RE: (SC2WG2.609) New contribution N2705

From: Peter Constable (petercon@microsoft.com)
Date: Wed Feb 18 2004 - 11:34:10 EST

  • Next message: Peter Kirk: "Re: Fwd: Re: (SC2WG2.609) New contribution N2705"

    > From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org] On Behalf
    > Of Michael Everson
    > Sent: Wednesday, February 18, 2004 4:48 AM

    > There are UPA "wildcard" characters used in analogous description of
    > vowel harmony: U+1D3D for instance. Letter x is hardly different.

    So, 1D3D means o or u? I suspect if you were proposing some kind of subscripted a/o ligature that was in conventional usage that Ken might be offering a different opinion.

    Note that I floated LATIN LETTER SMALL CAPITAL I OVER SMALL SCHWA and LATIN LETTER SMALL UPSILON OVER SMALL SCHWA in a proposal last year (cf. L2/03-190r), and they were tossed out on their serifs (these letterforms don't have ears) on the grounds that these should be considered rich-text representations.

    There is no question that the field of linguistics needs devices to represent variability in vowel qualities.

    One possible solution is to devise ligated forms based on symbols for the extremes of variation (which I gather is the case for 1D3D). If these become convention, they could be considered for encoding. (I would *not* consider these merely presentation forms.)

    If (non-sub/superscripted) ~ or / are used, those obviously can be represented.

    When we get into devises such as stacking reduced-size symbols within the x-height, or adding super/subscripted symbols as algebraic devices, things are less clear, and we need to proceed with caution.

    And whatever decisions we make in these cases I would not extend as precedents to mathematics or other fields using technical notations, including usage in linguistics for anything beyond representation of speech sounds.

    > >what is to guarantee that I won't find alternative representations
    > >of such formulations using "~" instead of "/", for example? Do
    > >we then also need a subscript tilde to handle that?
    >
    > We are not proposing to encode something which we have not
    > encountered, and do not accept the burden of proof about what you
    > will or will not find. That's not right, Ken. That's not how we have
    > ever proceeded. People come with requirements and evidence. We don't
    > throw that out because of what can be "imagined".

    It's not a matter of proposing to encode something which hasn't been encountered; it's a matter of trying to figure out how something should be categorized. Can the subscript / *really* be considered to be the same kind of beast as a, e and all the other symbols for speech sounds (albeit a fringe instance), or is it *really* the same kind of beast as ~ = > $B"*(J { } [ ] etc.?

    If the latter, then we would have to be consciously deciding that we want to permit an exceptional, non-precedent-setting case of encoding a subscript form of a symbol of this type. Perhaps we might be able to convince ourselves to do this *if* (i) it is clear that this is an established convention used widely in some branch of linguistics, and (ii) that this is particularly motivated because this particular symbol is one of a small fixed set used in a way that is closely integrated with modifiers that represent speech sounds.

    Peter
     
    Peter Constable
    Globalization Infrastructure and Font Technologies
    Microsoft Windows Division



    This archive was generated by hypermail 2.1.5 : Wed Feb 18 2004 - 12:24:03 EST