On Tue, May 27, 2008 at 10:53:23PM +0200, Ondrej Certik wrote:
>
> Hi,
>
> I think we should know allow using implicit imports in the sympy code
> as I think it only makes debugging harder because you don't know where
> the symbols are comming from.
> We discussed this on the mpmath mailinglist:
>
> http://groups.google.com/group/mpmath/msg/d97f4b9e195b77a3
>
> where Fredrik used the argument:
>
> > This is a good convention generally, but not really necessary to
> > follow when there's a small number of well isolated modules. The fact
> > that mpmath uses import * internally doesn't mean sympy has to. I
> > think any code in sympy that uses mpmath should import explicitly.
>
> which I found reasonable, so I didn't pursue this issue further. But
> now the problem with it has popped up in this issue:
>
> http://code.google.com/p/sympy/issues/detail?id=864
>
> had we used explicit imports in mpmath, the error message would be
> more clear, as python would shout at the import time that he cannot
> import it or something. Does anyone have some good arguments for ever
> using implicit imports? I think it is necessary even for a small set
> of isolated modules.
I think implicit imports like "from numpy import *" are handy for
interactive sessions.
But for source code I too prefer using explicit imports because it is
then much clearer (say e.g. after several months) what gets from where.
Explicit is better than implicit after all :)
--
Всего хорошего, Кирилл.
