Subject: Fw: UT_Pair car and cdr
From: Mike Nordell (tamlin@algonet.se)
Date: Fri Jun 01 2001 - 00:00:51 CDT
From: "Dom Lachowicz"
>
> Ok, everyone here has some valid points. It's important to notice design
> patterns and similarities used by languages, whether that language be
Lisp,
> C++, Java, etc... each has its own ways of generally solving any specific
> problem set. Many of these patterns are shared between languages.
>
> However, C++'s STL has an analog of our UT_Pair - it's std::pair. Lisp's
car
> and cdr are a great analogy to use here, but no better (and arguably much
> worse) than std::pair. Seeing that we are a C++ app, we should strive to
fit in
> with the C++ mantra and design patterns and other such conventions.
>
> Mike or Pat, please change car and cdr to match std::pair's impl (1st &&
2nd).
> If you're feeling ambitious, code in a version which is implemented on top
of
> the STL. Look at UT_Vector for an example of how to do this.
>
> What's in a name? A lot, actually.
>
> Take care,
> Dom
>
> Quoting Mike Nordell <tamlin@algonet.se>:
>
> > Patrick Lam wrote:
> > [Mike informing of the C++ way]
> > >
> > > I don't like first and second. I think we should keep 'car' and
> > > 'cdr'. Irrational? Yes. So what?
> >
> > Perhaps if we start to think about other maintenance programmers down
> > the
> > road...
> >
> > AbiWord is implemented in C++. As such, I'm not especially humble in my
> > opinion that we should use established C++ conventions. Like has nothing
> > to
> > do with it.
> > When in Rome...
> >
> > To me it seems like plain obfuscation to call the two elements of a pair
> > "car" and "cdr". These TLAs has absolutely no meaning in C++, not to
> > mention
> > the fact that they are *completely wrong* to use in a "pair". Those
> > other
> > maintenance programmers will also ask this question again and again. Is
> > that
> > a Good Thing? Even in the light that using "first" and "second" would
> > remove
> > all shadows of a dubt?
> >
> > I hope this answers your "So what?".
> >
> > /Mike
This archive was generated by hypermail 2b25 : Fri Jun 01 2001 - 00:11:47 CDT