[Tutor] class method problem

[David Hutto]

On Sat, Sep 25, 2010 at 9:16 PM, Steven D'Aprano <steve at pearwood.info> wrote:
> On Sun, 26 Sep 2010 08:13:23 am David Hutto wrote:
>> Since I had nothing else to do, but practice, this looks much better:
>>
>> def find(word, search):
>>       if search in word:
>>               print True
>>       else:
>>               print False
>
>
> For some definition of "better".
>
> If I called a function:
>
> find("anti-disestablishmentarianism", "lish")
>
> and got back an answer:
>
> True
>
> I'd feel ripped off and cheated. That would be like going to Google,
> typing in something into the search box, and Google comes back with:

OP wanted to find if a in b. Does my function do that?...Yep

>
>    Yes, we found your terms on the Internet, but we won't tell you
>    where. If you would like to find something else, we won't tell
>    you where that is either.

It's not supposed to be a four line version of Google.

>
> Aside from the name of the function, which is deceptive because it
> doesn't describe what the function does, the names of the arguments are
> also poor. The first argument is not necessarily a word.

Badly named yes, but I know it just finds a string in a string

 Nor is there
> any need for it to be -- it can be any text. The second argument is
> poorly described as "search" -- search is a verb.
>
> A better function signature might be:
>
> def search(text, target):
>    # Search for target in text and print whether it is found or not.
>
> Even this is slightly misleading, because "text" doesn't need to be an
> actual string. It could be any sequence, such as a list. But this gives
> the *intention* of the function, which is to do text searches. So this
> is (in my opinion) an acceptable compromise between intention and
> generality.

Semantics, are only important to those who didn't write it.

>
> Now on to the code itself. The body of the function is needlessly
> verbose. You say:
>
> if search in word:
>    print True
> else:
>    print False
>
> This is so simple we can trace the entire function by hand. Say we call
> search("I like spam and eggs", "spam"):
>
> (1) target in text? => True
> (2) take the if branch
> (3) print True
>
> Now say we call find("I like spam and eggs", "cheese"):
>
> (1) target in text? => False
> (2) take the else branch
> (3) print False
>
> Can you see the common factor? The object which is printed is always
> precisely the same object generated by the `in` test. So we can
> simplify the body of the function:
>
> def search(text, target):
>    print target in text
>
>
> But this is so simple, there's no point to wrapping it in a function!
>
> Small functions that do one thing are good, up to a point, but when the
> function is so small that it is just as easy to include the body in the
> caller code, the function is pointless. It's not like "search(b, a)" is
> easier to write or remember than "print a in b" -- if anything the
> opposite is the case, because I would never remember which order to
> pass the arguments.

The reason I put it in a function, was the same reason the OP put it in a
class, to eventually expand on the initial program being created.(also
I had it in a snippets file, accompanied by an instance for it, a way
to conveniently call
it later if needed)

I do know I could have shortened it further, I'm not all omniscient
like yourself.

Breath deeply, and remember it's someone elses code who hasn't had
the same experience writing code that you have.

But thanks for pointing it out, I'll know to eliminate the excess in the
future, knowing that the all seeing steven is watching my every
function.

>
>
>
> --
> Steven D'Aprano
> _______________________________________________
> Tutor maillist  -  Tutor at python.org
> To unsubscribe or change subscription options:
> http://mail.python.org/mailman/listinfo/tutor
>