant. Heck, Borland/w OWL was never compliant. People just used the "best" one out there.
What happens when someone else says their version of Java is Java, too? Sun telling everyone what Java is didn't help them to much when Microsoft released their own version of Java, did it? Oddly enough though, Microsoft Visual C++ sticks to the ANSI standards and is compatable with almost anyone elses C++ compiler.
Funny thing is, if it doesn't stick to ANSI C++, is it C++? Is it as compatable? So now we have one supposedly ansi C++ compiler, and a bunch of supposdly non compliant. But you know what, people still call it C++.
Let's take a worse one. HTML. It's not a programming language, but everyone calls HTML, HTML. Every brower likes to claim support. but in the end, everyone extended it.
And btw, there are multiple "c"'s. There's J&R, (someone & R, i forget), borland, turbo, MS and ansi c. You write in ONE c, you can't cross compile as easily. Thus you have automake, autoconf and crap.
Sun just sued MS for distributing something called java which was incompatable. They also sued for not providing an implementation of Java. Heck, freebsd is suffering from this. They can't distribute "java" because they didn't pass the little certification of code to be called java. They have to distribute patch sets.
C isn't as portable as perl, python, php or java. C and C++ suffer from either incomplete or random interpretations of the specifications.
perfect example is TurboC and that stupid conio.h file. I came to the unix worled after learning TurboC and was confused for months until i learned what ncurses was.
Class action lawsuit against sega for what? It's not like they lied about what was there, did they? Maybe I'm wrong on that part. But if someone wants to go to a pay system vs free, that's their choice.
Re:Why do people do this?
on
I, Spammer
·
· Score: 1
But you realize, snailmail spammers are affected the same way. N% of people don't reply to or are affected by spam. It's out of what how uneven it is, no? So adding a 100th of a percent to the 99% prolly wouldn't hurt.
But then again, there's lies, larger ones and statistics.
There are a LOT of devices that are jack of all trades, master of none.
Look at the iPod. It's not a great PDA. It never will be. Apple won't persue it very hard.
Look at those annoying scanner-fax-printer things. Great if you dont' care much about poor performance of one or even two of the functions. It's the same reason why the ones that really do everything right tend to be very expensive.
Counter-point, the tivo-directv combo box. They were meant for each other. It's rare when devices should combine. Radio's and music players are the same way. Look at receivers and walkmen. Surprised I don't see more cd players/w radios.
BTW, the GBA functionality is almost always optional. Look at wind waker. It's simply a bonus to have the linkup. Same/w sonic.
Uh.. no.. you can't predict that something will be fast enough to catch up on lost time and be fast enough to continue working.
If something takes in the long run 10 years from now, or 9 years from 2 years from now.... well you do the math. But no one knows what to expect. So making the best out of what you have is always good.
A few % increase can be the difference betwee 9 or 10 years, which is even more visible in parallel processing. If one machine is one whole minute faster a week, and you have 1000 machines. you just gained a few hours. Remember, it's all in perspective.
As for a total waste of his time? Who cares? He was bored. We all have our hobbies. As long as he makes a living, and he comes out having had a good time, learning or what have you, be happy he did it. At least SOMEONE did.
Just a note on 1. It's not a panacea. You always get that one dork who decides to cc you on a forum post or something silly. But it does slow it down dramatically.
Slashdot Mirror
Tnx
Ever hear of arguing a possibility and not a chosen solution?
People like to assume that int pointers are always the same size, thus mixing them with void pointers. So you get people doing stupid things like cross assigning them since they fit into each other.
Under languages that don't provide the use of pointers or allow mix-matched assignment doesn't suffer this.
--
Turns out you are right on the flock issue.
Er? You dont' have explicit pointers in Java. You don't have to worry about pointer sizes. The implementator of the java vm on choice machine has to worry about that.
You ARE old. :)
Given up? :)
Problem is, now you have split versions of C. You dont' have that problem with java. Foreshadowing your 1.3->1.4 reference, yeah, they can change the spec, but when you write a java 1.3 app, it's consistent across all java 1.3 jvms. Now we have C89, 99 K&R etc.. cross compilation becomes harder since every compiler does something different regardless of what it implements.
I rather my vendor follow a standard than me having to sort it out. It makes things less difficult. It also makes things easier to go to any box running jdk 1.4 and run 1.4 code. Or i can run 1.3 code on that 1.4 box. Heck, run 1.0 code on it too.
Nono.. i mean what if I write a c++ compiler and leave out...say.. template pointers. Everything else works. Can I say it's C++? What version?
Even AFTER 89, there was still that problem.
