- From: MURATA Makoto <eb2m-mrt@asahi-net.or.jp>
- Date: Wed, 14 Dec 2011 22:48:28 +0900
- To: www-style@w3.org
2011/12/14 Florian Rivoal <florianr@opera.com>: > On Wed, 14 Dec 2011 13:54:37 +0100, MURATA Makoto <eb2m-mrt@asahi-net.or.jp> > wrote: > >> 2011/12/13 Håkon Wium Lie <howcome@opera.com>: >>> >>> Koji Ishii wrote: >>> >>> > I see defining some i18n values first and then try to look for >>> > general @rule in future is a good incremental approach. But if the >>> > WG sees it differently and resolves not to include values that are >>> > only used in some scripts, it's very unfortunate for me, but >>> > deferring full-size-kana looks better way than defining >>> > @text-transform without taking enough time to think about it. >>> >>> I'd rather not define script-specific values without a generic >>> mechanism -- it will lead to difficult discussions about which scripts >>> should be prioritized and why. But I think it's possible to define >>> @text-transform within a reasonable time -- this could be a win-win >>> scenario. >> >> >> Although your desire for a universal solution is very sensible, I >> do not think that such a solution can be invented without creating >> and using ad-hoc solutions first. > > > I don't think we can create a (good) generic solution without > considering the individual use cases it is trying to solve, and > see if it solves them well. But I disagree that we have to solve > (ie. spec, write tests for, implement, ship..) these use cases > individually first to be able to create the generic solution. > > >> I am also worried about universal-but-hard-to-use solutions. >> For example, some subset of sed or perl would be a very >> universal solution. But isn't such a subset overkill? Hard >> to use and hard to implement? > > > I don't believe the solution I am proposing will be that hard. > But that is also why I have started making a draft for it. After > working on it for a little while, if our conclusion is that it > is too complicated and is a bad solution, then maybe it should > be abandoned. But I don't think it will be that hard to use. If grapheme clusters, word boundaries, and Unicode normalizations are incorporated, the result will be very complicated. Note that Unicode regular expressions Level 1 (Unicode Technical Standard #18) significantly simplifies grapheme clusters and word boundaries. The smallest generic solution is one-to-one mapping of UCS code values. I would be a small subset of your "convert". I think that it would be very appropriate as Level 1 of text transformation. Regards, Makoto
Received on Wednesday, 14 December 2011 13:49:03 UTC