Shortest? (54) creative solution for The End of Other by StefanPochmann - python coding challenges

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

Shortest? (54) solution in Creative category for The End of Other by StefanPochmann

checkio=lambda W:any(w.endswith(tuple(W),1)for w in W)

April 21, 2015

Comments:

StefanPochmann on April 21, 2015, 8:19 p.m.
Thanks to <a href="http://www.checkio.org/mission/end-of-other/publications/james.verran/python-27/endswithtuple/">james.verran</a> and <a href="http://www.checkio.org/mission/end-of-other/publications/MadCow234/python-27/the-end-of-other/">MadCow234</a> for pointing what `endswith` can do.

veky on April 21, 2015, 9:16 p.m.
LOL, wonderful. :-) Giving start index without end index to endswith is just beautifully weird. :-D

StefanPochmann on April 21, 2015, 9:47 p.m.
Also weird: With end index, endswith might answer True even though the string doesn't end with the other :-)

veky on April 22, 2015, 4:02 a.m.
Well, that's the intended usage. The point is that slicing a str-like object (for example bytearray) might be expensive since it has to make a copy (one of rare implicit copies in Python), so endswith has its own virtual slicing that doesn't copy. So, you should use <pre class='brush: python'>a[s:e].endswith(b) ~~~> a.endswith(b, s, e) </pre> And it's not so weird: it says "a ends with b, when sliced from s to e". The whole point of virtual slicing is that the original string as a whole doesn't end (or start) with the other. But I guess that when someone was designing startswith method, it seemed obvious that people will more often tweak s, not e, so it was made possible to give only s. Then the whole interface was just copied to endswith, where probably (on its own) it would make more sense to give just e, not s. But then the interface would be too inconsistent. All in all, I'd say that slicing from the start only in the endswith method call just to avoid reflexivity false positive is unintended usage, but nevertheless, it is documented and consistent to work exactly in that way. Congratulations! :-)

StefanPochmann on April 22, 2015, 5 a.m.
<blockquote> it says "a ends with b, when sliced from s to e" </blockquote> I understand that, it's just that the latter part isn't in the method name. Does it really have to make a copy? Even for the immutable strings and when the slice step is 1? Reminds of <a href="http://blog.golang.org/go-slices-usage-and-internals">Go slices</a>, which do not copy the contents. The special case of start=1 also seems quite natural to me, btw, that simply makes it a test for proper suffixes. And now that I just checked, I'm slightly disappointed that set.issubset doesn't offer a way to ask for proper subsets. But I get your point about start vs end arguments. I guess [,start [,end]] is just standard and special cases aren't special enough. Edit: Ok, "<" does proper subset (of course). I'm at peace again :-)

veky on April 22, 2015, 5:10 a.m.
<blockquote> I understand that, it’s just that the latter part isn’t in the method name. </blockquote> Not everything should be a part of name. Should int be called intwithbase? Some methods are really used much more often with default values of some parameters. <blockquote> Does it really have to make a copy? Even for the immutable strings and when the slice step is 1? Reminds of Go slices, which do not copy the contents. </blockquote> Read carefully what I've written: for bytearrays it has. So bytearray.endswith has to have virtual slices. Once you have them in the interface, it would be stupid not to allow the same mechanism for str. <blockquote> The special case of start=1 also seems quite natural to me, btw, that simply makes it a test for proper suffixes. </blockquote> When you view it that way, it makes sense. You can even go one step further: <pre class='brush: python'>a.endswith(b, True) ~~~> test for true suffix :-D </pre> <blockquote> And now that I just checked, I’m slightly disappointed that set.issubset doesn’t offer a way to ask for proper subsets. </blockquote> As I said in some threads before, operators exist for a reason. Method calls are for Java fans, who don't mind smaller expressiveness as a price for greater verbosity. :-D

StefanPochmann on April 22, 2015, 6:03 p.m.
<blockquote> Not everything should be a part of name. </blockquote> Yes, and I wasn't complaining and I don't have a problem with it. Just saying at first sight such modification of the meaning can look unexpected. <a href="http://www.checkio.org/forum/post/3520/if-word2endswithword1-1/">Proof</a> :-) <blockquote> Read carefully what I’ve written </blockquote> I did, and I was talking about your "slicing a str-like object [...] has to make a copy". In the case of of str itself and step=1, I don't see why it couldn't internally optimize by re-using instead of copying. But it's ok, I'll just accept that it does copy. <blockquote> a.endswith(b, True) </blockquote> Excellent :-)

