I know that a very big dictionary is likely to be bad for the system performance. But nevertheless, you can now import Ralf's Modern Greek dictionary (version 0.1; GPLv3) into simon. On my computer, the import took less than 3 minutes. Training with this dictionary is currently not recommended because of phoneme issues.
Some words about the creation of this dictionary: After getting a spelling dictionary, I converted with iconv the unreadable Greek characters from ISO-8859-7 to UTF-8 (all my dictionaries are UTF-8 encoded). Then I generated the phonemes with eSpeak. In the last step, I converted the eSpeak phonemes into IPA.
simon displays the Greek words correctly:
You can imagine what great advantage UTF-8 has to offer: no encoding issues any more. No crappy characters should occur.
My concept is to offer pronunciation dictionaries that are as user friendly as possible. You don’t need to understand the details. You just have to know that you have to do two steps:
1. Download Ralf's Modern Greek dictionary.
2. Import as PLS dictionary into simon.
Edit November 10, 2009: For the conversion from eSpeak phonemes to IPA phonemes I used the following XSLT style-sheet: espeak2perfectipa-greek.xsl
This style-sheet could be edited by a native speaker to improve the quality of the IPA phonemes.
Tags: iconv, ISO-8859-7, node55, UTF-8
