Similar to the classic word game 'Scrabble', but with unicode support for multiple languages and character sets. The game is currently distributed with 88 dictionary resources for playing Language[i]- Language[j] 'Scrabble'. For example, if configured to use the French-English dictionary, then the distribution of available tiles will be computed based on frequency of occurance of each character of Language[i] (French), and for each submission the corresponding definition will be given in Language[j] (English).

Here is an old video with no sound.
Broswe code on github.
Help wanted!

TuxWordSmith can be ported to JavaScript and run in the browser, just like TuxMathScrabble. The Python version could use lots of work, too. Contributors, collaborators and helpers welcome! Please coordinate via the forum.

Unfinished work generating MxN overlapping entries. So far I have managed to get up to 5x5 shared blocks, but computation time goes waaay up with 5, let alone 6. But, the new fitting algorithm for TuxMathScrabble HTML5 is well suited for this, or a variant of the algorithm, at least. It involves non-linear tensor-transforms in imaginary Hilbert spaces, so it's gonna take a little while ...