If you really don't think the two are any different, maybe the problems are in part caused by the fact that you don't know how to use C++ properly
The two are different, but mostly the same with respect to memory management. You have automatic "crope"; but you could have the same in C. The problem is passing around objects in all directions, letting users of objects possibly keep a reference to them and still timely destroying them.
Unfortunately, Kai got purchased by Intel, and (from what I see on their site [kai.com]) they seem to be dropping the
other platforms to support only intel. Really, the Intel compiler is really the Kai compiler, but only for Intel.
No, Intel, like some other chip/OS manufacturer (former DEC w/ Alpha), has its own compiler. Which is heavily worked on, to get decent SPECint/SPECfp figures ; with the added bonus that they can both test the chips and the compilers to see what chip design ideas are good and what are wrong "good ideas" in practice.
Re:Coming from a store owner...
on
The Euro
·
· Score: 1
The only way I can see someone tracking you via a Euro is this: you go to an ATM to withdraw money. The bank knows you and can read the Euro as it is dispensed. You spend the bill. Assuming the merchant deposits all cash to their bank, the bank could resolve the bill from the merchant to the bill dispensed from the ATM assuming it is the same bank or banks share info
In addition, one has to contrast this to what happens when one buy using a credit card (which are smart cards in my country, and make real time authentification to the bank)... Bank notes don't really look worse.
Re:Coming from a store owner...
on
The Euro
·
· Score: 1
The new anti-counterfeit measures contained in the Euro. This may seem like a good thing, but the larger Euro demoniations contain coils electromagnetically charged to a certain serial number. This can thus be tracked, and as much as consumers are worried about their privacy, merchants are worried about ours in respect to competitors.
I don't understand this at all. All bank notes I have, including british pounds, have a serial number written on them.
Additionally, I'll puke before I will allow my browser to decide to open an extra "pane" on the screen when I use Google, which is every few minutes for me.
Agreed, they should had a button "Disable this feature forever" in this pane. I was relieved later to discover that it could be disabled in the "Preferences" menus.
MT: I use Linux for work and sometimes I use Windows to play games.
You know this really should scare lokigames and the other linux game manufacturers...Here is there target market publicly stating that he uses windows for games.
Why? He is using Windows because many good games are there ; it just shows the reverse : there is a market of Linux users that are ready to play games, but who have to reboot.
There was a story on the news about a week ago about an 8 year old boy in Afghanistan who only wants to be able to sleep without fear--first of the tabliban, now of the bombing.
I don't know how this matches the point you were trying to make. After this, who will the boy tend to dislike/hate now ?
One of the few things that might make a dictatorship look brighter is a foreign, external agression.
You listed 14 packages, and I could probably come up with another 10 or 20 if I search around, but this is in stark contrast to the hundreds of open sourced packages available for *nix.
However these are the major ones. Add Perl, Python, Ruby, the Gtk-win32 alpha, xemacs, SDL, and you really have almost everything on Windows that a Unix developer has on Unix.
No, most shells (that I've seen) use crappy filename completion. Compare how it does it to something better, like a good web browser's address field.
Which shell do you use ? bash (both on Unix and Windows) is pretty decent.
Shells make you guess if you have enough characters for the completion, and force you to hit the tab key to trigger it.
Remember that 10-15 keystrokes are worth a click. I often double-tab right at the beginning to see the list of files. I hate clicking to select, especially since I run usually in maximal resolution (1600x1200) and mouse control is harder.
Browsers bring up a most likely choice (which could be based on a variety of factors - alphabetic order, frequency of use, explicit preference, etc. as you like)
You need to separate file display and file selection. First issue a ls command: "ls -lS"
sorted by decreasing size, "ls -ltr" sorted by inverse time, "find . -cmin -10",...
A bash expert can do "Ctrl-r" in bash allow to search a string in a previous command in the history, then cut with "Ctrl-k", then "Esc->" and paste with "Ctrl-y" at the end of the line, all this with an amazing speed.
BTW, it is almost unthinkable to use bash without using the Ctrl-r feature, which is a search in your previous history.
