Forgot your password?
Programming IT Technology

Ask Guido van Rossum 202

Posted by Roblimo
from the land-of-the-pythons dept.
This week's interview is with Guido van Rossum, a man who, as they say, needs no introduction. (Not around here, at least.) To learn a bit more about him, check his personal page. You might want to ask him about Python 2.1, which was released today. One question per person, please. We'll send 10 of the highest-moderated ones to Guido about 24 hours after this post went up, and will run his answers as soon as he gets them back to us.
This discussion has been archived. No new comments can be posted.

Ask Guido van Rossum

Comments Filter:
  • by Anonymous Coward

    Please mod this down. It's a fair question, but it's kind of uninformed. It certainly doesn't pertain specifically to Python.

    (First of all, Python doesn't have ++ and -- operators.)

    Even C goes through revisions. Some C programs that work on more recent C compilers won't work under older ones.

    Any language newer than C is likely to have this problem, because computing (let's face it) is exploding. The number of standard libraries people want is always increasing. Java definitely has this problem. Just one call to a v1.3 API prevents your program from running in a v1.2 JRE. Likewise, Perl has new language features and libraries with each release. And so on.

    Either you accept change, and you get all the benefits of progress; or you standardize on one Final version, and you get the benefit of language stability. Most people prefer for Python (and Perl; and Java; etc.) to keep improving. Go figure.

    For your Debian box at home, I recommend upgrading to (at least) Python 2.0.

    -- jason

  • by abischof (255) <alex AT spamcop DOT net> on Tuesday April 17, 2001 @08:33AM (#285871) Homepage
    Considering that you named the language after the comedy troupe, what's your favourite Monty Python sketch? Personally, my favourite is the lecture on sheep aircraft [], but I suppose that's a discussion for another time ;).

    Alex Bischoff
  • I've got to disagree with you wrt the C API -- I find it quite pleasent (and though I haven't worked with TCL, I do have prior experience with the C APIs of Perl and, to a limited extend, Scheme). Threading, on the other hand, is a real issue (has Stackless Python resolved it?) and I'm as interested as you to find out when it'll be fixed in the official Python implementation.

    Of course, there's also JPython, which has fantastic integration between Java and Python code, and resolves the threading thing... so if yer extending Java code, it's really a no-brainer (IMHO).

  • Forget that other guy -- of course Python is interpreted, just like everything else: it all comes down to machine code, which the CPU interprets.

    There has been efforts, most notably the (currently inactive, I believe) static typing SIG [], which would make it possible to make code that could be efficiently translated to machine code (also known as "compiling"). Static typing is by no means required to translate to machine code, but it would probably make that faster.

    An interesting option might be something like Squeak [] did, with a subset of the language that can be efficiently translated into C (and then compiled). This is something like what the static type SIG was aproaching -- not quite Python, but something that could run in the conventional CPython interpreter, wouldn't necessarily require any knowledge of C or assembly, but could run fast.

    Of course, it is also possible to simply compile Python into system code just as it is (well, if someone wrote the compiler). There is nothing that makes Python inherently interpreted. The problem, however, would be in the efficiency of that code. A naive attempt to do this would probably be slower than CPython. OTOH, compilers for Scheme (which have many similar issues as Python) have produced very fast code (on par with C). In particular, Stalin is very fast (though there are others which are very fast as well) -- sometimes faster than C. Python doesn't have any of the huge flaws that languages like, say, Tcl have in this regard. Some operations are difficult in compiled environments -- like getattr or eval -- but they could still be possible, or they could be left out (since much code doesn't use them). OTOH, all the best Python code uses them.

  • There has already been work towards this, particularly in the last few months.

    distutils fulfills some of the functionality required for a Python CPAN. I think it has reached a state of relative maturity -- it may be asked to do more later, but it does what it needs to well right now.

    There has also been some discussion and implementation of methods to describe and upload modules, which would probably be included in distutils or something related when they mature. I think at that point, along with a little infrastructure on the web, Python will have its own little CPAN. I know CPAN does more than this, but not all that much more, and it's something that should be grown into.

    I couldn't find what I was really looking for on the subject, but maybe this thread [] would be a starting place. There's more stuff elsewhere as well. This message [] (from that thread) gives a nice overview of what's necessary for CPAN, I think. And I guess the whole discussion starts here [].

  • Almost every modern high-level application language today supports, and perhaps forces the Object Oriented Paradigm (OOP). Students are encouraged to define and use objects whenever possible. Python's standard library seems nothing but objects. Java is the same. Newer technologies, like SOAP/Microsoft .Net/blahh are the apex of this concept.

    When I refer to the OOP, I mean the drive to decompose all programming into components, not necessarily the individual concept of having classes with methods and constructors and context-sensitive results, etc

    Application developers accept the OOP as the only way and consider those who refuse it to be uncivilized coder barbarians. Clearly, one can only bring sanity to programming via the OOP. But what is it really bringing?

    Most of the ideas that the OOP promotes are good programming practice anyway. That is absolutely not what I hate about the OOP.

    In the very ideal cases, you can create a reusable object/module that other people can enjoy. This is very rare, though. UNIX is a good example of an easy to use interface that allows for massive code reuse. The Win32 API, while affording code reuse, has a miserable interface that makes Windows programming a chore.

    Without going into a huge tirade; Modularity on that level is good. It's worth it to struggle with the interface, because the alternative is 100 man-years worth of functionality that you need to implement. On a much reduced level, trying to deal with the OOP just doesn't seem worth it.

