[Idle-dev] unicode variable names

Neil Hodgson neilh@scintilla.org
Thu, 6 Apr 2000 12:52:24 +1000


> > 2) It seems like a very easily implemented, natural and
> >    desireable adjustment, is it?
>
> There are major problems with this if you consider code portability.
> WHen I was 18 and first started to program, I decided to write all my
> code (including variable names, comments and string literals) in
> English, even though my mother tongue is Dutch.  I have never
> regretted this.  Much more of my code than I ever dreamed has found
> its way all over the world.  If I had used Duch in my programs, they
> would never have crossed the borders.

   This is easier for Dutch people than for people whose mother tongue is
further removed from English. I've worked on Japanese code and found that
choice of program identifiers was often poor, which I attribute to the
burden placed on Japanese programmers in having to use roman characters for
identifiers. They can try to think up a good English word for an identifier
(which takes too long and may require a dictionary to make sure its OK) or
try to write a Japanese word in roman characters. I've not seen anyone use
the latter technique, possibly because its never done in other fields of
life where roman characters are only used for imported words.

   If this is a problem for professional programmers, who have generally
learnt good enough technical English skills to read English journals, how
much of a burden is it for the Es in CP4E?

> > 4) Is it desireable to allow unicode variable names in
> >    sourcecode, like in Java. Is it desireable for JPython?
>
> Since most computer systems used by programmers have hardly any
> support for entering and displaying Unicode, this would again hamper
> source code portability, so I'm not a big fan of this.

   NT/W2K has quite decent support for Unicode entry and display which can
be reasonably easily used in programmer tools. On Windows at least, IDLE
supports Unicode display and Pythonwin will real soon now.

   Neil