Also a classic is to use "find . -cmin -10", for finding the files that were changed less than 10 minutes ago ; handy when you extracted an archive in the wrong directory, and everything is messed:
"find . -cmin -10 | xargs rm -rf".
The filters such as "*.c", "*.h" are useful, also.
In some rare cases, I need to select manually many files, in a non-obvious way: then it's easy: "ls >/tmp/s && emacs/tmp/s", and the right amount of "Ctrl-k" does the job, faster than clicks.
Then I can directly do a "tar czvpf//mybackupmachine/backup/another.tar.gz $(cat/tmp/s)", instead of clicking a billion of aggravating times on menus, buttons, confirmations, etc...
The key of shells is first that you can combine commands, second that you have full history (with the mandatory Ctrl-r) that you can cut and paste in your new commands.
No one is perfect; no one remembers everything; and many people who might love to use CLI features are put off by this moronic insistance that they spend time learning how to use them down to their depths before they can do thing one
I think that's not the point. Maybe 99% of the users will never have the (costly) time to learn CLI ; the same way I don't have the time to learn properly MS Word, and use it in a crappy way. But saying that, even for the remaining 1%, CLI are inefficient is plain wrong.
As for the comment below re: keyboard shortcuts, testing reveals that they are for the most part slower than using the mouse. People don't consciously realize it, but unless the command is incredibly common and ingrained as only a few are, they pause and try to remember it.
When keyboard shortcuts are a matter of life or death, not only you learn them, but you learn to learn them so that to remember them efficiently. A example of GUI with many shortcuts (it was 100% keyboard shortcuts before newer version) is Blender. I spent a few hours learning the modeler and the keys, and never used it after, but I'm amazed when I come back to it, 2 years later, that I still remember most the keys. Of course, for a big system like Emacs, it's harder, that's why there is Meta-x with long names, apropos with "Ctrl-h a", and a 4+ page dense "reference card", all with the keyshort cuts, not to mention search in the index of the big emacs help ("Ctrl-h i", "Ctrl-s emacs:", "enter", "i", )
If you look, from the developer point of view, Python and Perl work times are similar to those of Lisp, along with program sizes. Of course, from the speed point of view, in the test, none of the scripting language could compete with Lisp.
Anyway some articles by Prechelt are interesting too (as many other research papers ; found via citeseer for instance)
As geekoid mentions below, if an employee
costs the company $50/hr (wages, payroll taxes, benefits, utilities, etc.) then a free(beer) program would have
to need 10 hours to learn to make the program cost more than MS Office.
But this argument ca be reversed. If using the full features of MS Office spares you a total of 10 hours on 2 years (that's like 2-5% in productivity), then it is worth it.
It's to the credit of the coders that they produce such impressive results so fast, and it'll certainly be interesting
reading when the full details are out. But let's not try to read too much into it this time, OK? Haskell is not
suddenly a million times better than OCaml was last year, just because OCaml doesn't feature in the top list
this time around. Functional programming still may not be the best approach for writing low-level instrument
control code or operating systems.
So, before the millions of posts start arriving, I make a small plea: don't treat this as an objective (no pun
intended) programming language comparison!
I don't think many are running out shouting "Haskell is the best programming language". Or that functional languages are great for everything... However for the 4th time in a row, they did extremely well, on a fun and non-obvious programming task. Usually, *some* people claim functional languages are slow or hard to program with, or only useful for toy programs: to some extent the contests proved the contrary.
I don't agree that the contest favored functional languages at all : the first one (1998) clearly didn't (how many chess or go programs are written in functional languages ?). This one didn't ; it was actually a document/optimization processing contest: it actually favored intelligence, and clever algorithms. Maybe such algorithms can be better be expressed in functional languages ?
Don't treat this as an objective (no pun
intended) programming language comparison!
It isn't, but do you have more objective programming language comparisons ?
Re:Hate to say, sounds like a dot-bomb strategy...
on
HP Buys Compaq
·
· Score: 1
Someone explain to me just how these gargantuan companies are going to turn a profit on IA-64? Like all Intel
processors, the Taiwanese clone makers will have a motherboard out a week before the chip comes out at 1/10
the price that HomPaq will be willing to sell it at.
