Magnum opus creative solution for I Love Python! by veky - python coding challenges

Enable Javascript in your browser and then refresh this page, for a much enhanced experience.

Magnum opus solution in Creative category for I Love Python! by veky

(lambda _:globals().__setitem__(_.lower().translate(dict(enumerate(( '_'
,None),1<<5))),lambda:_))("""\
In my wildest dreams, I never imagined I'd have to write _essays_ here
on CheckiO. Yes, I know: who's talking? Really, I've written many mind-boggling
things here already, and I have puzzled many folks reading my solutions.
It's only fair that I puzzle myself now. :-\
So, what to write? You have already seen enough to realize that I consider
Python one of the most excellent inventions our civilization has ever made, not
only in the tier of programming languages. It still puzzles you: why one would
think so? Puzzling is good, because in this jaded culture, it is one of the
rare ways ideas can still be learned. Of course, my essay won't magically
enlighten you. But neither will Zen of Python by itself. Those are only small
chunks of an infinitely rich mosaic that's waiting for you to discover it on
your terms...
While you are trying to "digest" this essay, and wondering if it is an
essay at all, think: we must know, we will know. Good luck!"""[::73])

Aug. 22, 2014

Comments:

veky on Aug. 22, 2014, 10:10 p.m.
<ul> <li>Note the smiley at the end of line 1. ;-)</li> <li>Find how many Python keywords are in the essay. ;-)</li> <li>What's the connection between 73 and CheckiO (for old ones here;)?</li> </ul>

veky on March 3, 2015, 8:27 a.m.
... And I must note the amusing fact that right now it has exactly 42 likes. :-D

kryksyh on April 13, 2015, 8:34 p.m.
Oh, so sorry for ruining this..

veky on April 13, 2015, 9:52 p.m.
No problem. :-) Maybe I can get it to 153 likes. :-D

bryukh on April 14, 2015, 4:34 a.m.
<blockquote> What’s the connection between 73 and CheckiO (for old ones here;)? </blockquote> I'm need a hint :-)

veky on April 14, 2015, 6:42 a.m.
You need a hint? Too much šljivovica? :-P <a href="http://www.checkio.org/mission/the-best-number-ever/">http://www.checkio.org/mission/the-best-number-ever/</a>

bryukh on April 14, 2015, 7:54 a.m.
Oh, I forgot about this Oduvan's mission.

bryukh on April 17, 2015, 2:59 p.m.
...and I should return this mission back :-)

veky on April 17, 2015, 4:46 p.m.
Absolutely. It was the best mission ever. Almost as good as 73 itself. :-DD

veky on Jan. 14, 2016, 8:08 p.m.
...and right now it has 73 likes. Wonderful. :-D

viktor.chyrkin on Feb. 19, 2022, 2:26 p.m.
Nice! I never noticed that!

ciel on Sept. 2, 2014, 4:53 a.m.
God. I love it. Maybe you can beat Ruby TRICK contest ^^

veky on Sept. 4, 2014, 1:36 p.m.
Maybe. It would help if I knew what that was. :-D

ciel on Sept. 4, 2014, 3:18 p.m.
See <a href="https://github.com/tric/trick2013">https://github.com/tric/trick2013</a> . Perhaps I should have referred IOCCC instead :p

veky on Sept. 5, 2014, 4:58 a.m.
Perhaps. At least it's googlable. :-) Look, I see no point in writing obfuscated C. It's easier than writing nonobfuscated one. :-D I don't know Ruby in such depth, but I suppose that's also true there. Python, on the other hand, is where "(un)readability counts". ;-D

nickie on Sept. 4, 2014, 9:39 p.m.
The trick was great! The essay, on the other hand... :-)

veky on Sept. 5, 2014, 6:27 a.m.
Once you know the trick, you're welcome to make a better one. ;-)

veky on Sept. 10, 2014, 10:45 a.m.
Well, I hope you like <a href="http://www.checkio.org/mission/i-love-python/publications/smilicic/python-3/first/">this essay</a> better. It's a much more usual personal story of a relationship with Python. ;-)

zlomovsky on March 3, 2015, 4:04 a.m.
Thank you very much for this awesome idea! I've created the "<a href="http://www.checkio.org/mission/i-love-python/publications/zlomovsky/python-3/my-star-wars/">Star Wars</a>", which is based on your code.

