Post

Aha! Huffman on the Speccy. 12000 words is pretty impressive. Given that you are stem searching, I'm not sure a doubling of word count would be untenable. You are still miles from being able to include the entire 200MB SOWPOD dictionary (zips to about 600k).

Indeed! I've got a 1.7MB dictionary I use for puzzle games - still relatively massive. I suppose you could compress it down to 5 bits per character as you only need to store 26 characters, no punctuation. Maybe it already does that.

Text, attributes, a couple of UDGs, a dictionary, a remarkably effective anagram solver. I'd be fascinated to know whether a 128k version with a larger dictionary would be materially better.