Le 7 mars 2012 17:58, Dominikus Dittes Scherkl <[email protected]> a écrit : > Am 07.03.2012 01:52, schrieb Ken Whistler: >> On 3/6/2012 4:25 PM, Leo Broukhis wrote: >>> What about Grapheme_Extend class characters placed out of context? It >>> would be nice to see a dotted box in cases like AׁB >>> (U+0041 U+05C1 HEBREW POINT SHIN DOT U+0042) > In fact, I do see a dottet circle between A and B with the point on it.
Me too. In fact this is a good fallback solution, which remains perfect as long as noone really need to give some distinct meaning to the cluster A+SHIN DOT, and render it without that circle with a dedicated font. In my opinion it would be very confusive with <A, COMBINING DOT ABOVE>, but this could be <A, COMBINING DOT INSIDE>. And until now, we've not seen an actual use of the dotted circle with diacritics to mean something else (with the litteral symbol dotted circle by itself) than this fallback. If this ever occurs, renderers could still use some other fallback mechanism (including by changing the color(s) of the dotted circle itself, or by making it visible only when it is hovered by a mouse tool, on an interactive media and with a mechanism supported by the renderer itself to control various rendering aspects or options). I don't think that Unicode should standardize those fallback mechanisms, and the dotted circles seen in charts or in the book or just a local-only convention which is explained locally (in an informative way, not a normative way), but that does not need to be enforced to every other media or in every fonts and renderers.
