I notice strong distinctions between those languages that have been developed by actual programming language researchers and those who hack together a language to scratch an itch.
"Actual programming language researchers" are typically not even interested in designing languages for general purpose use. Often they are just trying to explore a particular idea of aspect of programming.
Usually the latter turns out to be some baroque conglomeration of features piled on features, creating a very top heavy feeling to the language, while the former classification languages all have a purity to them, e.g. smalltalk, lisp, and c.
I think you're going to have to define your terms if you want to make this point. How were Kernighan and Richie programming language researchers but not Van Rossum and Wall? By now, the latter two have spent about ten years of their lives thinking about almost nothing other than programming languages.
Some would defend the "hack languages" as a means to Rapid Application Development, but Smalltalk has been shown to be the most productive language,
Could you provide a reference to back up that claim? I'd like to see evidence that Smalltalk fares well at system administration or text processing.
and Ruby/Python/Perl all seem to me to have a BASIC odor to them;
Now you are really grasping. That claim isn't even solid enough to refute.
...I'm wondering if people are afraid to learn a new way of speaking?
I don't know...are you? If you are into purity and elegance, I would suggest you give either Python or Ruby a real try. There are many Lisp fans that like both and Ruby is especially popular amoung Smalltalk users.
Your post indicates that you have a pretty limited view of the programming language space. PHP is "Perl and C/C++ wrapped into one" because it has "classes", "great string manipulation", "support for CGI" and "now GTK" These are all supported by Python, Perl, Ruby and a host of other languages. PHP is just one among many. PHP's real virtue is that it is totally embedded in the web server environment. It is hard to justify it purely in terms of language features as you seem to want to. It is a mediocre language embedded in a great dynamic web pages environment.
And anyhow, the existence of these many server side languages do not really have any impact on the need for languages on the client side. Yes, PHP can generate JavaScript, HTML, Flash and other stuff that works on the client-side. But really Curl is competing with those client-side languages, not with PHP. PHP could just as easily generate Curl if it turned out to be better than JavaScript et. al. So PHP is great at what it does but not really relevant to the question of Curl's utility or viability.
...is to generate press releases, news articles and Slashdot threads. I've never heard anyone use it. I just hear people pointing it out to each other and saying: "Tim Berners-Lee backs it." As if that automatically makes it more interesting than all of the other languages (client and server side) out there.
Pay as you go looks like a pretty clear way to kill a programming language to me...
It is amazing how many Slashdot readers will slam things without understanding them. I suggest you read Mono's.NET FAQ and see if there is anything in there that would lead to "Lost Privacy. Stupid Security bugs. Pay-to-play software."
The very first two questions address your concerns.
You're confusing things by bunching together trademarks and copyrights under the obfuscatory category "Intellectual Property":
Microsoft says that the GPL is against "intellectual property rights." I have no opinion on "intellectual property rights," because the term is too broad to have a sensible opinion about. It is a catch-all, covering copyrights, patents, trademarks, and other disparate areas of law; areas so different, in the laws and in their effects, that any statement about all of them at once is surely simplistic. To think intelligently about copyrights, patents or trademarks, you must think about them separately. The first step is declining to lump them together as "intellectual property".
If the Taliban called itself a communist movement, that's how the media would report it. If they called themselves a Christian sect then that's how the media would report it. They call themselves fundamentalist muslims so that is how the media reports it!
Typical American Internationalism
on
Internet2 Update
·
· Score: 1
In June, the consortium announced it now had member universities in all 50 states.
Oh, well then. Everything important is covered. Great!
Rather than forcing you to do it Guido's Way, you can do it the Perl Way, or the Smalltalk Way, or the Functional Way... or any combination of the above.
Right...Python forces you to do it Guido's way. You can't use Python like Perl even though it has a strong, Perl-compatible regular expression library. And you can't ues it like Smalltalk even though it has OO and can even support meta-OO types of programming. And it doesn't support functional programming even though it has map, filter, lambda and list comprehensions.
I would suggest you actually use Python before you start guessing about what it is like.
Re:My article was modified slightly when posted.
on
Why not Ruby?
·
· Score: 1
I believe that every language has its place
That explains why you asked the question. You think about languages differently than most people.
