| Brooks Moses 2006-05-31, 7:09 pm |
| Richard E Maine wrote:
> Jugoslav Dujic <jdujic@yahoo.com> wrote:
>
> "Definable" is at least close. To me, it is a much more appropriate term
> than "lvalue". I challenge you to find anyone who hasn't specificly
> learned the term from the C language or some derivative thereof, but can
> tell you even vaguely what the term "lvalue" means. And even among
> language lawyers, you'll have to read a fine line to carefully explain
> that an lvalue is not a value, in spite of the similarity of spelling.
> (It is not a value as the term "value" is defined in Fortran anyway). On
> the other hand, I think that most educated English speakers, even
> nonprogrammers, could at least come up with something like "capable of
> being defined" as a definition for "definable".
I'm not at all convinced that this use of pre-existing English words is
always a benefit, though. Certainly a word when it gets adopted into
standardese becomes jargon in the sense that it has a specific
narrowly-defined meaning that's limited to the field, regardless of
whether it had a previous meaning or not. Sometimes the existence of a
previous meaning is useful, and sometimes it isn't -- consider the
confusion between assumed-size and automatic-size arrays, for instance,
or the confusion about what default initialization means. Quite
possibly that confusion would be substantially decreased if the standard
had not used pre-existing English words.
So I'd have to say that I don't consider this a point against use of
"lvalue". The similarity to "value" might be, though I personally never
found it confusing.
- Brooks
--
The "bmoses-nospam" address is valid; no unmunging needed.
|