best practice(s) for singular word or plural words?

Question

This is about best practices in general, not specific for a single language, database or whatever

We all have to deal with generated output where you can be reporting "one products" or "two product". Doesn't read very well... Some just solve this by using "one product(s)" or "number of products: (1)" and others might have other solutions.

Things could be even more complex in different spoken languages! In French, when you have zero products, you would use the singular form, not the plural form! (Zero product) Other languages (Chinese, Japanese) might even lack these grammatical differences or have more than two different words to indicate something about the number of products. (A plural and a greater plural, for example.)

But to keep this simple, let's focus on the languages that have both singular and plural words.

When setting up a new project, which also has to generate reports, how do you deal with singular and plural words? Do you add two name fields in your database for singular and plural form? Do you add additional rules in the code to transform words from singular to plural? Do you use other tricks?

When working on a project that needs to track singular and plural forms, how do you deal with this?

Answer 1

Number of products:  1 
Number of products:  4
Number of products:  FILE_NOT_FOUND

Trying to use natural language for reporting quantitative data is just too much of a hassle.

Answer 2

With english applications it's usually easiest and most efficient to store the singular and create the plural by using a group of if statements.

if( count > 1 ){ suffix = 's'; }

Answer 3

Normally I would send my text through some kind of formatter, which reformats the values you want to display to a human readable text. This also could modify your "product" text. Java has the MessageFormat class for this, which supports such modifications. See the examples at [1].

[1] http://java.sun.com/j2se/1.5.0/docs/api/java/text/MessageFormat.html

Answer 4

I'd recommend taking a look at gettext in general and ngettext in particular. Maybe even if you're not going to translate your application. Just head to this part of the documentation. It has implementation for more or less all languages and even if your language of choice lacks this support, nothing stops you from borrowing the ideas.

Answer 5

Read and implement this; report back when you're done (in a few years). Personally, I'm satisfied with the (s) approach ;) (though it goes without saying this doesn't work for all languages).

Answer 6

In Perl, this is comprehensively solved by Lingua::EN::Inflect. It uses a large dictionary and carefully handles all the exceptions to rules. It also does things like 'a' or 'an', and handles comparisons as well!

See the paper for the gory details.