    The problem with the OOP is that it encourages all code to be tiny little modules with it's own unique domain, which helps complicate the code both visually and in terms of execution.

    For every beautifully designed reusable component, you have a thousand more that are confined to a single project and do nothing but add complexity and visual noise to an otherwise simple idea.

    The Objects Everywhere philosophy seems to promote complexity, rather than simplicity. The less code, the easier it is to understand, the better off it'll be. Python does achieve many of these goals, but I can't understand why the push to OO'ize it all.

    Bad programmers can write bad code no matter what, and there's a vast army of bad programmers out there, but I'm not sure I've ever seen good code that employs the Objects Everywhere ideal.

    Does this make sense? What are your thoughts?

  • Don't know where you got this, but he didn't start with Pascal. Pascal is *very* type-strict, but Python doesn't even require you to explicitly declare variables! He started off with ABC, a small language he had designed earlier for the purpose of teaching beginners how to program (similar to BASIC in that sense). That is essentially why Python uses indentation to denote block rather than begin's, end's or braces, and why statements end with a newline rather than a semicolon: Sensible and intuitive to a beginner, especially since it promotes good coding style (or enforces it, depending on your point of view), but a little odd for a programmer already fluent in languages such as C, Awk, Perl, etc.

    Python is my favorite language, btw.
  • Actually neither type inference or static typing is required for a compiled implementation of a language. Common Lisp requires neither, though you can declare type information for greater efficiency. Instead of variables having an intrinsic static type, values have a type, and variables can refer to any kind of value. Typically the type of the value is encoded in the reference to it. So you might use the top three bits of a word to encode the type and the lower 29 to be a pointer to the value. Explicit declared typing information does help the compiler though- you can sometimes double the performance of a compiled Common Lisp program by using type declarations.

  • Good programmers indent nested blocks, but that's just to make the code easier to read. The parser doesn't care.

    That's really the beautty of Python's whitespace blocks. Both the programmer and the parser are looking at the same cues for block nesting. I am sure you have debugged C code that was missing a brace (or worse that had one misplaced) but was still indented "correctly." Your mind thinks that everything is hunky dory because the code "looks" right. This doesn't happen very often if you are a skilled C coder (with an intelligent text editor), but it does happen. And it happens a lot with newbie programmers. I was teaching my little brother Perl at one point, and he had all sorts of trouble with braces. However, when we switched to Python there was no longer any need for him to think about which braces matched up. I didn't believe that Python's significant whitespace was a good idea either, at first, but I am a believer now.

    Oh yeah, and since the Vim % command (jump to matching bracket) doesn't work with Python code, do you know of a macro to replace it?

    I am not a vi user, so I can't help you there.

  • It is well known fact, that the latest versions of
    Python has licence, that is incompatible
    with GNU GPL. I know that RMS himself has tried to
    negotiate to rectify that problem. But what is
    current situtation? Do you see any chance to resolve the problem? How likely is it, that the
    problem will be resolved?
  • Guido,

    I am a long time Python developer and fanatic. I have long enjoyed using python because of its Object Oriented nature and its faster and more open development process than its main competitors.

    While I believe Python is a great language, there is one thing that I really am missing from Python. Currently, Python uses a very antiquated reference counting method that is less than ideal. Are there any plans to re-implement the garbage collection mechanism for Python in the future? May I suggest using the more advanced generational methods of garbage collection?

    Good work so far, and thanks for all you have done!

    Jonathan LaCour

    Developer, Student

  • Guido,

    Python is great, and I use it all the time, but I have an annoyance that I would love to see corrected.

    Currently, Python's standard modules provide much functionality, but are a mish-mash of submitted modules that are largely uncategorized. Java provides a much better organization and naming scheme that could be carried over to Python.

    For example, when importing the Python HTTP modules (httplib), currently, I do one of the following:

    import httplib
    from httplib import *

    Java has a categorization structure imposed to make things a bit clearer. I would prefer if this same kind of structure was imposed in python. So, the above would then become something like one of the following:

    from import *

    I think that this could really enforce a cleaner organization and make it easier to identify where modules belong, and what they do, just by looking at their organizational location.

    What do you think? =)

    Jonathan LaCour
    Developer, Student

  • What other languages do you use? Why do you use them? Are any of them better than python? When can I expect Parrot to be released? ;-)

  • Would you hold up the rewriting of python to use a stackless implementation because Jython or other implementations will not be able to implement it? Where is the line between pushing python forward, and holding it back so that it works nicely with everything else ?

    Thanks for a great language.
  • Queues can be done using Python lists; linked lists could be done using tuples or lists. AVL trees and B-trees are available in persistent forms (through the BerkeleyDB library, for example), and there's a BTrees package inside the ZODB. PEP 218 [] proposes a built-in type for sets, but there doesn't seem to be much agreement on what people want from a set type. Some people want fast intersection operations and some don't care; some want; some people want an O(1) membership test and some people don't care; it seems difficult to write a set type that's equally speedy for all possible applications.
  • Note that Python 2.0 included a cycle-detection algorithm, so reference cycles do get cleaned up now. See the Modules/gcmodule.c file in the Python source distribution, and the GC module docs [].
  • by Tumbleweed (3706) on Tuesday April 17, 2001 @08:49AM (#285886)
    Do you have any plans to merge Stackless Python with Python? If yes, when? If not, why not?
  • Since the original question referenced VB and Kylix I'm assuming the poster wants a 'GUI builder' type of IDE. I believe the most common Python approach to GUI widgets is through Tk. In that case the question might be, are there any Tk GUI builders that handle Python. I don't know the answer, I'll leave the searching as a reader exercise. There are a couple of basic IDEs for Python, the Windows version even comes with the install. These are about on par with Komodo, faster but with different features.