Most people can't afford to learn every language that comes along, and port all of their projects over, and rewrite their extension modules and.... there are a finite number of programmer resources in the world and most don't consider the thrill of a new language to be worth the pain of incompatibility between their old code and their new code.
I have done projects in Python that I do not think I could have done in Perl or Ruby because of Python's clean namespaces, clear OO and portable, robust, native thread support. So there are practical reasons to learn Python. There are probably also projects that would be MUCH easier in Perl than in Python (e.g. those that are solved by a module on CPAN, or by Perl's great database support).
But on what sort of project would Ruby be uniquely qualified? When I understand that, I'll know why I should learn Ruby. I'll also be able to judge whether Ruby is the next big thing or just another language.
Maybe Ruby is the next big leap. But you haven't really made that case. All you said was that it had slightly better OO than Python. Help me to understand it in terms of the problems I have to solve. Show me a problem that would be hard to solve in Python but easy to solve in Ruby.
His followers called him Mahasamatman and said he was a god. He preferred to drop the Maha- and the -atman, however and called himself Sam. He never claimed to be a god. But then, he never claimed not to be a god. Circumstances being what they were, neither admission could be of any benefit.
This is one of my favorite science fiction books. Zelazny always married the mythic to the scientific. Aren't the Nine Princes in Amber also akin to Greek Gods? Pantheistic religions have always believed that earth is a reflection of heaven.
Lord of the Light is his most extended fugue on that theme. The book isn't very politically correct, though. The bad guys are Hindu gods and a follower of Christianity. The hero follows the path of Buddha on purely pragmatic grounds.
Python has not "taken off" as Java did or Perl did in their day because it is not synonymous with an exciting new platform like applets or CGI. Not that applets or CGI matter today but they were neverthless great publicity for those languages.
Python is too general purpose to get attached to any particular niche like that. It is just a little bit better for most taks than most of the competition. If you don't believe me you'll have to try it out yourself.
I edit Python in VI (well, VIM) every day. It isn't very difficult.
We're talking about a programming language here. Don't you think that there are bigger issues than end of block delimiters? Polymorphism. Extensibility. Zen?
This argument for python white-spacing could also be applied to any/all of the languages that required statments to begin at a certain column. We all know how productive that requirement was.
Your argument is silly for two reasons. First, the column-oriented languages were designed to make life easier for the computer, not for the human.
Second, actual programmers used those languages and hated the column-orientedness. Tens or hundreds of thousands of Python programmers use Python and like the whitespace rules. It is only people who haven't used it (like the original poster) who complain that it has theoretical problems.
It is extremely annoying to have people who haven't tried it claim that it can't work (even though it demonstrably does) because it is vaguely reminicient of other languages where a totally different feature did NOT work.
This post is factually incorrect. The Python interpreter is multithreaded. It does not have "free threading" but it does hae threads.
The C API is actually a joy to work with compared to Perl or Java. TCL is a very simple language with very simple data types (i.e. no built-in ojbects) and that makes it's embedding API very simple. Python's C API is pretty damn nice for such a relatively sophisticated language.
Python is not going to be rewritten from scratch. Guido has stated that this isn't a goal anymore. Various bits like the API and the threading may get better over time incrementally if people with the right interest pitch in.
There are two different issues here. One is standardization. The other is change.
Standardized languages change. In fact, during the standardization period that they probably change more than unstandardized languages. After standardization they still change, just more slowly. For instance "C" has recently undergone a revision. If you depend on new features from a new C compiler, your code won't work with older ones. C'est la vie.
It's important to recall that the reason that Microsoft benefits from owning both the most popular applications AND the most popular OS is that it sells both. Both are huge cash cows.
Let's imagine that Microsoft moves its apps to.NET. It isn't clear how they make money from the ".NET" part of the equation? Will they sell the.NET framework? Sell the development tools? It's hard to see how either option could be near as profitable as selling an operating system.
Microsoft likes to bundle mostly unrelated technologies under a single banner in order to "simplify the message."
The ".NET Framework" is a virtual machine and standard library similar to the JVM..NET bytecodes are not tied to any particular hardware so in a few years shrinkwrapped software may not be specific to Intel or Alpha.
The primary difference between the.NET framework and the JVM is that.NET is supposed to be multi-language whereas Sun unashamedly promotes Java as the "One True Programming Language." Multi-language applications written for.NET can be a lot more integrated than COM or CORBA. There is a single exception architecture, a well-defined debugging architecture etc.
