I was under the impression that old English manuscripts did use different glyphs for the two sounds of th.
Jony > -----Original Message----- > From: Peter Kirk [mailto:[EMAIL PROTECTED] > Sent: Thursday, July 31, 2003 8:30 PM > To: Jony Rosenne > Cc: [EMAIL PROTECTED] > Subject: Re: Hebrew Vav Holam > > > On 31/07/2003 11:31, Jony Rosenne wrote: > > >This argumentation applies equally well to th (which should > be at least > >two Unicodes in English), gh (how many?), etc. > > > >Jony > > > > > > > >>-----Original Message----- > >>From: Ted Hopp [mailto:[EMAIL PROTECTED] > >>Sent: Thursday, July 31, 2003 4:58 PM > >>To: Peter Kirk > >>Cc: Jony Rosenne; [EMAIL PROTECTED] > >>Subject: Re: Hebrew Vav Holam > >> > >> > >> > >> > >... > > > > > >>I think of holam male as an indivisible glyph that happens to > >>look like a vav with a dot centered above it (or above its > >>stem, if you will, but that's just how it might vary from > >>font to font). It's much the same as a lower-case 'i' not > >>being a dotless i glyph with a combining dot. (Sometimes an > >>'i' is just an 'i'.) I wouldn't call the dot anything but a > >>dot, certainly not a holam male. > >> > >>Let's encode Hebrew, not dots. It may mean changes to what > >>SIL, UniScribe, and others are doing, but there's no free > lunch here. > >> > >> > >> > >> > As a native speaker of English, I certainly think of th and gh as > sequences of two glyphs, not as indivisible combinations, so > that is the > difference here. > > But a better example might be French e, e acute and e grave. > These are > three separate letters which need three different ways to > encode them. > Whether the accented versions are encoded as one character or two is > unimportant as long as they are distinct. Similarly we have three > letters, vav on its own, vav with right holam and vav with > left holam, > and so we need three ways of encoding them. > > As for the character name, I am forced to consider these entirely > meaningless except for being unique and stable, as UTC has refused to > correct demonstrable mistakes in these names, including at least one > Hebrew accent. So I would actually prefer to use a meaningless random > string of characters because at least that is more or less guaranteed > not to be misleading. > > -- > Peter Kirk > [EMAIL PROTECTED] > http://web.onetel.net.uk/~peterkirk/ > > > >