veky on March 3, 2015, 8:25 a.m.
Well, you took only "function name is processed return value" from it. It would be much more funny if you tried to take the [::73] idea. ;-]

zlomovsky on March 3, 2015, 8:36 a.m.
The sad thing is, I don't know the background of "73" story. I was trying to search a bit, but found nothing.

veky on March 3, 2015, 10:44 a.m.
There once was a CheckiO mission of "What is the best number?". It was inspired by <a href="https://www.youtube.com/watch?v=TIYMmbHik08">this</a>, of course. (That's why it's exactly 73 up there. But you can do the same idea with another stride.:)

zlomovsky on March 3, 2015, 1:59 p.m.
I will try later. :) By the way, is there is any way for direct modifying "globals" dictionary? For example, I can do this: <pre class='brush: python'>d = {'a':'b'} d = __import__('collections').defaultdict(lambda:lambda:"I love Python!", d) print(d['blah-bla']()) # I love Python! print(d['blah-bla-bla']()) # I love Python! print(d['a']) # b </pre> if I would modify "globals" in such way, I could write a solution for this challenge where I don't need to define the name for the function.

veky on March 3, 2015, 5:21 p.m.
You are asking dangerous questions. :-D You don't want to modify globals() object, you want to modify it's type. And that's a lot harder to pull off. Note that in your second line, name d is rebound to a different object than the d in line 1. It's kinda like saying "I don't want ord to be a builtin function, I want it to be a class. It can still be called like before, of course." It's hard to even think how that would work. :-D

zlomovsky on March 3, 2015, 7:32 p.m.
Yes, I know, in this case, `d` transforms from `<class 'dict'>` to `<class 'collections.defaultdict'>`. I just had an interest if it's possible to do the same thing with 'globals'. :) I like do experiments. They give me better understanding of things and their bounds. Also, while experimenting, I learn things a lot faster. I see no reasons for preventing `ord` to be a class if it will entertain somebody somehow. This is an example of "ilovepython" <a href="http://www.checkio.org/mission/i-love-python/publications/Arandomusername/python-27/class-way/">as a class</a>. I just wanted to know, is it possible to modify `globals` in a such way? :) I'm not talking about moral aspects of the problem. As for me, it's like talking about differences between painting in oil and painting over printable coloring pages.

