Re: Latin ligatures and Unicode

From: peter_constable@sil.org
Date: Tue Dec 28 1999 - 10:11:40 EST


       MC>Michael gave better evidence for the need of a ZWL. Although
       he had to resort to ancient scripts like Runic, he demonstrated
       that, in some cases, ligatures may not be defined by any kind
       of rule, being simply a free decision by the author. In these
       cases, there must be a way to encode this scribal caprices, or
       the software will have no other choice than using a "default"
       or "best-fit" rendering.

       But ligatures that are in free distribution with non-ligated
       forms are exactly the kind of thing that ought to be specified
       using style/formatting information applied to the run of
       characters: it does *not* have anything to do with the meaning
       of the text but is purely there for presentation purposes.
       Ligation in this case *can* be handled using such a mechanism,
       selected by the user using some UI device, with the
       OS/rendering engine applying a feature that applys the
       appropriate substitution. In such cases of ligation, it should
       not be a requirement that the ligature be retained when
       exchanged via plain text any more than should the choice of
       typeface, point size, bold, italic, etc. Such presentation
       information does not belong in plain text.

       The "wachstube" example in Fraktur is the only valid example
       I've seen so far of where user-determined control of ligation
       is essential. As Mark Davis has pointed out, there are at least
       two candidate solutions:

       1) use of ZWL and/or ZWNL
       2) use of out-of-band style information selected via some UI

       I.e. it is not clear that the solution to this problem belongs
       in Unicode.

       Marco has suggested an alternate solution for "wachstube" using
       a character that indicates the morpheme boundary, suggesting
       the use of U+200B ZWSP; this can certainly solve the problem,
       but wouldn't a soft hyphen be a better choice?

       Peter



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