Hmm... Well... I'm sure most customers who rans servers (this is the target of IA-64), are not going to try to save $500 by byuing a cheap taiwanese multi-processor motherboard. Just consider how much money the server can represent, and how much the administrator is paid to maintain the server.
Compare and contrast the progress air travel,
which had heavy private involvement, has made
this century, with the progress made by space
launch, which was mainly developed and is dominated
by the government.
The space technology looks way better.
They're still flying cargo to
ISS at $ 5000/pound or more with the Space Scuttle.
So ? Shipping 1 pound from this side of my building to the other side is way cheaper than all those costly $$$ airplanes. This is sure because those inefficient planes are build by bureaucratic companies which try to suck as much money from their customers as possible and to hand it to their shareholders.
Is that really what you want pharmaceuticals
research to look like?
They got to the moon and returned, which was never done, and which was still not done 30 years later by any private company, because there is no profit.
I want pharmaceuticals research to find cures that have not been found, even if there is no profit.
Why should the public care about an unknown programmer kept in jail for an obscure reason for a whooping two weeks, while you and the slashdot crowd, don't seem to even care about the innocent people in the death row, who are regulary freed (about 5 per year) after years, or dozen of years ? Not to mention those who are actually executed.
Re:IMHO: Perl-Python-Ruby
on
Why not Ruby?
·
· Score: 1
First, you are right that I owe some backup for the "faster" contention.
Thus there is no proof. In fact on some tasks, Python is faster, on other Ruby is faster, and more importantly, neither is faster than the other by an order of magnitude.
As for Stackless Python only being about continuations you are wrong. Stackless Python also allows for OS level threads
in Python. This allows for system calls not to block ALL interpreter threads. As for the utility of continuations, I too
haven't heard anything compelling, but my imagination can come up with vague ideas about many many concurrently
running state machines with out using threads or explicitly building a framework by hand.
Continuations break some semantics, and always slow down code, whatever the implementation. The showstopper of continuations is that they can't be implemented in Jython, and they break user C extentions, each of them and hundred of times more useful in the average case.
Guido has OVERLOADED the meaning of white space.
And why is this a tragedy ? Your reasoning is essentially: "AAAAAAAAArrrgh ! I'm *NOT* accustomed to *THIS*! I *cannot* stand *this*! It *must* disappear". This is mostly irrationnal since Ruby code is essentially Python code with "end" added at every end of block and vice-versa.
Oh yes, and note that Ruby is using newline as a statement delimiter, something in most other procedural/oop languages isn't done. Ruby is overloading newline. So what's your point?
Getting reference counts on variables was hard. Python has the same issue. Ruby doesn't have this
issue due to it's mark-sweep memory management.
Python 2.1 has some GC ; you are mistaken if you think Python programmers couldn't write a mark-sweep GC. Guido & friends, have choosen deliberatly to use GC, to avoid problems with C extensions. Simple example: imagine you have a 128 MB array which include a reference to one C++ object, which could (or couldn't) have a reference to a Ruby object. Are you sure Ruby doesn't need to scan the 128 MB array ? And reading memory at random might lead to segfault.
Lexical scopes...Python doesn't have them. It has the namespace scope (global) and the function scope.
Python 2.1 has lexical scopes as an option, and Python 2.2 will have them.
I must admit I am doing some amount of speculation
here, but I believe (with a tiny bit of support) that because blocks were relegated to second class language constructs,
Guido didn't want lexical scopes (prolly to confusing for the slope headed Python programmers;).
Blocks aren't first class language constructs. But this is a choice. Like, in Ruby, functions aren't first class objects.
However, Python has many idiosyncrasies.
Wow, and Ruby has absolutly none ?
It has its own share of historical
oddities like tuples (constant arrays, why not constant hashes, or constant variables, or even a 'constant' modifier on all
types)
But Ruby has none, 100% sure guaranted ? And Python oddities are so numerous that they cripple every single line of Python code, turning it into a unreadable mess ? Is this what you are contending ?