veky on April 23, 2015, 4:47 a.m.
<blockquote> Yes, and I wasn’t complaining and I don’t have a problem with it. Just saying at first sight such modification of the meaning can look unexpected. Proof :-) </blockquote> I don't see how it's different than someone seeing `int("52", 7) ~~> 37` (if they don't know what second parameter means) and complaining it's unexpected. People should just read the docs. <blockquote> I did, and I was talking about your “slicing a str-like object […] has to make a copy”. In the case of of str itself and step=1, I don’t see why it couldn’t internally optimize by re-using instead of copying. But it’s ok, I’ll just accept that it does copy. </blockquote> Ah, I see where the misunderstanding was. I should have said: "slicing a str-like object (for example bytearray) might be expensive because of having to make a copy", meaning that it is possible that it is necessary to make a copy. \Diamond\Box{copy}, you know? Two nested modal operators are not easy to say, especially when English is not my first language. :-) In any case, I'm sure you agree that bytearray.endswith needs to have virtual slicing (its slice must copy, since people are already accustomed to copy mutable sequences with `[:]` - notwithstanding the fact that bytearray acquired .copy method a few versions ago), and since Python is duck-typed, it would be weird not to enable the same mechanism on other types having endswith, when it has a clear semantics. BTW Golang's slices are not without their problems: see e.g. <a href="http://soryy.com/blog/2014/common-mistakes-with-go-lang/,">http://soryy.com/blog/2014/common-mistakes-with-go-lang/,</a> Growing fat from leftovers. Python rather errs on the side of "you don't pay for what you don't use" in terms of memory. "Flat is better than nested" for garbage collection, too. :-)

StefanPochmann on April 25, 2015, 2:40 p.m.
<blockquote> I don’t see how it’s different than someone seeing int("52", 7) ~~> 37 </blockquote> Ok, I agree it's really close. I guess this still feels different to me because both parameters are written with digits, so I don't have the effect of a "word" drawing the attention away from the "number". Or because I distinguish between numbers and their written representations and am aware that there's always a base involved (or something else, in case of Roman notation etc). <blockquote> Ah, I see where the misunderstanding was. </blockquote> Thanks for the clarification, and sorry for being a nitpicker :-)

veky on April 25, 2015, 11:12 p.m.
<blockquote> there’s always a base involved (or something else, in case of Roman notation etc). </blockquote> Well, Frege wouldn't agree, but his view doesn't have many adherents nowadays anyway. :-] <blockquote> sorry for being a nitpicker </blockquote> If someone told me I'd live to see a day when someone apologized to me for being a nitpicker on CiO... :-DD

StefanPochmann on April 26, 2015, 12:19 a.m.
<blockquote> Frege wouldn’t agree </blockquote> I didn't know anything about that and only read <a href="http://en.wikipedia.org/wiki/The_Foundations_of_Arithmetic">this</a> now, which didn't help. But if anything, I get the feeling he'd agree. Maybe my point wasn't clear enough: I see numbers as abstract things. If you ask me to write down a number, then strictly speaking, I cannot do it. I can only write down a representation of it (and I have to choose one). Does Frege or anybody disagree? <blockquote> If someone told me I’d live to see a day when someone apologized to me for being a nitpicker on CiO… :-DD </blockquote> :-) I nitpick too much and sometimes feel bad about it :-(

veky on April 26, 2015, 12:01 p.m.
<blockquote> I can only write down a representation of it (and I have to choose one). Does Frege or anybody disagree? </blockquote> Frege agrees that you can only write a representation, but disagrees that you can choose one in a meaningful way. :-) There can be only one. :-D

StefanPochmann on April 26, 2015, 3:41 p.m.
I can choose between representations "4" or "IV" or "Four" and in each case, you know which number I mean. What is Frege's problem?

veky on April 26, 2015, 5:38 p.m.
That all of these are flawed, of course. :-) He really thought that "what numbers really are" should be reflected in the representation. I don't think I can defend his view properly, since I also find it ridiculous, but I remember it was something like: logic is the only medium in which numbers can be represented, since it is the only one before them. Language and other cultural protrusions are inadequate, since they rely on the idea of sequence (of symbols), which relies on the idea of number.

StefanPochmann on April 26, 2015, 7:36 p.m.
That dude needs to get a life.

veky on April 26, 2015, 7:58 p.m.
It's funny how computer geeks and old philosophers probably symmetrically think that the other group of people should really get a life. :-D

StefanPochmann on April 26, 2015, 10:21 p.m.
Well I know for a fact that he doesn't have one.

veky on April 26, 2015, 10:37 p.m.
That's just coincidental. Probably. :-)

StefanPochmann on April 26, 2015, 10:53 p.m.
His days were numbered.

oldjaponec on Oct. 3, 2019, 7:26 a.m.
not bad +4

oldjaponec on Oct. 3, 2019, 7:35 a.m.
I like your solutoions

leggewie on June 4, 2021, 10:45 a.m.
amazing

Different kind of sets

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.