> > I thought glyph variants were supposed to look at least somewhat similar. > > Any reference to this similarity in appearance as a condition ? > > (Is the S�tterlin �e� a glyph variant of standard latin �e� then ? It does > not ressemble any other "e" I know but rather an "n".) > > P.A. Okay, let's go to the standard:
3.2.D2: Consistency with the representative glyph does not require that the images be identical or even graphically similar; rather, it means that both images are generally recognized to be representations of the same character. Representing the character U+0061 LATIN SMALL LETTER A by the glyph "X" would violate its character identity. So I wasn't entirely correct. If they were Han ideographs they wouldn't be unified, becuase of R3 (the abstract shape, Unicode 3.0) and probably R2 (unrelated in historical derivation). (That is, the somewhat similar rule clearly applies for Han ideographs.) But talking about character identity, anyone using a S�tterlin font expects the e to appear as it does. The S�tterlin e is clearly an e in character identity. But as a mathematician, there is little variation I would accept for U+2203 as an acceptable glyph shape. There is quite a bit more that I would accept as conforming, but stupid. But all of them I would count as acceptable would look like a backwards E. IMO, any other glyph shape violates the character identity. Another rule which isn't written into Unicode but I like (don't know if Everson and Whistler and others will), is the font clarity rule. Given a font minus one character, I should be able to predict what that character will look like. If I have a S�tterlin font or a Fraktur font, I know what the � will look like. It may look nothing like it would in Times New Roman, but it will look right for that font. If there are two seriously different glyphic choices for a character in one font style, then you have inappropriately unified the two characters. -- ___________________________________________________________ Sign-up for Ads Free at Mail.com http://promo.mail.com/adsfreejump.htm