veky on March 4, 2015, 8:38 a.m.
I understand you're not talking about morality. But I don't think you understand quite well what are you talking about. :-) First, there is a fundamental difference between rebinding a name to a different object, and changing the object bound to a name. It's not a problem to write `class ord` to behave the same as a builtin. But it will be a different object ("different" not in sense of "changed", but "another"). `id(ord)` will change after you write it. What you're asking for is much more complicated: you ask if it's possible for an object returned by globals() to change its type, yet retain its identity. Of course, you might say, "I don't necessarily want this - if there's a canonical name for globals(), let's just rebind it." There the problem is "canonical". You see, there is no such thing as "THE globals namespace" in Python. Every module has its namespace (stored in a module`.dict` attribute), and from its perspective, it IS "the" globals. Ok, so you say, can we `import something` and then rebind `something.dict = defaultdict()`? No cookie, says Python: `dict is a read-only attribute`. Which makes sense in a way: after all, module's `dict` is where its attribute lookup is usually performed. Rebinding that attribute would invalidate many assumptions of functions like `getattr` and others. Ok, so you go ballistic: let it stay the same, it can keep its hashtable and everything, I just want to graft on the default_factory from defaultdict. <pre class='brush: python'>something.__dict__.__class__ = defaultdict </pre> Then you get a much more obscure error message, `TypeError: class assignment works only for heap types`. That's another ugly truth: native Python types have a certain layout in memory that's optimized for quick retrieval and memory usage (read about `slots` for an aspect of that layout controllable from Python), but is not really pluggable into the rest of "custom defined classes". One user-defined class can be transplanted onto instance of another user-defined class (usually, and with much gnashing of teeth;), but this charade doesn't work with "native" types whose implementations are a bit cut down in the name of performance. (dict doesn't seem so important performance-wise, until you realize that all name lookups everywhere except in local scopes of functions, and half of parameter-passing when calling functions, happens via it.) Ok, there could be a third way. dict has a hook for handling missing keys, it's a special method aptly named `missing` (by the way, it's how defaultdict works:). But as defaultdict shows, it is really meant to be implemented in subclasses of dict, not in dict itself. Trying to assign <pre class='brush: python'>something.__dict__.__missing__ = whatever </pre> will just tell you `AttributeError: dict object has no attribute missing`. It has a hook, but nothing to hook on it. :-) So, sorry but no. To complete the story, I must say that classes had a bit of same problems with their namespaces, but there people were much more motivated to customize the creation of a namespace, even more because they already had a natural place to add it: metaclasses. So in PEP 3115, metaclasses protocol got one more hook, `prepare`, which by default returns `dict()`, but can return whatever it wants, where class statement will be executed. (One important case is when it returns `OrderedDict()` - that way class doesn't lose the order in which things were defined inside it. `Enum` from Py3.4 uses that.) Why am I telling you that? To make you see that in principle, such things are possible. You just have to invent "metamodules", that will hook into `import` statement the same way metaclasses hook into `class` statement, and customize their creation. And then you have to write a PEP and persuade Guido to invent/accept a nice syntax for that (classes have it easy, since base class list naturally can accept keyword parameters). My suggestion: <pre class='brush: python'>import something in defaultdict(whatever) </pre> Of course, after that you can forget about `something` and do what you originally wanted: <pre class='brush: python'>import __main__ in defaultdict(lambda:...) </pre> But please don't. Even if you persuade Guido, it would uglify the language and degrade performance, with a really not justified gain. :-)

zlomovsky on March 4, 2015, 11:03 a.m.
Things you've written about, I was trying to do (or have obtained from documentation) before I've asked the question. If I will try to describe what I was trying to do before asking the question, we will read the story which is comparable to the Leo Tolstoy's "War and Peace" in its volume. :) We could save a lot of time by answering just "no" when the answer is "no". :) Also, right now, I'm not ready to discuss inventions and PEPs because I have no suitable competency yet. But I describe how I see your proposition to write the PEP (exactly the PEP about meta-modules): it was your idea, you have described your idea to me and then you have asked me do not do things you have described. Please do not do that. I didn't want write PEPs at all, why you are giving me advice do not do that? I feel myself like a mannequin which has been dressed up with a pink trousers by its owner and then criticized by its owner for wearing pink trousers. It is not a good approach to make a discussion.

veky on March 4, 2015, 11:47 a.m.
I think most of our misunderstandings come from language problems, but I'm not quite sure. I'll try to continue discussion in that direction. First, about "saving time": don't think I write this only for you. Many more people than just two of us read this thread. About 150 people follow me here. At least 50 of them read all the comments I write. ;-) Second, I'm sorry you feel that way, but it wasn't my intention to dress you in pink trousers. I told you how exactly things you want can be accomplished in principle. I think it's instructive, at least potentially. And then I told you my personal opinion of that potential change. Since you are so ambitious that it's possible you actually write that PEP, I just wanted not to be responsible for that. :-) If you wanted to say that writing a PEP is a stupid way to accomplish what you wanted, I'd like to know if you know a better one.