ASP.NET is Microsoft's Web development platform. It has a concept of "web services" which is basically distributed computing re-invented based on Web-ish technologies.
The win32 GUI APIs have been replaced with "Windows Forms".
There isn't really a lot in.NET that points to "a future where people rent rather than buy applications." Obviously applications have been moving to the desktop to the Web for years and.NET has additional features that will allow developers to take that a few steps farther. But the new.NET features are useful in traditional applications also.
"Rent" versus "buy" is basically a marketing and distribution decision. You could do the same thing with Java bytecodes or even Active-X controls.
Some people gather around languages because its the language they learned. But other people choose languages that support their philosophy of computing or the world in general.
Perl encourages syntactic creativity (some would say idiosyncracy). Python encourages cleanliness and simplicity (some would say conformity). Functional languages encourage a mathematical view of programming.
You could argue that just as in the real world, it is your *responsibility* to promote the philosophy that you believe in. That doesn't mean you should be blind to the virtues of languages that conform to other philosophies, but it also means that you should not be ashamed to promote yours. Our industry will only advance by embracing the philosophies that work better. Remember "goto considered harmful?"
If you are programming in a modern programming language like Python or even JavaScript, with a portable GUI framework like Tk, Swing or Mozilla, it really sould not matter what operating system you are using.
Rather than promoting "Unix for schools" we should promote "platform-independent software development for schools." We should stress that if you focus on that which works across platforms you come to understand better the universal themes of computer science rather than the specifics of an OS.
Once the operating system becomes more or less irrelevant, schools will of their own volition choose the operating system that is cheaper, more secure, easier to maintain programatically and so forth.
ActiveState did not port Python to Windows. Python has run on Windows for almost a decade. As far as I know, it has run on Windows since versin 1. ActiveState does support some platform-specific enhancements to Python for Windows and other platforms. ActiveState also supports platform-independent extensions to Python.
It has nothing to do with Netscape's release schedule. Microsoft sends representatives to the W3C. They agree with representatives of other companies on what standards to make. They write the standards. Everyone else goes home to work on their implementations, expecting the standard to arise in the browsers eventually. Microsoft releases something totally unrelated instead. Everyone who participated in the process got shafted and so did the people who have to create web pages that work with multiple user agents.
Whatever we may think about the RIAA, MP3s and the future of information, it is sensationalist to claim that an injunction against KNOWINGLY linking DIRECTLY to illegal material would shake the foundations of the Web as we know it.
If Yahoo accidently links to pirated information, that's a totally different issue in terms of the law. Intent can matter.
Beware slipperly slope arguments. They have their place but critical thinking can often indicate that the slope is not so slippery.
Slashdot Mirror
That's a real killer argument against a language. The depth of your insight astounds!
Your post indicates that you have a pretty limited view of the programming language space. PHP is "Perl and C/C++ wrapped into one" because it has "classes", "great string manipulation", "support for CGI" and "now GTK" These are all supported by Python, Perl, Ruby and a host of other languages. PHP is just one among many. PHP's real virtue is that it is totally embedded in the web server environment. It is hard to justify it purely in terms of language features as you seem to want to. It is a mediocre language embedded in a great dynamic web pages environment.
And anyhow, the existence of these many server side languages do not really have any impact on the need for languages on the client side. Yes, PHP can generate JavaScript, HTML, Flash and other stuff that works on the client-side. But really Curl is competing with those client-side languages, not with PHP. PHP could just as easily generate Curl if it turned out to be better than JavaScript et. al. So PHP is great at what it does but not really relevant to the question of Curl's utility or viability.
...is to generate press releases, news articles and Slashdot threads. I've never heard anyone use it. I just hear people pointing it out to each other and saying: "Tim Berners-Lee backs it." As if that automatically makes it more interesting than all of the other languages (client and server side) out there.
Pay as you go looks like a pretty clear way to kill a programming language to me...
It is amazing how many Slashdot readers will slam things without understanding them. I suggest you read Mono's .NET FAQ and see if there is anything in there that would lead to "Lost Privacy. Stupid Security bugs. Pay-to-play software."
The very first two questions address your concerns.