I'm sure former Java users now using Perl or Python could say the same thing. Of course Ruby is much more
than a scripting language. In fact, I really wish I could just totally stop programming Java and just use Ruby (since it can
solve the same problems), but I really don't think that is possible now since Ruby is so new (to the United States.) And of
course Java is pretty much the corporate mantra these days.
I think Python is much more suitable for Java programmers for a very simple and pratical reason : Jython, the Java implementation of Python. Jython can use Java packages with zero overhead, and you can inherit Java classes in Python, and inherit from Python classes in Java. That's extremely convenient.
Re:Why not not switch?
on
Why not Ruby?
·
· Score: 1
There are probably Perlies that think Perl is a bit too loose, and Pythonettes that think Python is a bit too strict.
Ruby uses setjmp and longjmp for its
exception handling, so if you call rb_raise, you'd better make sure you don't have any objects sitting on your stack that
might need to be destructed! I don't think Perl or Python do much better at this (though I could be wrong).
Well for Python, an exception is raised by a function by setting global (or thread-global) variables, and returning NULL. So you have to test manually (or to wrap around with C++ "try" statements), but it's better.
In my experience, basic performance tuning can easily create an order of magnitude improvement regardless of the language. And
in my experience, most engineers don't know how to do basic performance tuning.
Or they don't bother to introduce bugs in tested code, just to improve a performance that is already good enough.
Slashdot Mirror
The two are different, but mostly the same with respect to memory management. You have automatic "crope"; but you could have the same in C. The problem is passing around objects in all directions, letting users of objects possibly keep a reference to them and still timely destroying them.
No, Intel, like some other chip/OS manufacturer (former DEC w/ Alpha), has its own compiler. Which is heavily worked on, to get decent SPECint/SPECfp figures ; with the added bonus that they can both test the chips and the compilers to see what chip design ideas are good and what are wrong "good ideas" in practice.
In addition, one has to contrast this to what happens when one buy using a credit card (which are smart cards in my country, and make real time authentification to the bank)... Bank notes don't really look worse.
I don't understand this at all. All bank notes I have, including british pounds, have a serial number written on them.
Python 2.2 has actually useful lambdas, they are created with the Python construct "def".
Agreed, they should had a button "Disable this feature forever" in this pane. I was relieved later to discover that it could be disabled in the "Preferences" menus.
You know this really should scare lokigames and the other linux game manufacturers...Here is there target market publicly stating that he uses windows for games.
Why? He is using Windows because many good games are there ; it just shows the reverse : there is a market of Linux users that are ready to play games, but who have to reboot.
You define up front what a sane config is (THIS is the meat of the problem) and express it in a DTD or Schema.
You can't.
I don't know how this matches the point you were trying to make. After this, who will the boy tend to dislike/hate now ?
One of the few things that might make a dictatorship look brighter is a foreign, external agression.
However these are the major ones. Add Perl, Python, Ruby, the Gtk-win32 alpha, xemacs, SDL, and you really have almost everything on Windows that a Unix developer has on Unix.
Which shell do you use ? bash (both on Unix and Windows) is pretty decent.
Shells make you guess if you have enough characters for the completion, and force you to hit the tab key to trigger it.
Remember that 10-15 keystrokes are worth a click. I often double-tab right at the beginning to see the list of files. I hate clicking to select, especially since I run usually in maximal resolution (1600x1200) and mouse control is harder.
Browsers bring up a most likely choice (which could be based on a variety of factors - alphabetic order, frequency of use, explicit preference, etc. as you like)
You need to separate file display and file selection. First issue a ls command: "ls -lS" sorted by decreasing size, "ls -ltr" sorted by inverse time, "find . -cmin -10", ...
A bash expert can do "Ctrl-r" in bash allow to search a string in a previous command in the history, then cut with "Ctrl-k", then "Esc->" and paste with "Ctrl-y" at the end of the line, all this with an amazing speed.
BTW, it is almost unthinkable to use bash without using the Ctrl-r feature, which is a search in your previous history.
Also a classic is to use "find . -cmin -10", for finding the files that were changed less than 10 minutes ago ; handy when you extracted an archive in the wrong directory, and everything is messed: "find . -cmin -10 | xargs rm -rf".