    Chris Cothrun
    Curator of Chaos
  • Incidentally, Haskell [] can also use indentation to indicate structure.
  • by Luke (7869)
    I was asking about Guido's thoughts.

    I've already downloaded the source and compiled it on Cygwin. Flawless and easy install.
  • by Luke (7869)
    Where do you think I got the source?

    It wasn't from your particular link, however, but I thank you for your gracious help.
  • by Luke (7869) on Tuesday April 17, 2001 @08:10AM (#285891)
    Thoughts on Ruby?
  • Given Python's more limited exposure to the commoners, I've had varying amounts of success getting companies that I've worked for to actively use Python. Those that I have convinced, have grown to love it, and all is good with the world.

    What are the five most compelling arguments that _you'd_ offer to a company doing web development, system administration, etc... to use Python over other interpreted languages such as Perl, TCL, etc...

  • b) If you move control blocks horizontally (e.g. after removing a preceeding conditional), your editor (EMACS with python-mode for me) may introduce subtle errors when re-tabbing.

    Nah. Not if you have emacs configured correctly. Unfortunately I'm in Windows right now (long story) so I can't refer to my own Emacs settings. If you have a non-ancient version of Python mode, Emacs actually handles the Python whitespace rather well.

    And make sure when you move a block horizontally you make the block the emacs region then hit (I think) Ctrl-C then < or >. It's also on the Python mode menu.

    One good tip for maintaining sanity in a professional Python dev shop is to have EVERYONE USE TABS, always, all the time. One tab character == 1 Python indentation level. Then each person can set their tabs to 4 or 8 or whatever space for visual display, and everyones happy. Just make sure peoples editors are set to keep those as tabs and not convert back to spaces. If you set up Emacs properly (very easy) then this becomes a no-brainer. Hitting Tab on a line should (almost always) produce the right indentatin for a given context.

    Mixing up 4-spaces-as-indentation and tabs in a single file is a recipie for disaster (subtle bugs.) Aside from these issues, I actually like Pythons use of whitespace. I find it easier for long term maintainability not to mention the initial conceptualization / prototyping.

    Email me if you need those .emacs settings to work with Python and tabs properly.


  • I noticed that you left our beloved country and went to the US.

    I was wondering: why did you do it? How do The Netherlands (or Europe for that matter) and the US compare? Don't you miss the good bakeries, cannabis and licquorice? Do you experience Americans as being shallow (I've heard that comment more than once)? Or is this just a case of "Cherchez la femme?"

    That's a lot questions together, but I'm not so much interested in specific answers. I'd like to hear your general experience of US vs. Europe.

  • Is there any chance that stackless [] and microthreads [] might be integrated into the main python distribution?

  • In my opinion, one of the most awful features of Python as a language is the fact that it thinks of whitespace as a highly significant feature in programs, so much so that you always have to properly indent all blocks of code that are under a particular control structure (such as if statements or for loops), instead of having a 'begin'-'end' token pair to do the job, such as what we have in C, Perl, Java, Pascal, and almost every other block structured programming language I know about. It takes away the programmer's freedom to style his or her code, forcing them to conform to somebody's idea about how programs ought to visually look like. I've heard of "bondage and discipline" languages, but this is arch-B&D... not even Pascal is so anal!

  • by MAXOMENOS (9802) <maxomai AT gmail DOT com> on Tuesday April 17, 2001 @09:04AM (#285897) Homepage
    The Free Software foundation mentions the license that comes with Python versions 1.6b1 and later as being incompatible with the GPL. In particular they have this to say about it:

    This is a free software license but is incompatible with the GNU GPL. The primary incompatibility is that this Python license is governed by the laws of the "State" of Virginia in the USA, and the GPL does not permit this.

    So, my question is a two parter:

    1. What was your motivation for saying that Python's license is governed by the laws of Virginia?
    2. Is it possible that a future Python license could be GPL-compatible again?

    ObJectBridge [] (GPL'd Java ODMG) needs volunteers.
  • Python can spawn multiple threads, yes; but Python can only interpret bytecodes in one thread at a time. What does this mean? Well, your C extension can go about its merry way in its own thread, but the moment you are back in Python-land everyone is competing for the one and only interpreter thread. As I said in my initial post, it's a workaround. It is well known that Python does not completely and properly support threads, as Tcl (and many other script languages) does. Guido has said this is a deficiency in the current C implementation and will be addressed in the future.
  • Perhaps I could sum up the C API aspect of my question as follows:

    JPython rocks. Working with JPython in Java rocks. When, oh when, will the C implementation of Python compare favorably to the Java version?

    None of you who have worked with both C and Java Python can tell me that there is any comparison at the API level. JPython rocks all over CPython in that regard.
  • The Python interpreter is NOT multithreaded. The interpreter can manage, and spawn, threads, but when interpreting the Python P-code itself there is one and only one thread. A "global interpreter lock" is not multithreading; it's a kludge. Guido has said as much in the past, and hinted that this would be addressed in Python 3000.