You're confusing things by bunching together trademarks and copyrights under the obfuscatory category "Intellectual Property":
GNU GPL and the American Way
If the Taliban called itself a communist movement, that's how the media would report it. If they called themselves a Christian sect then that's how the media would report it. They call themselves fundamentalist muslims so that is how the media reports it!
In June, the consortium announced it now had member universities in all 50 states.
Oh, well then. Everything important is covered. Great!
Right...Python forces you to do it Guido's way. You can't use Python like Perl even though it has a strong, Perl-compatible regular expression library. And you can't ues it like Smalltalk even though it has OO and can even support meta-OO types of programming. And it doesn't support functional programming even though it has map, filter, lambda and list comprehensions.
I would suggest you actually use Python before you start guessing about what it is like.
That explains why you asked the question. You think about languages differently than most people.
Most people can't afford to learn every language that comes along, and port all of their projects over, and rewrite their extension modules and .... there are a finite number of programmer resources in the world and most don't consider the thrill of a new language to be worth the pain of incompatibility between their old code and their new code.
I have done projects in Python that I do not think I could have done in Perl or Ruby because of Python's clean namespaces, clear OO and portable, robust, native thread support. So there are practical reasons to learn Python. There are probably also projects that would be MUCH easier in Perl than in Python (e.g. those that are solved by a module on CPAN, or by Perl's great database support).
But on what sort of project would Ruby be uniquely qualified? When I understand that, I'll know why I should learn Ruby. I'll also be able to judge whether Ruby is the next big thing or just another language.
Maybe Ruby is the next big leap. But you haven't really made that case. All you said was that it had slightly better OO than Python. Help me to understand it in terms of the problems I have to solve. Show me a problem that would be hard to solve in Python but easy to solve in Ruby.
His followers called him Mahasamatman and said he was a god. He preferred to drop the Maha- and the -atman, however and called himself Sam. He never claimed to be a god. But then, he never claimed not to be a god. Circumstances being what they were, neither admission could be of any benefit.
This is one of my favorite science fiction books. Zelazny always married the mythic to the scientific. Aren't the Nine Princes in Amber also akin to Greek Gods? Pantheistic religions have always believed that earth is a reflection of heaven.
Lord of the Light is his most extended fugue on that theme. The book isn't very politically correct, though. The bad guys are Hindu gods and a follower of Christianity. The hero follows the path of Buddha on purely pragmatic grounds.
In this government-less universe, who stops the fishermen from overfishing a species into oblivion?
Who steps in to stop emotional and sexual abuse of children?
Who sets emissions limits for pollution producers?
Who negotiates with foreign powers on behalf of the residents of the United States?
Who defends against heavily armed totalitarian countries elsewhere in the world?
Who tries to smooth the wild swings of the economy?
Government cannot and should not be limited to a narrow reactionary role. It needs to be able to act in the public interest.
Python has not "taken off" as Java did or Perl did in their day because it is not synonymous with an exciting new platform like applets or CGI. Not that applets or CGI matter today but they were neverthless great publicity for those languages.
Python is too general purpose to get attached to any particular niche like that. It is just a little bit better for most taks than most of the competition. If you don't believe me you'll have to try it out yourself.
I edit Python in VI (well, VIM) every day. It isn't very difficult.
We're talking about a programming language here. Don't you think that there are bigger issues than end of block delimiters? Polymorphism. Extensibility. Zen?
This argument for python white-spacing could also be applied to any/all of the languages that required statments to begin at a certain column. We all know how productive that requirement was.
Your argument is silly for two reasons. First, the column-oriented languages were designed to make life easier for the computer, not for the human.
Second, actual programmers used those languages and hated the column-orientedness. Tens or hundreds of thousands of Python programmers use Python and like the whitespace rules. It is only people who haven't used it (like the original poster) who complain that it has theoretical problems.
It is extremely annoying to have people who haven't tried it claim that it can't work (even though it demonstrably does) because it is vaguely reminicient of other languages where a totally different feature did NOT work.
This post is factually incorrect. The Python interpreter is multithreaded. It does not have "free threading" but it does hae threads.
The C API is actually a joy to work with compared to Perl or Java. TCL is a very simple language with very simple data types (i.e. no built-in ojbects) and that makes it's embedding API very simple. Python's C API is pretty damn nice for such a relatively sophisticated language.