The filters such as "*.c", "*.h" are useful, also.
In some rare cases, I need to select manually many files, in a non-obvious way: then it's easy: "ls > /tmp/s && emacs /tmp/s", and the right amount of "Ctrl-k" does the job, faster than clicks.
Then I can directly do a "tar czvpf //mybackupmachine/backup/another.tar.gz $(cat /tmp/s)", instead of clicking a billion of aggravating times on menus, buttons, confirmations, etc...
The key of shells is first that you can combine commands, second that you have full history (with the mandatory Ctrl-r) that you can cut and paste in your new commands.
No one is perfect; no one remembers everything; and many people who might love to use CLI features are put off by this moronic insistance that they spend time learning how to use them down to their depths before they can do thing one
I think that's not the point. Maybe 99% of the users will never have the (costly) time to learn CLI ; the same way I don't have the time to learn properly MS Word, and use it in a crappy way. But saying that, even for the remaining 1%, CLI are inefficient is plain wrong.
As for the comment below re: keyboard shortcuts, testing reveals that they are for the most part slower than using the mouse. People don't consciously realize it, but unless the command is incredibly common and ingrained as only a few are, they pause and try to remember it.
When keyboard shortcuts are a matter of life or death, not only you learn them, but you learn to learn them so that to remember them efficiently. A example of GUI with many shortcuts (it was 100% keyboard shortcuts before newer version) is Blender. I spent a few hours learning the modeler and the keys, and never used it after, but I'm amazed when I come back to it, 2 years later, that I still remember most the keys. Of course, for a big system like Emacs, it's harder, that's why there is Meta-x with long names, apropos with "Ctrl-h a", and a 4+ page dense "reference card", all with the keyshort cuts, not to mention search in the index of the big emacs help ("Ctrl-h i", "Ctrl-s emacs:", "enter", "i", )
However there is more data now, as, Prechelt itself widdened the study, and published in 2000 An empirical comparison of C, C++, Java, Perl, Python, Rexx, and Tcl (a detailed technical report is here).
If you look, from the developer point of view, Python and Perl work times are similar to those of Lisp, along with program sizes.
Of course, from the speed point of view, in the test, none of the scripting language could compete with Lisp.
Anyway some articles by Prechelt are interesting too (as many other research papers ; found via citeseer for instance)
As geekoid mentions below, if an employee
costs the company $50/hr (wages, payroll taxes, benefits, utilities, etc.) then a free(beer) program would have
to need 10 hours to learn to make the program cost more than MS Office.
But this argument ca be reversed. If using the full features of MS Office spares you a total of 10 hours on 2 years (that's like 2-5% in productivity), then it is worth it.
"Corporations have to make short term decisions that maximize profits for their shareholders "
Really?
Yes.
"profits for their shareholders or they can be sued"
Is that so?
Yes.
So, before the millions of posts start arriving, I make a small plea: don't treat this as an objective (no pun intended) programming language comparison!
I don't think many are running out shouting "Haskell is the best programming language". Or that functional languages are great for everything ... However for the 4th time in a row, they did extremely well, on a fun and non-obvious programming task. Usually, *some* people claim functional languages are slow or hard to program with, or only useful for toy programs: to some extent the contests proved the contrary.
I don't agree that the contest favored functional languages at all : the first one (1998) clearly didn't (how many chess or go programs are written in functional languages ?). This one didn't ; it was actually a document/optimization processing contest: it actually favored intelligence, and clever algorithms. Maybe such algorithms can be better be expressed in functional languages ?
Don't treat this as an objective (no pun intended) programming language comparison!
It isn't, but do you have more objective programming language comparisons ?
Someone explain to me just how these gargantuan companies are going to turn a profit on IA-64? Like all Intel
processors, the Taiwanese clone makers will have a motherboard out a week before the chip comes out at 1/10
the price that HomPaq will be willing to sell it at.
Hmm... Well... I'm sure most customers who rans servers (this is the target of IA-64), are not going to try to save $500 by byuing a cheap taiwanese multi-processor motherboard. Just consider how much money the server can represent, and how much the administrator is paid to maintain the server.
