At one point, Latin was the lingua franca[1], but those days are long past. French and Russian were used in the mathematics, but even there, mathematical papers are being published largely in English nowadays as much as possible, even if the author is a native speaker of another language.
The reason this is practical is because the sciences all speak their own language anyway. Every chemist, regardless of their native language, knows that this expression[2] denotes the combustion of propane. Likewise, mathematics uses its own symbols and relationships.
I remember the story of someone - perhaps Erdos? - co-authoring a paper with another mathematician, despite neither of them speaking a single natural language in common.
A well-written computer program may have English identifiers for variable names, but this is really cosmetic more than anything else. (If you don't believe me, go ask your lexer!) Don't confuse the fact that we usually attach English identifiers to variable/function names with the myth that our programming languages are somehow "English" in nature.
[1] An ironic term in this case, for sure!
[2] http://upload.wikimedia.org/math/a/7/6/a76ae3507ba0d03474726...