zlomovsky on March 4, 2015, 1:03 p.m.
I was interested, is it possible or not, because I can write another creative solution for this task if it's possible. If no, I just will not try to do that. I see no problem here. I see no reasons to do that if the language give me no chances to do that. Again, I was trying to do that for learning purpose, not for language modifying purpose. There are a lot of differences between these two approaches. While I want just to see language bounds, you are still thinking I want to break this bounds. When you imagine situations and then imagine how I would act in this situations, it's exactly criticizing me for wearing pink trousers which I'm not wearing actually. Was it your intention or not, just do not do assumption on what I was trying to say. If you do not understand all I wanted to say or you have some doubts about meaning of some parts of my text, just ask for details, do not imagine this details without me. Maybe somebody else will not notice this or will swallow such a behavior, me not. <blockquote> Many more people than just two of us read this thread. </blockquote> Right now you're talking with me. If you want write for the auditory, you could write a monologue, instead of contributing to the discussion. <blockquote> If you wanted to say that writing a PEP is a stupid way to accomplish what you wanted, I’d like to know if you know a better one. </blockquote> I'm not wearing pink trousers. I wanted to say that writing PEP is a stupid way to learn Python. To learn Python you need learn it and discover its abilities and bounds. Writing PEP is a way to extend or limit language bounds which is far from learning process. I think it's an obvious thing. Anybody can see such a difference. So when you are trying to imagine that I wanted to say something other, it's like wearing something on a mannequin which will endure anything. Again, if you want write monologues, I just can leave you here, alone. If you want to discuss, please describe your own original thoughts, not thoughts about imagined meaning of my words. <blockquote> Since you are so ambitious that it’s possible you actually write that PEP, I just wanted not to be responsible for that. :-) </blockquote> I think, in this our situation, you do not see difference between "a wish to know, is it possible or not" and "a wish to accomplish this thing at any price". Writing such a PEP is your idea, I don't want this idea to be applied to me at all. I do not want the "you are so ambitious that it’s possible" will be applied to me as well. It is an imagined thing. If you think that it is not imagined thing please describe, which facts forced you to think so.

veky on March 4, 2015, 2:13 p.m.
First, about auditory: sorry, but too many times I've seen that the thing I wrote to somebody was read and was useful to somebody else. If you don't want to participate in the discussion that way, it's your right, but keep in mind that CiO doesn't have private message feature (yet). I think it's a signal that we should try to discuss things keeping in mind that wider context. And I think you're too easily offended. If it's so obvious to everybody else that you didn't want to see if your cute idea was possible at all, just if it's possible in a very specific "CPython 3.3 - CheckiO edition", and I am the only one that got that stupid idea, then what's the problem? Other people will easily see you're right and I'm wrong. Of course, I (in my subjective opinion) don't think it's so clear cut. You're asking me to explain why. I'll try. In your "by the way" post, you asked a question that's really deep and demands a thorough knowledge of Python internals. IMO, it's very hard to imagine a beginner asking such a question. If I read you right, you say that it was just an accident, you were playing and stumbled on this. Ok, let's go from there. I try to brush it off easily, saying it's a dangerous question, and it would break a lot of assumptions about how internals work. If the previous event was just an accident, this surely would be enough. I thought every beginner would get the hint. But not you. You maintained that what you want is sensible and understandable, and directly said you don't care if it's right or wrong, just if it is possible in any imaginable way. So I accepted the challenge. I tried to think of every possible way someone might approach such problem, and explained why it didn't work. (You speak about War and Peace length... sorry, but at least one of us had to write that text. Otherwise how would we know we're speaking on the same level of detail?) So, what's the next step? For someone really determined, it's the process of "writing a PEP" (although called that way, it doesn't really begins with writing the PEP document - it involves discussing the idea on various mailing lists and so on). I don't know why you are so aversive to that idea. I would understand if you just said "No, I don't want to do that", but to say things in style of "how could you POSSIBLY even THINK that I would do such a THING" is overreacting. [Imagine if I told you "Yes, you can have globals() as a defaultdict, but you must first install a special extension to CPython and say `from future import customizable_vars` at the top of your code". It would be completely fine if you said "No, that's too much for me". But would you react this violently?] Official documents say, more or less, that if you tried using current language features to accomplish something you want, and it didn't really work, and you asked knowledgeable people if you missed something, and they told you that it's really impossible with current language, you should write a PEP - if only so that it could be rejected and the matter put ad acta. And many much more bizzare PEPs have been accepted just because people demonstrated a real need and showed why current language features don't fulfill that need. Py3.5 will have a core language support for an operator that no builtin type implements (nor intends to implement), just because a fraction of mathematicians were successful in that process. Maybe my biggest problem, or rather the great difference between two of us, is that I've known Python for a long time, and I've seen it change in front of my eyes. In my view, those bounds you're speaking about are not so firm. After all, Python here on CiO is different in many details from Python on my computer at home - even as core (yield from, Enum, max(default),...), not to mention stdlib. Bounds surely do exist, and breaking them is impossible while we have BDFL - for example, you cannot add braces to Python. :-D (That example is most often quoted because it's funny, but there is more serious stuff that's also off the table.) I just wanted to tell you that customizable modules `dict` is not such a bounds-breaking idea. It's completely imaginable how to solve it in current language vision, it's just that nobody has of yet demonstrated a strong enough need to prevail over the need for performance.