    Reading the Zope list, and searching the archives, you will see Python's poor threading support bemoaned repeatedly as a roadblock to scaling a single Zope process on multi-processor systems.
  • by Xar (11113) on Tuesday April 17, 2001 @09:09AM (#285901) Homepage
    It's been a while since I've seen any mention of Python 3000--aside from the recent April Fools joke, that is. I love Python as a language, and use it both professionally and personally. But, Python's current implementation is lacking; the interpreter is not multi-threaded, causing large Python applications (such as Zope) to implement various workarounds that only partly address the problem; and the C API is rather...unpleasant. Working with the Tcl C API in an embedding situation is much, much better, IMO. Will Python 3000 address any of these concerns? Any information on a timeline, or current status?
  • It seems that you are personally involved in getting the new Python license GPL compatible so that Python can be mixed with code released under the GPL.

    The Apache people seem to have given up on ever getting their license compatible with the GPL since it seems to much work to get the language about the trademarked words right. So they just accept that people writing GPLed software can never use code distributed under the Apache license. It does not seem to hurt the Apache project to much.

    Why do you care so much about Python being GPL compatible? And what work should still be done?

    I loved the original CWI license by the way, it was short and to the point. Is there any way to get that back as standard license? And is was GPL compatible as a nice bonus.
  • First, I feel obligated to state that I think Python is totally tits. I love it. Thank you.

    Reading "What's New in Python 2.1", I'm curious about nested scopes. Given that name lookup has always been the cause of poor performance in Python (or so I've heard), it would initially seem that introducing nested scopes would further reduce the speed at which Python scripts run if you use this feature.

    Is there any information available about Python's performance with this addition?
  • by Xerithane (13482) <xerithane.nerdfarm@org> on Tuesday April 17, 2001 @09:06AM (#285904) Homepage Journal
    First off, as a disclaimer I have never actually written anything in Python. But, I have read up on virtually all the introduction articles and tutorials so I have a grasp on syntax and structure.

    I have been doing C development for 9 years now, and I know a plethora of other languages including shell scripting, perl, PHP (for scripts). Now, each language uses 'normal' grouping for control structures (if, for, etc).

    What was the logic behind creating a whitespace-based syntax rule? And why do you feel it is good, please refrain from the readability answer because that is all I get from those people I know who know Python.

    I find, because of my background, it is much easier to read code that uses braces ({}) than whitespace because my mind automatically looks for them. After maintaining legacy code that extends a life span of 20 years from it's first line of code, I have some concerns about the longevity of any Python code. So, my second question is, how well do you see Python holding up for 20 years and why do you think it will hold up that long?


  • Why include a lambda form at all in it's current, useless, state? Was some Schemer holding a gun to your head, or what?

    Oh, and many many thanks for a beautiful language.

  • More to the point, it's the primary reason (at least as far as I can discern) for not including the beautiful stackless patches into the core distro. Which is a shame, as Stackless Python is a thing of beauty.


  • It's a feature of some object-oriented-scheme-like languages (Dylan, ScriptX), that might be expressed in Python.

    Let's not forget the language that made multiple dispatch popular (for certain values of popular): Common Lisp. CLOS is still the best expression of multiple dispatch around (although Dylan sure is nifty.)

  • Note Exactely python now has ++, --, etc. how every they are a new feature, (1.6 or 2.0) thus older python implemenations don't have them. as a side now python is quite downward compatable I think it is almost 100% between 1.5.2 and 2.1 and not much is going to change between 2.1 and 2.2 which is much better than say JAVA 1.0 and 1.1, or even java 1.1beta3 and 1.1
  • What was the logic behind creating a whitespace-based syntax rule? And why do you feel it is good, please refrain from the readability answer because that is all I get from those people I know who know Python.

    Well, that's the answer, i'm not sure why it isn't acceptable. One of the main stated goals with Python is that they didn't want a language that had completely different formatting depending on who wrote it, so they made formatting part of the language. This makes it much easier for non-programmers (like me) and beginning programmers to pick it up.

    I find, because of my background, it is much easier to read code that uses braces

    but someone with no programming backgound wouldn't have that bias, so if you're inventing a new language, why feel hindered by older syntactic conventions?

  • Well, here's MY attempt to defend the whitespace-based syntax.

    Would you allow a junior programmer joining your team to write code like this?

    int someFunc(int x) {
    if( x < MIN_VALUE ) {
    return x;}
    else { int i, v, new_x=0; for( i=0; i<x; i++ ) {
    &n bsp;v = process(i);
    new_x += v;
    return new_x;}

    I certainly wouldn't. In c (or c++, or java, or whatever language that snippet was written in), whitespace is significant, but it is significant ONLY to the human reading the code, not to the compiler.

    If I could, I'd ask the compiler (or the source-code-control checkin program) to require that the indentation match the braces. That would prevent stupid stuff like the above, and also prevent the more insidious errors like this:

    int someFunc(int x) {
    if( x < MIN_VALUE ) {
    re turn x;
    } else {
    in t i, v, new_x=0;
    fo r( i=0; i<x; i++ )
    &n bsp;v = process(i);
    &n bsp;new_x += v;
    re turn new_x;
    where my failure to use {} around the for() clause has created an error (ie, new_x += v should be inside the loop, and you can TELL that from the indentation).

    So in python, you get exactly this feature. The language verifies that the ACTUAL block structure of the compiled code equals the INTENDED block structure as indicated by the programmer:

    def someFunc(int x):
    if x < MIN_VALUE:
    re turn x
    ne w_x=0
    i= 0
    wh ile i < x:
    &n bsp;v = process(i)
    &n bsp;new_x += v
    re turn new_x

    In my opinion, the python code will hold up BETTER than over long life spans than c-style code, at least if you consider only the effects of whitespace-based block structure. I say this because the use of indentation to indicate block structure is older and more widespread than the use of braces for that purpose -- compare Algol, Lisp, APL, and nearly everything else except early (column-sensitive) Fortran and assembler.

