From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Sat Sep 29 2007 - 18:25:43 CDT
Michael Everson wrote:
> At 11:34 -0500 2007-09-29, vunzndi@vfemail.net wrote:
> >So before this thread is stopped. Who is going to write the proposal
> >to encode <><?
>
> See U+1319F in FPDAM5
If I look at the current FPDAM5 documents, I can see that it adds a new
large block of Eyptian hieroglyphs. But, independently of the encoding by
ISO of these characters, there remains the question of their properties that
will affect the way they are rendered (or interpreted).
I can't see those details in the referenced documents N3237, L2/07-097 and
N3345... But these details will be needed for inclusion in the Unicode
standard. Did you start a discussion somewhere about their intended
character properties and the content of the future presentation chapter in
the amended Unicode standard?
The main questions about the proposed hieroglyphs are:
(1) About their directionality:
- (1.1) What will be their default directionality: left to right, top to
bottom, other? Inherited from the text before them, or from a leading
direction override format control? Will new direction overrides be
introduced to support the vertical layout if the default direction is
horizontal? In horizontal boustrophedon style, will it support the inversed
ductus on even lines (i.e. with mirrored glyphs, like for Old Greek rendered
in boustrophedon style) or will the ductus be kept constant (non mirrored
glyphs)?
- (1.2) How can a text encode a change of direction explicitly (if a single
leading control format does not encode the boustrophedon or spiral style
implicitly for the rest of the text): with an explicit format control
inserted everywhere a direction break occurs or after each explicit line
break control?
- (1.2) Which characters should be mirrored, or rotated when displayed in
boustrophedon style or vertically? Can this implicit reorientation be
overridden for specific characters or spans of characters?
- (1.3) How will line breaks work? Does it preserve the directionality or
does it swap it?
- (1.6) How will the BiDi algorithm be affected?
(2) About the cartouches:
- (2.1) Characters are added to encode the beginning or end of a cartouche
(with a representative glyph showing the left or right part of the
cartouche) but no combining character is added to modify the enclosed
hieroglyphs. Does this mean that the upper or lower parts of the cartouche
is inherited from the previous hieroglyph, starting by the character
encoding the beginning of the cartouche?
- (2.2) Is the presence of cartouches affecting the position of candidate
line breaks (if the text is not rendered in a boustrophedon style, either
horizontally or vertically, or in a spiral style: in both styles, the line
breaks do not affect seriously the way in which the cartouches are rendered,
given that they can run continuously, because the whole text looks
effectively like a single zigzagging line, where the text baseline is not
necessarily oriented in the same direction as the glyph baseline when the
glyph orientation is kept)?
Also, it seems that wordbreaks are not a difficulty for hieroglyphs (but
there may be exceptions with special symbols appearing around hieroglyphs,
notably the cartouche delimiters (acting like surrounding punctuations, but
also possibly prohibiting some line breaks in the middle if the text
direction is kept constant between two lines), or specific punctuation
signs, or hieroglyphic numbers and date symbols.
For scientific documents discussing the hieroglyphs script, it may be
acceptable to use the convenient unidirectional style (as used in Latin or
Arabic), but for rendering complete hieroglyph texts, supporting vertical
layouts or boustrophedon or spiral layout will be needed (additional
constraints for better rendering will soon appear like the need for full
line justification in boustrophedon or spiral styles, but this is a much
less critical issue).
Also for the spiral layout (if someone wants it), the rendering of a contour
outline delimiting the direction for reading may be needed (to help
readers), but this may be addressed by style without affecting the encoding
itself, as this contour may or may not be present. This contour outline
should not be confused with cartouches that are part of the encoded text and
have their own glyph.
This archive was generated by hypermail 2.1.5 : Sat Sep 29 2007 - 18:30:07 CDT