Re: ZWJ, ZWNJ and VS in Latin and other Greek-derived scripts

From: John H. Jenkins (
Date: Fri Jan 26 2007 - 23:43:54 CST

  • Next message: "Re: writing Chinese dialects"

    On Jan 26, 2007, at 5:31 PM, Ruszlan Gaszanov wrote:

    > Asmus Freytag wrote:
    >> However, from that insight to the particular mechanisms of encoding
    >> that
    >> you advance is a large step, and I don't follow you there.
    > The point I'm trying to make is that rather then investigating every
    > possible ligature in Latin, Greek, Cyrillic, Coptic, Gothic etc.,
    > trying to figure whether it is a mere stylistic variation or
    > distinct character in each particular orthography variation, and
    > then encoding them as recomposed characters in the latter case, we
    > could just provide a standard mechanism of encoding any arbitrary
    > ligature (like the standard mechanism we have for applying any
    > arbitrary diacritics to any arbitrary letter).

    But that's a straw man. Nobody is suggesting that every possible
    ligature be encoded. Nobody is suggesting that *more* ligatures be
    encoded. The ligatures which are in Unicode are there to cover a
    specific need which does not generalize to ligation in general.

    There is general agreement that having ZWNJ mark where ligation must
    not take place is a reasonable suggestion. There is also general
    agreement that having ZWJ mark exceptional places where ligation is
    obligatory is reasonable.

    The major push-back is from people who object to the idea that ZWJ
    should be used to mark where ligation would be desirable.

    And the big can of worms is how to deal with particular styles of text
    (like fraktur or older typography in general) where the standard
    ligation rules are not as simple and regular as they tend to be in
    modern western typography. My own feeling on the subject is that once
    you decide to set a text in fraktur (or italic or roman) type, you're
    making a stylistic choice which need not be supported in plain text.
    I do, however, look forward to a UTN or UTR on the subject.

    John H. Jenkins

    This archive was generated by hypermail 2.1.5 : Fri Jan 26 2007 - 23:45:29 CST