I meant crap as in stuff.. junk. Not as in pieces of shit
But while the core very core things may be portable in some parts, we have so many extensions to obfuscate the language. You can't really extend java past JNI, or perl past XS. And even then, you aren't extending the language itself further. You've just added new functions/methods.
I notice you dodged the whole pointer thing. And flock...
Heh, gcc doesn't govern portability. For instance... take networking. Bsd, Linux, windows.. they all implement it differently. Maybe recently it changed, but at least 4 years ago, they were different.
When you need things like autoconf to detect the size of a pointer (32bit vs 64) or if flock works, it's not portable. sorry charlie.
I'd try and equate it to something, but analogies suck.
You pay the cable company for the connection to them and using them.
You inadvertantly pay for commercials by watching them, which fund shows as well.
Now mind you, money doesn't always flow like this, i.e. PBS or PPV easily, but that's how things work.
Heh, what are YOU babling about. gcc isn't ansi c++ compliant. Heck, Borland
Funny thing is, if it doesn't stick to ANSI C++, is it C++? Is it as compatable? So now we have one supposedly ansi C++ compiler, and a bunch of supposdly non compliant. But you know what, people still call it C++.
Let's take a worse one. HTML. It's not a programming language, but everyone calls HTML, HTML. Every brower likes to claim support. but in the end, everyone extended it.
And btw, there are multiple "c"'s. There's J&R, (someone & R, i forget), borland, turbo, MS and ansi c. You write in ONE c, you can't cross compile as easily. Thus you have automake, autoconf and crap.
Sun just sued MS for distributing something called java which was incompatable. They also sued for not providing an implementation of Java. Heck, freebsd is suffering from this. They can't distribute "java" because they didn't pass the little certification of code to be called java. They have to distribute patch sets.
THere is. Java is backed by a company saying java is java. Perl is backed by a core group.. same with python, php and ruby.
C and C++ are backed by ansi, who don't really seem to wield power of financial or social kinds when saying "wtf are you doing, THIS isn't c++"
Don't forget when you ran out of symbols. Game just ended :)
C isn't as portable as perl, python, php or java. C and C++ suffer from either incomplete or random interpretations of the specifications.
perfect example is TurboC and that stupid conio.h file. I came to the unix worled after learning TurboC and was confused for months until i learned what ncurses was.
*grumble*
'careful that Disney doesn't sue you on that process of layering :)
Class action lawsuit against sega for what? It's not like they lied about what was there, did they? Maybe I'm wrong on that part. But if someone wants to go to a pay system vs free, that's their choice.
But you realize, snailmail spammers are affected the same way. N% of people don't reply to or are affected by spam. It's out of what how uneven it is, no? So adding a 100th of a percent to the 99% prolly wouldn't hurt.
But then again, there's lies, larger ones and statistics.
Nonono.. that's what the handles are for. Just carry the bloody thing to your friend's house :)
There are a LOT of devices that are jack of all trades, master of none.
/w radios.
/w sonic.
Look at the iPod. It's not a great PDA. It never will be. Apple won't persue it very hard.
Look at those annoying scanner-fax-printer things. Great if you dont' care much about poor performance of one or even two of the functions. It's the same reason why the ones that really do everything right tend to be very expensive.
Counter-point, the tivo-directv combo box. They were meant for each other. It's rare when devices should combine. Radio's and music players are the same way. Look at receivers and walkmen. Surprised I don't see more cd players
BTW, the GBA functionality is almost always optional. Look at wind waker. It's simply a bonus to have the linkup. Same
The "Darwin Awards" sorta counter-prove your point. :)
Note. If you have an SMP system, and your cpu is faster than your geforce... why shell out for a new video card.
Uh.. no.. you can't predict that something will be fast enough to catch up on lost time and be fast enough to continue working.
If something takes in the long run 10 years from now, or 9 years from 2 years from now.... well you do the math. But no one knows what to expect. So making the best out of what you have is always good.
If someone did the research already, then I don't have to. If I didn't have space/money for 1001, then the point is moot, eh?
'cept when you mess with the settings that screw up performance. Easy one. Try lowering your clock speed. :)
A few % increase can be the difference betwee 9 or 10 years, which is even more visible in parallel processing. If one machine is one whole minute faster a week, and you have 1000 machines. you just gained a few hours. Remember, it's all in perspective.
As for a total waste of his time? Who cares? He was bored. We all have our hobbies. As long as he makes a living, and he comes out having had a good time, learning or what have you, be happy he did it. At least SOMEONE did.
Just a note on 1. It's not a panacea. You always get that one dork who decides to cc you on a forum post or something silly. But it does slow it down dramatically.