The space technology looks way better.
They're still flying cargo to ISS at $ 5000/pound or more with the Space Scuttle.
So ? Shipping 1 pound from this side of my building to the other side is way cheaper than all those costly $$$ airplanes. This is sure because those inefficient planes are build by bureaucratic companies which try to suck as much money from their customers as possible and to hand it to their shareholders.
Is that really what you want pharmaceuticals research to look like?
They got to the moon and returned, which was never done, and which was still not done 30 years later by any private company, because there is no profit.
I want pharmaceuticals research to find cures that have not been found, even if there is no profit.
Why should the public care about an unknown programmer kept in jail for an obscure reason for a whooping two weeks, while you and the slashdot crowd, don't seem to even care about the innocent people in the death row, who are regulary freed (about 5 per year) after years, or dozen of years ? Not to mention those who are actually executed.
Everyone is welcome.
Thus there is no proof. In fact on some tasks, Python is faster, on other Ruby is faster, and more importantly, neither is faster than the other by an order of magnitude.
As for Stackless Python only being about continuations you are wrong. Stackless Python also allows for OS level threads in Python. This allows for system calls not to block ALL interpreter threads. As for the utility of continuations, I too haven't heard anything compelling, but my imagination can come up with vague ideas about many many concurrently running state machines with out using threads or explicitly building a framework by hand.
Continuations break some semantics, and always slow down code, whatever the implementation. The showstopper of continuations is that they can't be implemented in Jython, and they break user C extentions, each of them and hundred of times more useful in the average case.
Guido has OVERLOADED the meaning of white space.
And why is this a tragedy ? Your reasoning is essentially: "AAAAAAAAArrrgh ! I'm *NOT* accustomed to *THIS*! I *cannot* stand *this*! It *must* disappear". This is mostly irrationnal since Ruby code is essentially Python code with "end" added at every end of block and vice-versa.
Oh yes, and note that Ruby is using newline as a statement delimiter, something in most other procedural/oop languages isn't done. Ruby is overloading newline. So what's your point?
Getting reference counts on variables was hard. Python has the same issue. Ruby doesn't have this issue due to it's mark-sweep memory management.
Python 2.1 has some GC ; you are mistaken if you think Python programmers couldn't write a mark-sweep GC. Guido & friends, have choosen deliberatly to use GC, to avoid problems with C extensions. Simple example: imagine you have a 128 MB array which include a reference to one C++ object, which could (or couldn't) have a reference to a Ruby object. Are you sure Ruby doesn't need to scan the 128 MB array ? And reading memory at random might lead to segfault.
Lexical scopes...Python doesn't have them. It has the namespace scope (global) and the function scope.
Python 2.1 has lexical scopes as an option, and Python 2.2 will have them.
I must admit I am doing some amount of speculation here, but I believe (with a tiny bit of support) that because blocks were relegated to second class language constructs, Guido didn't want lexical scopes (prolly to confusing for the slope headed Python programmers ;).
Blocks aren't first class language constructs. But this is a choice. Like, in Ruby, functions aren't first class objects.
However, Python has many idiosyncrasies.
Wow, and Ruby has absolutly none ?
It has its own share of historical oddities like tuples (constant arrays, why not constant hashes, or constant variables, or even a 'constant' modifier on all types)
But Ruby has none, 100% sure guaranted ? And Python oddities are so numerous that they cripple every single line of Python code, turning it into a unreadable mess ? Is this what you are contending ?
I think Python is much more suitable for Java programmers for a very simple and pratical reason : Jython, the Java implementation of Python. Jython can use Java packages with zero overhead, and you can inherit Java classes in Python, and inherit from Python classes in Java. That's extremely convenient.
Huh????? Ruby is exactly as strict as Python is.
Well for Python, an exception is raised by a function by setting global (or thread-global) variables, and returning NULL. So you have to test manually (or to wrap around with C++ "try" statements), but it's better.
http://www.twistedmatrix.com/users/jh.twistd/pytho n/moin.cgi/IcfpPythonTeam
Or they don't bother to introduce bugs in tested code, just to improve a performance that is already good enough.