Python is not going to be rewritten from scratch. Guido has stated that this isn't a goal anymore. Various bits like the API and the threading may get better over time incrementally if people with the right interest pitch in.
There are two different issues here. One is standardization. The other is change.
Standardized languages change. In fact, during the standardization period that they probably change more than unstandardized languages. After standardization they still change, just more slowly. For instance "C" has recently undergone a revision. If you depend on new features from a new C compiler, your code won't work with older ones. C'est la vie.
Why not try out Python for the Palm instead? Python can do pretty much the same stuff Perl can do and some people even prefer it!
Slashdot did a Story on it recently.
It's important to recall that the reason that Microsoft benefits from owning both the most popular applications AND the most popular OS is that it sells both. Both are huge cash cows.
.NET. It isn't clear how they make money from the ".NET" part of the equation? Will they sell the .NET framework? Sell the development tools? It's hard to see how either option could be near as profitable as selling an operating system.
Let's imagine that Microsoft moves its apps to
Microsoft likes to bundle mostly unrelated technologies under a single banner in order to "simplify the message."
.NET bytecodes are not tied to any particular hardware so in a few years shrinkwrapped software may not be specific to Intel or Alpha.
.NET framework and the JVM is that .NET is supposed to be multi-language whereas Sun unashamedly promotes Java as the "One True Programming Language." Multi-language applications written for .NET can be a lot more integrated than COM or CORBA. There is a single exception architecture, a well-defined debugging architecture etc.
.NET that points to "a future where people rent rather than buy applications." Obviously applications have been moving to the desktop to the Web for years and .NET has additional features that will allow developers to take that a few steps farther. But the new .NET features are useful in traditional applications also.
The ".NET Framework" is a virtual machine and standard library similar to the JVM.
The primary difference between the
ASP.NET is Microsoft's Web development platform. It has a concept of "web services" which is basically distributed computing re-invented based on Web-ish technologies.
The win32 GUI APIs have been replaced with "Windows Forms".
There isn't really a lot in
"Rent" versus "buy" is basically a marketing and distribution decision. You could do the same thing with Java bytecodes or even Active-X controls.
Some people gather around languages because its the language they learned. But other people choose languages that support their philosophy of computing or the world in general.
Perl encourages syntactic creativity (some would say idiosyncracy). Python encourages cleanliness and simplicity (some would say conformity). Functional languages encourage a mathematical view of programming.
You could argue that just as in the real world, it is your *responsibility* to promote the philosophy that you believe in. That doesn't mean you should be blind to the virtues of languages that conform to other philosophies, but it also means that you should not be ashamed to promote yours. Our industry will only advance by embracing the philosophies that work better. Remember "goto considered harmful?"
Paul Prescod
If you are programming in a modern programming language like Python or even JavaScript, with a portable GUI framework like Tk, Swing or Mozilla, it really sould not matter what operating system you are using.
Rather than promoting "Unix for schools" we should promote "platform-independent software development for schools." We should stress that if you focus on that which works across platforms you come to understand better the universal themes of computer science rather than the specifics of an OS.
Once the operating system becomes more or less irrelevant, schools will of their own volition choose the operating system that is cheaper, more secure, easier to maintain programatically and so forth.
Paul Prescod
ActiveState did not port Python to Windows. Python has run on Windows for almost a decade. As far as I know, it has run on Windows since versin 1. ActiveState does support some platform-specific enhancements to Python for Windows and other platforms. ActiveState also supports platform-independent extensions to Python.
It has nothing to do with Netscape's release schedule. Microsoft sends representatives to the W3C. They agree with representatives of other companies on what standards to make. They write the standards. Everyone else goes home to work on their implementations, expecting the standard to arise in the browsers eventually. Microsoft releases something totally unrelated instead. Everyone who participated in the process got shafted and so did the people who have to create web pages that work with multiple user agents.
Whatever we may think about the RIAA, MP3s and the future of information, it is sensationalist to claim that an injunction against KNOWINGLY linking DIRECTLY to illegal material would shake the foundations of the Web as we know it.
If Yahoo accidently links to pirated information, that's a totally different issue in terms of the law. Intent can matter.
Beware slipperly slope arguments. They have their place but critical thinking can often indicate that the slope is not so slippery.