From: Werner LEMBERG (wl@gnu.org)
Date: Wed Feb 19 2003 - 14:54:58 EST
From: Barbara Beeton <bnb@ams.org>
Subject: re: [OpenType] PS glyph `phi' vs `phi1'
Date: Wed, 19 Feb 2003 11:56:03 -0500 (EST)
[Dear Barbara, I took the liberty to cite your message almost
completely while CCing the opentype and unicode lists.]
> the shapes of the two `phi's haven't changed since unicode 2.0; the
> change for unicode 3.2 is in the additional text. the naming in
> unicode of 03D5 as a "symbol" is the unicode technical committee's
> convention for indicating "an established variant that we have to
> include". while i disagree with the designation of 03D5 as a symbol
> to the exclusion of 03C6 (resulting in the note "in mathematical
> contexts ..."), the fact that both shapes already existed in unicode
> meant that they shouldn't be switched, since they had presumably
> been used in documents whose meaning could be corrupted thereby.
>
> i have to regard the unicode use as correct regarding codes and
> shapes. there *could* be an error in the annotations; i'm not
> familiar with the name "phi1". the only entity names i know are
> these:
>
> - isogrk3:
> - phis = straight phi
> - phiv = curly or open phi
> - isogrk1:
> - phgr = small phi, greek (shown as a curly phi)
> - there is no straight phi in this entity set
>
> unlike the main unicode names (which can't be changed -- a rule that
> ensures that iso 10646 will be identical to the relevant subset of
> unicode), the annotations can be changed, so i will forward your
> query to my contacts on the utc.
Thanks. As a conclusion it seems that both Adobe's mapping of U+03D5
and U+03C6 to glyph names and the Unicode annotation for U+03D5 is
incorrect (in case backwards compatibility is of importance).
The right mapping should be
phi 03D5
phi1 03C6
Werner
This archive was generated by hypermail 2.1.5 : Wed Feb 19 2003 - 15:40:04 EST