Re: Vai word boundaries and line breaking
Asmus Freytag <[email protected]>
| Newsgroups | gmane.text.unicode.devel |
|---|---|
| Message-ID | <[email protected]> |
On 11/30/2007 4:33 PM, Michael Everson wrote: > At 16:15 -0800 2007-11-30, Peter Constable wrote: >> > From: Michael Everson [mailto:[email protected]] >> >>> Modern Vai always uses spaces between words. As in the Bible I have >>> here. Manuscript Vai is less scrupulous about this, but the Right >>> Thing To Do is treat the syllables like Cherokee syllables... that >>> is, like Latin letters. >> >> OK; that's helpful. > > I'm glad. > >> > >L2/05-159R has a section on line breaking that indicates that line >>> >breaks are not permitted before SYLLABLE LENGTHENER or before NG. >>> >Looking at the beta UCD files, though, it appears that NG (A60B) has >>> >the same line breaking property as regular syllables (AL); it's not >>> >clear to me if that's intended or not. >>> >>> That is Ken's bailiwick I guess. >> >> Well, I think Ken is dependent, like me, on the script experts out >> there spec'ing the desired behaviour. > > OK, I suck at properties, so use small words and ask me again. > "Intended"? Line breaks should not be permitted before SYLLABLE > LENGTHENER or NG. So... the properites shouldn't allow breaking before > them. If something extends a word that itself consists of characters with type AL, then making the extender also AL has the effect of keeping the word together. AL is a good default for anything that's not punctuation and that's to be kept with words. Special properties are needed usually only if something can occur separated from a word (e.g. following a close paren, for example, or inside a number) and/or where it doesn't follow a 'word' made of AL, but things like ID (ideograph) that themselves don't bind together. So, any marks that occur _only_ as part of words can be AL. > >> > >The LENGTHENER (A60C) has linebreak = EX, as does FULL STOP (A60E); >>> >this strongly blocks break before. COMMA (060D) and QUESTION MARK >>> >(A60F) have linebreak = BA, which less-strongly blocks break before. >>> >(EX allows preceding break only after hard line break or ZWSP; BA >>> >also allows a preceding break after spaces (SP) or FFFC). >>> >>> I can't say I know how to evaluate this. >> >> Consider this an invitation for any of the people that were involved >> in the Vai proposal or with any with a connection to the primary user >> community, to review the Unicode 5.1 beta UCD files and provide >> feedback as needed. > > I'm afraid you're stuck with me. Ask questions I can answer, and I > will answer them as best I can.