zlomovsky on March 4, 2015, 4:18 p.m.
<blockquote> [Imagine if I told you "Yes, you can have globals() as a defaultdict, but you must first install a special extension to CPython and say `from future import customizable_vars` at the top of your code". It would be completely fine if you said "No, that’s too much for me". But would you react this violently?] </blockquote> CiO is giving me no chance to use such a possibility. But I've writing twice that I need to know it also for "writing another creative solution for this task". Here on CiO. (But mainly for learning purpose, of course.) I do not like when somebody is trying to pervert my words. This is what I was writing about. Right now you're acting better, because you write your own thoughts, not your thinkings about imagined sense of my words. When you show your point is awesome; when you use your point as an answer to the pink trousers is not good. <blockquote> I would understand if you just said "No, I don’t want to do that", </blockquote> Why I need to say so? It was your idea about PEP, not mine. Your idea was based on wrong thinking that I want extend Python. But I don't. I'm trying to learn it, not trying to change it, because such things will break my way of learning, will harm my understanding of the language. When I will understand Python, I maybe will thinking about such things (for example, if some parts of the language will give more disadvantages than advantages for me). Not now. <blockquote> If you don’t want to participate in the discussion that way </blockquote> I just do not want do participate in discussion when somebody is trying to discuss imagined sense of my words. Is it a forum message, is it a PM, is it a voice into my ear? It does not matter. Meaning matters. But now you have an awesome answer. Why you didn't publish this in the first place? <blockquote> In your “by the way” post, you asked a question that’s really deep and demands a thorough knowledge of Python internals. IMO, it’s very hard to imagine a beginner asking such a question. If I read you right, you say that it was just an accident, you were playing and stumbled on this. </blockquote> I just usually have more wide thinking and have more questions to an entity I learn, than others. I'm still being a newby to Python. I was playing and stumbled on this. 'Globals' is called a dictionary but I can't modify it as I can modify other dictionaries. Now I know that it's a special version of a dictionary which lies not among things I can fully administer from '.py' files.

veky on March 4, 2015, 7:21 p.m.
Ok, first part you've just made undiscussable by saying you'll interpret whatever I say there as a perversion of your words. I don't mind, there wasn't anything very valuable there anyway. :-P <hr /> So let's go to the second part. It's more interesting because it concerns Python. ;-) <blockquote> Why you didn’t publish this in the first place? </blockquote> Because I have no idea what you know and what you don't? You forget I didn't know you existed until a week ago. And I must say that from my perspective, you still got it wrong. "'globals' is called a dictionary but I can't modify it as I can modify other dictionaries." is wrong. It is a plain dictionary, you can do with it whatever you can with other plain dictionaries. It just has a specific method that returns it, and it is stuffed into a readonly attribute of a module. But that doesn't make it special. If I have a list `_l=[3,5]`, and a function <pre class='brush: python'>def myList(): return _l </pre> and a class <pre class='brush: python'>class myClass: @property def roa(self): return _l </pre> and an instance <pre class='brush: python'>myObj = myClass() </pre> then you have all these phenomena (more or less), but to conclude from that that "myList() is a special list" would be ridiculous. `l` is a perfectly normal list. You can do with it anything you can do with any other list (see `dir(myList())` for a list:-P of such things;). Rebinding a name of some object x to something else, is _not something you do to x. (Descriptors are a bit of exception, but not really.:)