    Of course you ALSO said that you liked having the braces, because it makes it easier for you to read. Fortunately, braces ARE allowed in python! The following example demonstrates their use. ;-)

    def someFunc(int x): #{
    if x < MIN_VALUE: #{
    re turn x
    else: #{
    ne w_x=0
    i= 0
    wh ile i < x: #{
    &n bsp;v = process(i)
    &n bsp;new_x += v
    re turn new_x

    -- Michael Chermside

  • by Lumpish Scholar (17107) on Tuesday April 17, 2001 @10:07AM (#285911) Homepage Journal
    (It has to be asked every once in a while.)

    Any movement away from Tkinter, and toward something else, as the pretty-much-standard programming interface for graphical user interfaces?

    Any movement towards a Tk library that *doesn't* use Tcl?
  • (nor "State" for that matter)

    Good call.
  • Do *you* think you need an introduction?
  • I don't think that's feasible due to the runtime nature of Python. It would almost certainly require a type of "engine" to handle these features (think about all the cool things you can do with classes that aren't possible in C++), which isn't much better than byte code.
  • Programmers frequently praise Python for its generally clean syntax, its ease of use (as in CP4E), and its "Batteries Included" philosophy. As a PyUser myself, I have to agree that these are fantastic features.
    I'm interested, though, in the conflict between library design and legacy habits. Would you suggest that C-isms like "popen2", "execlp", "sys.argv", and "socket(AF_INET, SOCK_STREAM)" really belong in a "clean" language in the long run?
    Is there any chance that we'll someday see a standard library extension that (much like Java's libraries) allows developer to program in English rather than UNIXglish?
  • by jilles (20976)
    I've recently come across Ruby and I must say that at first site it has all the features (after all it was inspired by python) that made python a success and some more. What are your impressions of this language and would you be willing to give up python for it or some other language?
  • There is a tool (whose name I've forgotton, unfortunately), that adds #{ and #} to python files as block delimiters (consider it as optional bracket syntax).

    The tool can convert back and forth to allow for use in indentation hostile environments. :) Can anyone give the name of the tool I'm thinking of?
  • I have seen a couple of Java based Unified Modeling Language [] tools but no Python support or implementation. It would seem natural to develop in python based on UML, so this must be a large gap in the python suite. What do you think of designing with UML and implementing in python?

    Thorn [] is an opensource UML editor written in Java with JPython scripting but no python code generation.

    ArgoUML [] is an opensource UML editor written in Java with no current python code generation

  • i notice that the release of 2.1 isn't signed or even checksummed. what will you do in the future to ensure that mirror sites don't supply versions infected with macro viruses or changes to the core? nobody has time to audit a 4 meg download, and everybody runs "configure" right away (downloading an executable and running it -- idiotic, eh?).
  • I have some friends that are extremely talented programmers that find python very difficult to use because it relies on whitespace for code block delineation. They're blind. With that in mind, is it still a great idea to use whitespace (not braces) just so the code will all look alike?

  • It was Guido's former employer, CNRI, which caused the license to be covered under the laws of Virginia. As Guido was working for CNRI, CNRI holds the rights for a significant portion of Python. If and when there is no longer any CNRI code in Python, most likely the license will be changed to something less GPL unfriendly.
  • Have you actually used Komodo?

    even compared to Visual Basic it really sucks... like.. REALLY sucks. I would be embarrased to ship a program built with that IDE builder.

    Would you like a Python based alternative to PHP/ASP/JSP?
  • by Pengo (28814) on Tuesday April 17, 2001 @08:18AM (#285927) Journal
    Do you know if there are any projects on it's way to compete with Kylx or Visual Basic based on Python.

    Would you like a Python based alternative to PHP/ASP/JSP?
  • by debrain (29228) on Tuesday April 17, 2001 @08:16AM (#285928) Journal
    Is it possible to make Python as fast as C/C++? In particular, is there a way to compile Python into system code (as opposed to byte code)? If there isn't, will there ever be?
  • I will admit that that method, dating back to Ada, Pascal, and the venerable ALGOL, does have its merits for long pieces of code. Typically, it enforces spacing as well, since single-line statements just don't look right in this style of language or may be against the grammar.

    I like Python's style and I like the ALGOL-derived style as well.
  • There have been a few other posts about this. See my earlier response to another post about that here [].

    I'll answer your post with a question. Why should a programmer be free to style their code anyway they want?

    The point of a programming language is to communicate instructions to a computer in human-readable format. I still maintain that people who complain about being forced to write readable code are the ones who write the most unreadable code. TMTOWDTI, Perl's motto, is the main reason that when the first Obfuscated Perl Code contest was announced, many Slashdotter's joked that it was redundant. Consistent style is a necessity for maintainability. It helps to allow others (and maybe yourself a year later) to understand what you were doing when you wrote a piece of code.

    Honestly, if you think everyone should be able to write code in whatever style they see fit, you've never worked on a large project before. Languages with more freedom just force people to place other personal standards on how their code must be formatted within their project or lose readability of code as multiple programmer styles conflict. Python forces a consistent standard across all development, making it an great relief to maintain.

  • by Valdrax (32670) on Tuesday April 17, 2001 @11:09AM (#285942)
    What was the logic behind creating a whitespace-based syntax rule? And why do you feel it is good, please refrain from the readability answer because that is all I get from those people I know who know Python.

