Don't re-invent the wheel.You could use for example http://emboss.sourceforge.net/apps/release/6.0/emboss/apps/nucleic_translation_group.html. Bioperl ships with parsers for that if you need the protein sequences in Perl.