zlomovsky on March 4, 2015, 9:27 p.m.
<blockquote> Because I have no idea what you know and what you don’t? </blockquote> To clarify things, I've started to learn Python about 1.5 months ago. I've done Codecademy course and I have (as it's calculated by WakaTime plug-in) 15 hours of coding experience. <hr /> So, based on your description, I was trying to get direct access to `l`, but since I do not know its name I'm calling it `myList` (I see that 'l' is a list in your example but let's imagine that it's a dict). In globals, the `globals()` function returns a link to the `l`-analog. I can work with `l` like with other dictionaries, the only difference is I can't redefine the `l` like I can redefine my own created variable `d`. I understand this. When I was asking the question, it was a question like "what the name of 'l' is? I need it to work with it directly. For example, change the object is linked to." Now I know that `_l` is "a readonly attribute of a module" when we are talking about globals. It can be used for any operations available for `dict` object, but it can't be redefined, I can't make it to be a string, I can't make it to be a list, I even can't make it to be a dict (new instance of dict I mean).

zlomovsky on March 4, 2015, 1:41 p.m.
Also. please, do not think that I was offended by your words, or something like this. I just see no reasons do talk with somebody who acts like this, because such talks are not productive and they are intended for spending a lot of time to the void without any result. It's just 'blah-blah-blah' which is not among things I like. I can do it once, maybe I can do it again, but it's too boring to do this on a daily basis.

veky on March 4, 2015, 7:24 p.m.
:-) It's nice to know that. I think these discussions are productive (although they would be more productive if they concerned Python more and psychological games less, of course), and you're free to stop anytime you want.

hanpari on March 12, 2015, 9:25 p.m.
You are too naive. In fact, Veky is not person but it is just checkio algorithm which prevent you to be much proud of yourself :) You have to be aware there is an order: 1/ God 2/ Guido 3/ Veky 4/ Poor you :) PS: Sooooo sorry, Veky, to mock you but being western citizen you probably never get to know real power of slivovice :) There is reference in Dracula of B. Stoker if you will read carefully. I believe it is chapter one.

veky on March 13, 2015, 6:01 a.m.
Don't underestimate my milieu. Croatia is in that sweet spot where we (as a culture) can claim to be western (and not laugh too much in the process:) and still enjoy the šljivovica sometimes. ;-) [Though, I personally, as a checkio algorithm, am not really designed to enjoy it very much, I don't need it. With so much code repetition in newbies' solutions I encounter here, I see double (and triple and quadruple...) even without it.:]]

suic on March 19, 2015, 10:31 p.m.
Whatabout: Veky is an algorithm (or machine) turning šlivovica into Python code? :D (When I read these comments about "Veky is an algorithm" and "In šljivovica veritas." it reminded me of Renyi's famous quote about coffee and mathematicians.)

veky on March 19, 2015, 11:22 p.m.
Considering how much šljivovica I've drunk, and how much Python code I've written, it seems I'm the most efficient machine ever. :-D [With an efficiency of about a mebibyte per deciliter.;]

hanpari on March 12, 2015, 9:14 p.m.
I have to admit in public there are only two names for me related to Python: Guido and Veky. Not necessary in this order. Stay cool, both of you :)

veky on March 13, 2015, 6:02 a.m.
In šljivovica veritas. ;-]]

hanpari on March 19, 2015, 8:58 p.m.
Veky, what about F#? You had your month as promised :)