    I fail to see why there would be any other reason. Furthermore, I fail to see why there should even need to be a better reason. Why do you have whitespace at all? There are only 2 real answers: easy parser writing and human readability.

    Python's style makes it easy to see blocks of logic. It also forces you to think about how your code is organized by exposing these blocks to you at all times. Braces, parentheses, brackets, etc. are easy to lose track of in complex single-line statements. You have to spend too much time thinking about whether or not you've got your puncuation matched up properly. Python eliminates this confusion by exposing logical blocks. Besides, properly formatted and readable C code should already be spaced out like a Python program. Python just eliminates the redundant punctuation.

    Typically, the people who complain the loudest about enforcing spacing in syntax are the same people who write those tangled, dense, single-line statements in C and Perl that inspired their respected obfuscated code contests. You don't need the ability to cram 5 lines of Python in 1 line of Perl. It just hurts maintainability, and there's really no compelling argument for keeping source code dense and compact anymore if it doesn't add speed and remove bloat. (Forgive me if I have unfairly tarred you with this brush, but this has been my general experience.)

    (In response to another post:)
    Also, I've never seen a source-control system mess with the spacing of a file before. That's just odd. Be consistent with using either spaces OR tabs and your Python code will be much easier to store. I'm not saying it doesn't happen. I'm just saying that bugs in certain tools that weren't written with Python in mind shouldn't be a black mark against the whole language.
  • I recently started using Python and I now use it for everything, including an interactive shell. Thanks.

    One thing I see in Python-land is that there is a tendency to implement everything in Python. I just submitted a design for a program that I want to write in Python and I used Sketch for diagrams, despite having CorelDraw for Linux right here. It seems that no Python programmer is happy with having a component written in another language. Many other languages don't interoperate as well as in Python but those programmers seem happy with mixing Perl and C, for example. Look at Zope. It has it's own web server. I know it's faster, but why do you think this is happening with Python specifically?
  • What use of Python have you found that surprised you the most, that gave you the strongest "I can't believe they did that" reaction?

  • No, Python is interpreted, it is a tradeoff, the speed you get from python is in terms of development, for very large projects you might see one line of python code for ten lines of C code or one line of python code for 5-7 lines of C++ code. Take your pick. Nevertheless, Python has an API that allows you to invoke functions written in C and C++. As programmers, you know the 80/20 rule, 20% of your program takes 80% of the time, you can code that 20% in C/C++, and the rest in python, then with the API call that 20%, thus saving incredible time in development.

  • I personally believe that Python bytecode is very efficent, take the case of Java bytecode, it is just as efficent, to over optimize it will put very serious constraints on what python does and can do, it is a simple tradeoff, the speed in python all lies in RAD.
  • From someone who has used the Python C API, I have to disagree with you, in no way do I find it unpleasant, it is so easy to work with, compare it to Perl C extension API, I haven't played with TCL C API, but you might be comparing apple and oranges, as we know tcl and python are on two different levels. A comparasion with Perl and Ruby will be more appopriate.
  • You have all the data structures you need, perhaps your skills are lacking. A list is already given with you, with that list you can implement a stack, queue, etc using the existing methods! A dictionary is given to you, with which you can implement a set/hashtables, the only thing that isn't native is trees, which is very easy. What I will like to see tho, is a Collections framework just as in Java, that would be cool.
  • I think it is very wonderful, I can't stop talking about JPython, the best of both worlds, mixing python and java, coding with pythons wonderful fun interpretter environment, and getting to use stable java libraries. Saves so much time with the java compile cycle, yet, you can write your java servlet or a swing application in python. Guido probably doesn't care, his work is python, but this shows java's potential.
  • Does the license matter? It is not like your source code or bytecode is subjected to the license. I say, if you are so GPL nuts, let the free software foundation write their own implementation.
  • I dobut he will, stackless is amazing, but I dobut it, I mean, there are many python implementations with many nifty things, how about Jpython? When does it stop? But I sure look forward to it tho, as far as it doesn't bloat it. :)~~

  • by segmond (34052)
    The most annoying thing in python for me, is when walking with classes, and having to reference all local variables and method with the self class. self.It self.gets self.annoying self.and self.I self.believe self.reduces self.readability. As you can see, lots of "selfs" over your screen is annoying, I know that there is a hack out there to remedy that, but is there any immediate plan to fix that in the future?

  • by segmond (34052) on Tuesday April 17, 2001 @10:41AM (#285953)
    Is there any plan to adopt wxPython ( as the GUI standard? The Tk interface looks butt ugly, and I am sure you have heard that many times, what do you see as the advantages and disadvatages for using wxPython over Tinker?

  • This is the greatest contribution of the perl community. A true example of code re-use.
  • I prefer the (new) php style of using keywords instead of spaces or brackets. I find it easier to keep track of multiple levels of indentation.
    if ($foo):
    while ($fee):
    buncha code

    If the buncha code fills up your screen it's easy to grok what's going on at the end of control structures. For every control structure php provides a keyword for the end.
  • "Wow. Imagine a Beowulf Cluster of morons."

    Oh man I can't resists...

    Where I work
    United States of America
    Washington DC

  • by wiredog (43288)
    Komodo []At Active State [].
  • by seanw (45548) on Tuesday April 17, 2001 @08:15AM (#285964)

    How do you see the relationship between jPython (the java implementation) and standard cPython (the original C language version) evolving? And do you see the advantages of either one (i.e. portability vs. speed) becoming especially pronounced in light of the recent trend toward distributed software (ala the MS .NET initiative)?

