RE: Behaviour of ZWJ that PR-37 has not considered

From: Peter Constable (petercon@microsoft.com)
Date: Tue Sep 01 2009 - 14:09:51 CDT

  • Next message: Harshula: "Re: Run-time checking of fonts for Sinhala support"

    The PRI-37 doc did not propose any particular behaviour for Cx+V+Z+C2 sequences other than the case in which Cx is a C1 consonant. Likewise, no behaviour was defined for C1+Z+V+Cx sequences except for the case in which Cx is a C2 consonant.

    I would expect that if such sequences were encountered then implementations would likely not produce "level 1" ligature conjuncts; I don't see any problem with that behaviour. Less clear is whether implementations would produce "level 2" behaviour or "level 1" behaviour; it would seem that "level 1" would be expected since, by definition, there isn't really any well-defined "level 2" behaviour (for Cx+V+Z+C2, Cx is not a C1; for (C1+Z+V+Cx, Cx is not a C2). That seems perfectly reasonable to me.

    You, however, are suggesting that C1+Z+V+Cx should produce "level 2" C1-conjoining when Cx is not a C2. At present, I can't think of anything in Unicode that would explicitly deprecate such behaviour, but I wouldn't expect it to just happen in implementations and I also don't think it's a good idea: it would mean that there are two different sequences that produce the same display. In general, that is not a good thing because it can lead to user confusion and can create a security risk. And the user confusion would be all the greater because the Z+V pattern would not behave consistently across all consonant combinations.

    So, I think I would be inclined to suggest a different position than either your "relaxed" or "rigorous" positions: as with your "rigorous" position, Z+V should not be disassociated from C2-conjoining, but instead of Z being ignored, it explicitly is not ignored; and since the sequence does not fit the patter for level 1 and level 2 behaviours then only level 3 behaviour, overt halant, is permitted. (Likewise for Cx+V+Z when Cx is not C1.)

    Peter

    -----Original Message-----
    From: Shriramana Sharma [mailto:samjnaa@gmail.com]
    Sent: Monday, August 31, 2009 1:22 AM
    To: unicode@unicode.org
    Cc: Peter Constable
    Subject: Re: Behaviour of ZWJ that PR-37 has not considered

    Hello all.

    No comments on my previous message on this thread:
    http://www.unicode.org/mail-arch/unicode-ml/y2009-m08/0197.html ?

    -- 
    Shriramana Sharma
    


    This archive was generated by hypermail 2.1.5 : Tue Sep 01 2009 - 14:13:52 CDT