I was thinking more of a bottom-up approach, where all possible readings for glyphs/ligatures are listed in a table and different words/languages can be more easily considered.
Then we could see at a glance which combinations form what looks like words, and then just finds translations for those "words" in any possible language.
Take a very simple example like "der mus del". Wiktionary gives translations in 15 languages for der, 18 for mus and 28 for del...
Edit: sorry david, I did not see your edit. It seems like we are thinking the same. I'll make a new thread for discussion.
Then we could see at a glance which combinations form what looks like words, and then just finds translations for those "words" in any possible language.
Take a very simple example like "der mus del". Wiktionary gives translations in 15 languages for der, 18 for mus and 28 for del...
Edit: sorry david, I did not see your edit. It seems like we are thinking the same. I'll make a new thread for discussion.

