Order of parameters in localization code

Question

In localization code (like T('English phrase with %s and %s parameters', param1, param2) the order of parameters in different languages is different. How to handle this?

Our localization engine seems not to support named or numbered parameters, so the parameters need to go in the same order as they are in code.

Answer 1

If your localization engine doesn't support named parameters, then there are a few options.

1. Get a new engine that does support named parameters
2. Don't let your engine fill in the parameters and feed the translated string to a second "printf" engine, which does support named parameters, to fill in the parameters.
3. Accept that you will get badly translated output and ensure that your translators know that the parameters may not be swapped during translation.¹

^{1. I have actually worked on a product where this choice was made, due to technical limitations in the printf/display engine.}

Answer 2

Pick a different localization engine, or roll your own if there are none for the language you use.

Parameters inversion is quite frequent in internationalization. That's why many format-type functions enable to specify the index of the parameter in the pattern.

For example, in Python, the two following lines print the same thing:

print('Hello, %s!' % ('John'))
print('Hello, {0}!'.format('John')) # Can specify the order

In C#, the most used pattern indicates in which order the parameters are taken:

string.Format("Hello, {0}!", "John");