something like Lingua::EN::Sentence might be a good start (but it only has a tiny list of acronymns/abbreviations)

Joost.

Thanks for pointing this out to me! I'll start looking at it.

Someone once told me that there are certain rules of thumb for determining if something is pronounceable.

For example: If no vowels it's an acronym/abbreviation. Certain letter combinations are never seen for a real word... I wish there was a way to find this rulebase.

If no vowels it's an acronym/abbreviation.

cwm is the exception to prove your rule...

vera is a text database of acronyms available at gnu.org. It would be pretty easy to take the initial character of your candidate acronym, lowercase it and search the vera.? file for a match.

my $meaning;
if ( /\b([A-Z]?)\b/ ) {
    open my $fh, '<', '/path/to/vera/vera.' . lc( substr $1, 0, 1)
        or die $!;
    while (<$fh>) {
        $meaning = <$fh> and last if /$1/;
    }
}
[download]

The problem with any/all solutions to this problem is that of acceptable usage. For example, there are technical documents who would refer to SCUBA but travel brochures that would refer to scuba. Similarly scientists work with LASER but companies sell laser devices.

You cant create rules about pronouncability because both 'laser' and 'scuba' are pronouncable. 'Sky' contains no vowels and so might/might not be an acronym. Qantas contains no 'u' after the Q, nor does Iraq. Qantas is the "Queensland and Northern Territory Aerial Service" however, noone uses that outside of a trivia game these days. Its considered a word. Iraq, of course, is from another language and doesn't follow English rules.

Databases like vera can help you with a list of known acronyms but in Real Estate advertising 'LUG' is a lock-up-garage whereas in a mechanical journal a 'lug' is a type of nut.

Given all this, when parsing user text I normally require that there be one or more lower-case letters in users input. That way I know they didn't just type it with the Caps-Lock button on. If it's all in capitals I'll ask them to change or confirm what they've entered.

Good luck.

Um, you have that backwards. An acronym can be pronounced and is accepted as a word: RADAR, LIDAR, SONAR, SCUBA, PATRIOT, and so on. IBM is merely an abbreviation (or initialism) because you say "eye bee em."

--
[ e d @ h a l l e y . c c ]

The dictionary definition (external link) of "abbreviate" means simply "to make shorter". From that definition, I would say that an acronym is a specific type of abbreviation.

The dictionary defintion (another external link) of "acronym" could be interpreted to mean what you say, but I don't think it has to be. The defintion given above says nothing about being able to pronounce the series of letters, just that it is a word formed by the first letters of the word series (or parts of the word series), which would include "IBM".

I typicaly think of an abbreviation as a shortening of a single word, such as "Inc."

----
send money to your kernel via the boot loader.. This and more wisdom available from Markov Hardburn.

I'm backing Halley on this one, and strongly disagreeing with hardburn and dragonchild.