Re: Latin Ligatures and Unicode

Date: Thu Dec 30 1999 - 12:30:44 EST

       John, are the varieties of Latin script where ligation is
       required for minimal legibility not merely typeface variants? I
       might choose a font that's not terribly legible unless I adjust
       the point size or leading, but clearly that doesn't imply that
       such info is required in plain text. Are the cases you're
       thinking about categorically different?


       From: <> AT Internet on 12/30/99 10:13 AM

       Received on: 12/30/99

       To: Peter Constable/IntlAdmin/WCT, <> AT
       Subject: Re: Latin Ligatures and Unicode

       on 12/30/99 12:43 AM, at

> I'd agree with Ken here. I've suggested three minimally
> necessary conditions for characters that control ligation in
> plain text:
> 1) that there are cases where ligation determines a (lexical)
> semantic distinction
> 2) that these cases involve ligatures that are not
> by context
> 3) that these cases involve ligatures that are not optional
> within the given writing system
> It sounds like the example Michael mentions meets only the
> second of these conditions.

       Actually, Michael's point encompasses two and three. That is,
       there are varieties of Latin writing which require ligatures
       for minimal legibility and where the ligature formation is
       sufficiently complex or arbitrary that it must be explicitly

       I think part of the problem is that Arabic and Devanagari are
       really bad analogs for Michael's point -- Syriac would be a
       better example. In Syriac, there are three distinct writing
       "styles", the two most important of which are Estrangela and
       Serto. Writing Syriac in an Estrangela style is relatively
       straightforward, but writing it in a Serto style requires a
       number of rendering rules. These rules are defined in the 3.0
       book even though they don't apply to the majority of actual
       Syriac writing.

       Michael would then be arguing that even though most writing
       styles for Latin don't require explicit ligation control to be
       legible, some do. We need to add ZWL and ZWNL to support those
       writing styles.

       John H. Jenkins

       (BTW, I'm not necessarily agreeing with Michael at this point,
       but I think I
       understand better where he's coming from.)

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:57 EDT