  • Python is a really snazzy language, but one limitation to its flexibility is the absence of static typing.

    It seems that (optional) static types would enable the creation of practical python compilers, and could also provide much more in the way of pre-execution error detection.

    I've heard rumours that static typing might be in the pipeline for Python. Is there any truth to these rumours, and if so, how might they be implemented?

  • by po_boy (69692) on Tuesday April 17, 2001 @08:08AM (#285975) Homepage
    First, thanks for Python!

    I have long wondered why the value None in python is named "None". It's pretty common in other languages to call that thing (or something very similar) "NULL". Were you trying to differentiate None and NULL in some way, or do you just like the way "None" sounds as you read code?

  • by po_boy (69692) on Tuesday April 17, 2001 @08:25AM (#285976) Homepage
    One of the reasons I still write some things in PERL is because I know that I can find and install about a zillion modules quickly and easily through the CPAN repository and CPAN module. I'm pretty sure that if Python had something similar, like the Vaults of Parnassus [] but more evolved that I would abandon PERL almost entirely.

    Do you see things in a similar way? If so, why has Python not evolved something similar or better, and what can I do to help it along in this realm?

  • I know first hand that you have an uncanny ability to remain calm and level-headed. I emailed you once a flame--it even contained several instances of "DIE"--that while done with some jest was quite insulting. Your reply was not only civil but combined with the fact that you even bothered to reply left me feeling quite embarassed (and rightly so).

    What is your secret for keeping your cool when discussions get heated, particularly when techies tend to be very loyal to their causes?
  • Don't you ever get sick of talking about Python? Are there times when you just want to talk about your collection of scotch-tape dispensers or your love of shuffleboard? Do you feel sad that people pidgeonhole you as "Guido van Rossum, the Python guy" or "Guido van Rossum, the father of Python", when perhaps you'd prefer to be "Guido van Rossum, the ballet dancer" or "Guido van Rossum, the Buddhist monk"?

    In all seriosity (?) are there times when you overload on Python and just have to get away for a while?

  • by Rexifer (81021) on Tuesday April 17, 2001 @08:15AM (#285982)
    How closely does the primary Python development team interact with the other language binding efforts (mainly Jython)? Anyone who's hung out in Slashdot seems to have a rabid attachment towards their native tounge, so to speak, and I've found it refreshing that the Python community "plays nice". Is there a lot of cross-pollenation(sic) between the groups?

  • Python is about simplicity of expression. When doing advanced programming it is nice to have a large set of encapsulated data structures. After coding a data structure 5-6 times it ceases to be instructional.

    You can look up data structures anywhere, but developing clean algorithms takes "skills".

  • by GrEp (89884) <[moc.liamg] [ta] [200brc]> on Tuesday April 17, 2001 @08:15AM (#285989) Homepage Journal
    I love python for making quick hacks, but the one thing that I haven't seen is a comprehensive data structures library. Is their one in development that you would like to comment about or point us to?

  • I've actually written a porn bot in python/glade ... "beelzebot" ... :) Its an automated usenet porn downloader/decoder.
  • by patnotz (112592) on Tuesday April 17, 2001 @08:27AM (#285999)
    While I love many things about python, one thing bothers me. It seems that it's the interpreter that defines the language rather than a language defining how the interpreter should act. One example is the "++", "--" etc. operators and another is the functionality path module. I often write scripts with 2.0 at work only to find thy don't run on my Debian 2.2 system at home (stable Debian's python is 1.5.2).

    Are there any plans to set some kind of standard language specification that will hold for a while?

  • I learned Python from the "Internet Programming" book, and I love the lean and mean Python. Because of the recent burst of development, it seems to me that lots of new features are being stuffed into the language. Is my perception correct? Is there a danger that Python will forfeit two of its best features, simplicity and cohesiveness, merely to appease a few complainers?
  • In the upcoming volume IV of Knuth's TAOCP, Seminumerical Searches for Hoaring Triplets, he conjectures the following about the Theory of Programming Languages:

    Computers are ultimately a tool for their users. As such, modern programming languages should allow the users of these inherently unwieldy tools means of effective and efficient control. In volume 1 of TAOCP, we proved that all computer uses fall into one of two categories: searching for pr0n and viewing pr0n. Thus, any modern computer language should have as its core requirement the shortest semantic expressions for searching and viewing pr0n. (As usual, I will cheerfully pay $2.56 to the first finder of any pr0n that I have not already found).

    Along these lines, I pose the following questions:

    1. How effective is python at finding me pr0n?
    2. What is the shortest python program that you know of that will allow me to veiw such pr0n?
    3. Do you have any good pr0n?
    4. Of Donald Knuth?

    Thank you.

  • There are a lot of "golden Python rules" or whatever you would call them, like "explicit is better than implicit", "there should be only one way to do it", that sort of thing. As far as I know, those are from old posts to the mailing list, often by Tim Peters, and they've become The Law afterwards. In the great tradition of Usenet advocacy, people who suggest things that go against these rules are criticized. But looking at Python, I see a lot more pragmatism, not rigid rules. What do you think of those "golden rules" as they're written down?

    What's your idea of the future of Python? Since the PEP process, a lot of new feature ideas have been put forward, and a lot of people feel uncomfortable with quick change to a good language (Python 2.1 is again excellent though, congrats). Do you think or hope Python will be finished one day? If not, isn't the alternative an endless string of added features? "Python 3000" was an idea of a sort of ideal Python that would be worked on, but as I understand Python will now evolve more gradually.

