begin quoting Andrew Lentvorski as of Thu, Oct 27, 2005 at 05:59:52PM -0700: [snip] > http://www.inference.phy.cam.ac.uk/dasher/ > > It's a system for entering words into interface-deprived devices.
Ah, yes, I remember hearing/reading about that now. It's heavily mouse-driven, and I'm heavily mouse-impaired. (I might have checked it out and found that a quarter-hour was still frustrating... or was that the semantic web?) You've recommened this before, right? > It turns out that Chinese is giving them fits: > > How would Chinese Dasher work? > > "We would not go directly for the ideograms, since there are too many of > them. We have to build up sentences using a sequence of symbols each of > which has small information content. " > > A better coding to allow them to put together glyphs would be a big help. I'd think you'd almost need to take a totally different approach. Start with [strokes] [nouns] [verbs] [modifiers] [common] [uncommon] or something like that. Go by what you are trying to *say* -- which might be an interesting way to train people on such languages: under each glyph could be the (local language) translation. [noun] -> [animal] [vegetable] [mineral] . . . That would take a heck of a data file, though. Hmmm... Maybe just do it by brush strokes. [horizontal stroke] [vertical stroke] [diagonal strokes] [curves] [box] [dots] ...heh. That's basically approach #2. I would guess that a phonetic approach would be a big problem in chinese, as I've always been told that there are multiple incomprehensible spoken dialects for one written language. Dunno if that's the best way to order the _glyphs_. I was thinking more along the lines of transliterating and munging the most common glyphs... Organizing ideograms sounds like an interesting approach. To bad I know, at best, only a few dozen ideograms, and haven't been making much progress lately. [snip] > >Okay. I'll have to ponder this for a bit. > > I believe that the original constraint was the fact that Windows used > wide strings (ie "String" == "array of byte pairs"). Thus, there was no > ability to move beyond 65536 glyphs because Windows couldn't cope with > the fact that a single glyph might be 2 *or* 4 bytes. ...while the Plan9 guys sat down and said "let's think about this over lunch". > However, the fact that Windows demanded wide strings helped a lot. An > entire generation of programmers has grown up without necessarily > thinking that "String" == "char[]" == "array of single bytes". I'm not entirely sure that's progress. Which, come to think of it, isn't suprising. Given two design or philosophical choices, I almost always end up on the side opposing M$ before learning which way M$ chose... C++, hungarian notation, UI design, ... and now wide chars. This is why I emphatically oppose "one solution for everyone" movements. I know the "one solution" that would be 'standardized' would be the one that I'd be unhappy with. -Stewart -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
