Re: Unicode implementation concept
Markus Scherer <[email protected]> Tue, 27 Jan 2004 11:52:56 -0800
| Newsgroups | gmane.comp.lib.icu.general |
|---|---|
| Organization | IBM |
| Message-ID | <[email protected]> |
Mark Evans wrote: > If character indexing is considered unimportant, > the idea has no merit. > > That opinion is misguided insofar as semantic > meaning arises only at the character abstraction > level. It seems a major bone of contention > in your circle, for mysterious reasons that > I cannot fathom. Character based indexing is > very, very much "normal software development." > If we disagree on its importance, so be it. We all agree that character (code point) _boundaries_ are important for some processing, as are "user character" boundaries (default grapheme clusters, in Unicode parlance) and word and sentence boundaries, etc. This is different from using character/code point _indexes_ for low-level string handling. Indexing and boundaries usually use code *unit* indexes which simply count the base units in strings (where strings are normally vectors of code units). Mark's emails illustrate this point nicely, in my opinion. > Be constructive now: suggest some test data > that will provide a performance/complexity testbench. This is beside the point. You are making an assumption that others don't. We have not spent much time trying to figure out good strategies (and test data) for code point indexing because hardly anyone is ever asking about it. ICU works under the same assumption as most other string handling software, and with that it works well and needs no special string data structure. Of course, as Mark mentions, there are possible optimizations for other reasons, like insertion into large strings. Such optimizations have nothing to do with code point indexing. > Since you are so antagonistic towards this > simple and elegant concept, why not dream up > the worst-case stress-test scenarios you > can imagine? I would be eager to run them. I did say early that I find your idea interesting and novel. It solves a problem that most people don't have. It may or may not be a good solution for that problem. markus -- Opinions expressed here may not reflect my company's positions unless otherwise noted.