  • Every language has good and bad points. The python documentation and website contains a lot of advocacy articles, which point to all the goot things (TM) in Python, but no mention of the bad points.

    What are the bad points in Python ?
    What kind (or size) of project should not be written in Python ?
    When should one use a different language ?


  • Every other programming language I know uses some kind of visible token to indicate code blocks -- curly brackets, begin/end, whatever. Good programmers indent nested blocks, but that's just to make the code easier to read. The parser doesn't care.

    In Python, indentation is the token. What's the rationale for this? Do you get a lot of flack from people who prefer the old-fashioned way?

    Oh yeah, and since the Vim % command (jump to matching bracket) doesn't work with Python code, do you know of a macro to replace it?


  • Brilliant! Use map to call a macro! Never would have occurred to me!

    Now all I need is a macro to actually do the function I described! That's the easy part, of course!


  • Follow-up questions are:
    What's your favorite Monty Python Quote and Movie?
  • (1) who would win in a fistfight, you or larry wall?
    (2) who speaks more languages, you or (that supposed linguist) larry wall?
    (3) if you could eliminate one planet from the solar system (besides the earth), which one would it be?
    (4) how do you feel about the euro?
  • What Is Python? Understanding Them : Pythons are relatively primitive snakes belonging to the subfamily Pythoninae within the family Boidae. Boidae, in turn, is one of 11 families in suborder Serpentes, The Snakes. Within the subfamily of pythons, arboreal pythons with heat-sensing pits along their lips (green tree pythons) are grouped seperately from terrestrial pythons that have heavy bodies and short tails
    (blood pythons). Each different type of snake eventually ends up with two names, one for the genus and one for the species. When isolated population exits that are still identifiable as the same type of snake, a third name, the trinomial, is added.
    The term primitive indicates that these snakes were some of the first snakes to evolve. Primitive snakes display features that link them to lizards. These features include a rudimentary pelvic girdle in the form
    of cloacal spurs, and lungs of equal sizes . Advanced snakes, like the rat snakes and whip snakes, have only one functional lung and no cloacal spurs.
    Pythons are divided into about 26 species, depending on which authority you accept. Pythons range in size from very big (Burmese and Reticulated pythons with the potential of over 20 feet and over 200
    pounds) to small (Children's pythons do not get much bigger than 24 inches '61cm' in length) . No matter what the size , they are all constrictors . Some burrow-hunting species have developed novel ways of
    using thei coils to catch prey within the confines of a burrow but they are still constrictors nonetheless . Most pythons are nocturnal hunters and some species have heat sensory pits along the edges of their lips
    to aid in finding warm-blooded prey .

    Pythons Versus Boas : Many people don't know the main difference between boas and pythons . Boas are termed ovoviviparous , this means their eggs inside the females are surrounded by a membrane
    instead of a hard shell like pythons . So when the boa babies born , babies break through the membrane to crawl away . Pythons are oviparous , this means the eggs are surrounded by a thin , parchmentlike
    shell . Female pythons will coil around their eggs and stay with them during the incubation period .

    Life Span : Over 20 Years But Much More In Captivity .

    Their Orginal Habitat : Africa , Asia and Australia , North America (A Little Amount) . All python snakes are tropical animals . They won't live under the temperature 22C (77F) . Under this fact the areas
    that they live must be near the deserts or tropical places like amazon . But mainly you can find them Africa , Asia and Australia , North America (A Little Amount) .

    Pythons As Pets : Pythons can be good pets if you care them enough .
    -- Floyd
  • Life Span : Over 20 Years But Much More In Captivity
    Wow, 20 years - it's pretty long for a programming language! I just hope that Guido will always keep it in captivity.
  • ... what book would you bring, what CD would you bring, and what language OTHER than Python would you be using (assuming that your computer suvives the sand and wet.)

  • How closely does the primary Python development team interact with the other language binding efforts (mainly Jython)? Anyone who's hung out in Slashdot seems to have a rabid attachment towards their native tounge, so to speak, and I've found it refreshing that the Python community "plays nice". Is there a lot of cross-pollenation(sic) between the groups?

  • Tim Peter's The Zen of Python [], reads:

    Beautiful is better than ugly.
    Explicit is better than implicit.
    Simple is better than complex.
    Complex is better than complicated.
    Flat is better than nested.
    Sparse is better than dense.
    Readability counts.
    Special cases aren't special enough to break the rules.
    Although practicality beats purity.
    Errors should never pass silently.
    Unless explicitly silenced.
    In the face of ambiguity, refuse the temptation to guess.
    There should be one-- and preferably only one --obvious way to do it.
    Although that way may not be obvious at first unless you're Dutch.
    Now is better than never.
    Although never is often better than *right* now.
    If the implementation is hard to explain, it's a bad idea.
    If the implementation is easy to explain, it may be a good idea.
    Namespaces are one honking great idea -- let's do more of those!

  • by janpod66 (323734) on Tuesday April 17, 2001 @08:35AM (#286065)
    What are the plans for the compilation of Python code to efficient executables? Python's object system is very dynamic, allowing anybody to add instance variables and methods to any object at any time--how are you planning on dealing with that during compilation? Performance-wise, how do you expect Python will compare to compiled CommonLisp or Smalltalk, which evolved along similar lines 20 years ago? And will there be a language standard, or will Python continue to be defined by what the C implementation does?
  • Technically Virginia is a Commonwealth, not a State (nor "State" for that matter).


Take care of the luxuries and the necessities will take care of themselves. -- Lazarus Long