Mark E. Shoulson wrote:
That said, I have nothing against using NBSP and various other tricks and winding up supporting this. Even the INVISIBLE LETTER might make sense in some settings (e.g. where you have something to be drawn in later but the diacritic is printed now, for some reason). Just that I don't considere qere/ketiv per se a very convincing argument in a plain-text domain.
For many plain-text purposes (searching, sorting, comparison, etc.) I would agree with you, and would expect the qere and ketiv to be separately encoded. But the fact remains that there is *also* a need to render the merged forms, to display them and to print them. This means that there need to be font mechanisms to arrange base and combining mark glyphs, and for better or for worse such font mechanisms interact directly with character strings, not with 'markup'. The only available input for glyph processing is glyphs, and the first level of input is via the font cmap table from the character string of the plain text. The famous 'higher level protocols' that are supposed to look after rendering are built on top of the plain text: they are not separated from that text.
By all means recommend that for most purposes ketiv and qere should be separately encoded: there are lots of good reasons to do so. But don't ignore the need to correctly display the merged forms, which is a textual problem requiring a solution that is at least in part character-level.
John Hudson
--
Tiro Typeworks www.tiro.com Vancouver, BC [EMAIL PROTECTED]
Currently reading: The Peasant of the Garonne, by Jacques Maritain Art and faith, by Jacques Maritain & Jean Cocteau Difficulites, by Ronald Knox & Arnold Lunn