veky on March 19, 2015, 9:41 p.m.
I did, and the only conclusion is that I probably need another month. :-) Yes, I've read much of that site, but still haven't found anything Python cannot do just as easily (or even easier). [Except static typing, but that's religion anyway.:] |> syntax still doesn't sit very well with me, implicit (un)currying I still find braindead, and ravings about elementary things like function wrappers, namedtuple and implicit interfaces (aka duck typing) are just childish. I guess it is really eye-opening to poor souls stuck in C# hell, but we in Python Kingdom have had almost all these goodies for a long time, we don't even think about them as luxury. My enthusiasm at the start was based on a wrong impression that it is a purely functional language, where you could eat your flexible typing cake and still have it, and where println is implementable without knowing category theory. :-)

hanpari on March 19, 2015, 10:01 p.m.
This is not about comparing Python and F#. Firstly, I think that site is worthy of interest no matter what programming language you were using. Secondly, I am not trying to make you switch to F# from Python. Guido would be sad. Thirdly, I am glad you have read much of that site because it seems to me appropriate... :) I mean I am glorifying both of you (you and mr. Wlaschin). Commenting from you on that site might be like a Battle of Titans :) By the way, if you want to test something, I have just found this: <a href="https://dotnetfiddle.net/">https://dotnetfiddle.net/</a> Good night :)

veky on March 19, 2015, 11:28 p.m.
As I said, site is fine for its intended purpose. :-) And a battle of Titans might be a good analogy, but not very useful there. BTW I don't think Guido would be very sad. He <a href="http://www.checkio.org/mission/hidden-word/publications/guido/python-3/generator/share/807b92bbcb07fbccff7fd25154fb6e61/#comment-9529">doesn't like</a> my style of writing Python. (Understandably - I wouldn't like it if I were him, either. We have very different goals.:)

hanpari on March 20, 2015, 5:42 a.m.
Thats funny link. Exactly this result I would expect :) By my logic was as follows: Veky is important for Checkio and Checkio is important for Python. Thus, Guido should be sad if he lost you because of F#. Regarding F# (putting Python aside) I want to react about your comment about static typing. In general, I would agree with you that static typing is annoying. But following Wlaschins way of doing things it seems to me that in F# static typing makes sense while in C# or Java is really obnoxious. And I really like the idea of measuring unit: <a href="http://fsharpforfunandprofit.com/posts/units-of-measure/">http://fsharpforfunandprofit.com/posts/units-of-measure/</a>

veky on March 20, 2015, 6:16 a.m.
Veky is surely important for CiO, but CiO is not very important for Python (at least according to BDFL). Guido wrote only one solution and about 10 comments here. :-] Of course static typing can be bad (in F#) and can be worse (in C#). :-) It can even be good sometimes, but that's not the property of language, that's the property of concrete task you're solving. F# comes very close to the ideal of having static types be just a little bit bad in the tasks where you don't need them, but there are languages that come closer - Haskell, for example. This "units" thingy is a nice example of a half-baked "solution" that could be a very nice library (v0.1:), but not in the core language. First it assumes all units are linear, and he gives an example of converting temperature units. That 1.8<degF/degC> number makes no sense at all, and he knows it. And why the hell `square 10<ft>` doesn't work? Square feet are not rocket science. It seems two different teams were writing type inference and unit support. Bleh. Honestly, <pre class='brush: python'>n + (LanguagePrimitives.FloatWithMeasure 1.0) </pre> is an abomination. But it's a natural consequence of no coherent design in the language.

hanpari on March 20, 2015, 9:59 a.m.
No agreement between us :) By the way, I am almost afraid to ask but... but... did you find something what you like about F#? Be positive, please :)

veky on March 20, 2015, 1:46 p.m.
Of course. As I said, it's nice to know someone cares about lost souls from C# wasteland. There are even things it does better than Python: for example, ranges that actually make sense. ;-]

hanpari on March 20, 2015, 6:26 p.m.
OK, I give you next month :)

GitHub

0 %

You ClassRoom Demo is Ready With the new tool CheckiO ClassRoom, you will be able to analyze your students' progress, have your own leaderboard, change missions order on the map, and so on.