[Prev][Next][Index][Thread]
Re: Long names are doom ?
Barry Margolin <barmar@genuity.net> wrote in message news:<MVwP6.27$_u.1005@burlma1-snr2>...
> In article <9elpia$1f19$1@ID-22205.news.dfncis.de>,
> Jochen Schmidt <jsc@dataheaven.de> wrote:
> >> - "long variables names are *hard* to read. And, you have to
> >> read though all the characters of every instance of them...".
> >
> >see above Example "genFunArgPrecOrder"
> >
> >> - "it degrades the legibility of a program to use identifiers that
> >> can't be easily remembered...."
>
> It should probably be noted that the functions with long names are usually
> *not* ones that are used frequently, so they don't impact readability or
> writability very much.
>
> And it's not uncommon to provide short abbreviations. In particular, note
> that Scheme has "call-with-current-continuation" (30 characters), but it's
> almost always invoked using its synonym "call/cc".
>
> And when you're having an online discussion about a long-named function,
> it's common to make use of an unofficial abbreviation. For instance, I've
> seen many emails about the CLOS MOP that refer to things like
> ENSURE-GENERIC-FUNCTION as E-G-F. This is not unlike legal documents that
> introduce a shorthand (e.g. "the Company") for a long term that will be
> used frequently (e.g. "The Widget Factory, Incorporated").
It should also be noted that in Dylan, there is no global namespace, and when
importing bindings from other modules, you can rename them however you like. So
in order to have fully descriptive names module authors can export nice, long
names, but when importing them you can abbreviate them however you like.
-- Brian Campbell
To really contact me, mail me at lambda (at) world (dot) std (dot) com
References: