From: "Peter Kirk" <[EMAIL PROTECTED]> > On 15/11/2003 09:08, Philippe Verdy wrote: > > >... (also in Hebrew and Arabic > >for vowel points and marks which should better be entered with a logical > >order after the base letter, even if they produce combining sequences to the > >application through the WM_CHAR event). > I'm not sure of your point here. In both Hebrew and Arabic vowel points > come logically after their base characters and so in the same order as > in Unicode - with a few insignificant exceptions. The canonical order of > points after any one base character is not the logical typing order, but > as canonically equivalent sequences are supposed to be processed the > same there is no need for a keyboard to ensure that it generates the > canonical order.
I was refering to the fact that a keyboard that allows you to enter Hebrew diacritics in any order, even when they are not entered after a base letter, will allow users to enter text which cannot be rendered successfully. When you are creating a text that will later be used in a legacy encoding with limited support for Hebrew in fonts, a keyboard that restricts you to enter precombined forms (even if they are not canonical) will work better: the sequence of letter + diacritic is only seen in the keyboard driver or IME, not in the text editor that uses it.
