Mathematica 9 is now available
Services & Resources / Wolfram Forums / MathGroup Archive
-----

MathGroup Archive 2008

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Pure functions?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg93234] Re: Pure functions?
  • From: Albert Retey <awnl at gmx-topmail.de>
  • Date: Sat, 1 Nov 2008 05:04:34 -0500 (EST)
  • References: <geee7b$a9g$1@smc.vnet.net>

Hi,

> So, what _is_ the real story on "pure functions"?

I think from your lengthy considerations there is a clear point:
There is a mismatch between the meaning of a pure function as used in
Mathematica and the definition of a pure function as given by wikipedia.

The Mathematica documentations uses 'pure function' as a description for
expressions with head Function, which in the wikipedia-sense are
'anonymous functions'. Using something like: f=#^2& as a counter example
is misleading, since here we only have a variable f that holds as its
(Own-)value a anonymous function, that doesn't make the
Function-expression have a name. A Mathematica 'pure function' can be
pure or impure in the wikipedia sense depending on what its body contains:

#^2&        is pure in the wikipedia sense
#*Date[]&   is impure (depends on external state)
Export[#]&  is impure (has side effects)

I don't think that there is any conceptual difficulty here at all, the
only question is, whether picking 'pure function' in the mathematica
documentation as a term for something that other people might call
'anonymous function' turned out to be a good choice or not. My
experience with computer science talk is that many of these terms are by
far not as 'common' as one might think or wish. When you then consider
that Mathematica has quiet a history, it would certainly cause much
confusion among long term users of mathematica if that would now be
changed in the documentation, even if nowadays 'anonymous function'
would be an excepted standard (which I am not convinced of just because
of wikipedia). So I'm afraid you will just need to learn to live with
that mismatch...

By the way, I personally would prefer 'anonymous function' because
unlike 'pure function' the term addresses quite clearly which aspect of
a function is considered, but I have never had problems with
understanding the mathematica documentation because of that.

hth,

albert


  • Prev by Date: Re: I can't evaluate cells in Mathematica
  • Next by Date: RE: Re: Re: 3D Graphics in a Palette or Dialog
  • Previous by thread: Re: I can't evaluate cells in Mathematica
  • Next by thread: Re: Pure functions?