From: Behnam (
Date: Sat Dec 08 2007 - 09:42:19 CST

  • Next message: Peter Constable: "RE: Questions regarding U+FD3E ORNATE LEFT PARENTHESIS and U+FD3F ORNATE RIGHT PARENTHESIS"

    I don't quite understand what anomality is being discussed here.
    It is true that ornate parenthesis were used in old fonts, old
    applications and old formats in text encoding. So was many
    presentation forms and ligatures!
    In modern text encoding these parenthesis are not being used. I
    suppose they are more useful in font design, substituting normal
    parenthesis in stylistic alternatives. The only anomaly I encountered
    was exactly this absolutely confusing mirroring effect of normal
    parenthesis when designing such substitution at font level.
    Otherwise, for an old text or even a modern text using ornate
    parenthesis, the idea that ornate parenthesis behave as they always
    did, without mirroring effect, should be welcomed.
    Now if you propose eliminating mirroring of normal parenthesis to
    eliminate the anomaly, I'm totally with you!


    On 8-Dec-07, at 9:25 AM, Kent Karlsson wrote:

    >> And Kent responded, for which I must supply a few corrections:
    >>> 1) These characters were relatively recently encoded in Unicode,
    >> False. They date from the original publication of 10646-1:1993
    >> (and Unicode 1.1 as well, of course).
    > I stand corrected. (I should have checked DerivedAge.txt.)
    > In another message Ken wrote (in a chapter of "Every character has
    > a histroy" ;-):
    >> The assignment of gc=Ps and gc=Pe was not a mistake. All
    >> left/right parenthesis and bracket pairs are given gc=Ps for
    >> the "LEFT" of the pair and gc=Pe for the "RIGHT" of the pair.
    > Perhaps not a mistake, but part of the oversight. Now that mirroring
    > cannot be changed, and the use for ORNATE parentheses is mostly in
    > RTL contexts, it would be more logical to swap these properties (and
    > properties that depend on the Ps/Pe values). In addition to adding a
    > note (in NamesList.txt) that these characters are intended for use
    > in RTL contexts. Anomalies resulting from using these in LTR contexts
    > would then have to be dealt with by other means (by someone editing
    > the text), if needed; rather than having to deal with those anomalies
    > in the expectedly more common case of RTL usage (which I venture to
    > guess isn't actually done).
    > Ever trying to fix things...
    > /kent k

    This archive was generated by hypermail 2.1.5 : Sat Dec 08 2007 - 09:45:49 CST