Hello Andrew. I'm aware of the issues with parsing japanese text. I'm the developer of Furigana Injector, which uses the Mecab lib. It uses a much larger dictionary (IPADIC, or UniDic), saved in a binary format in a trie data structure. A trie is the bee's knees for this sort of text search, as you've worked out. What I read about other people's trie classes written in Javascript, however, is that they don't perform so great and you seem to be confirming that. The amount of memory being consumed for the dictionary is many times (10x?) the starting amount the entire browser process required.
I can't see a way around this though, if using Javascript. The Rikaichan method of using one huge string and an index that contains offsets into the string looked clunky to me when I first saw it, but maybe that is the best optimization- to avoid creating javascript objects as much as possible. I'm going off to crbug.com to see if there's anything about binary dictionaries ... Akira --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Chromium-extensions" group. To post to this group, send email to chromium-extensions@googlegroups.com To unsubscribe from this group, send email to chromium-extensions+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/chromium-extensions?hl=en -~----------~----~----~----~------~----~------~--~---