Re: how to add all latin (and greek) subscripts

From: Ondrej Certik (
Date: Tue Jul 01 2008 - 14:40:57 CDT

  • Next message: philip chastney: "UTR 25 - Unicode Support for Mathematics"

    On Tue, Jul 1, 2008 at 6:07 PM, John H. Jenkins <> wrote:
    > On Jul 1, 2008, at 1:24 AM, Ondrej Certik wrote:
    >> Clearly, one needs to draw a line somewhere between the plain text and
    >> a markup. The unicode has drawn the line in such a way to put all
    >> latin letters to one side and "q" to the other side. That is just
    >> ridiculous in my opinion. Who will use the current "crippled" support
    >> for superscripts now? I think noone. So maybe we should propose to
    >> remove them from unicode altogether. I am actually not against,
    >> because the current state clearly encourages one to use it, as you
    >> could have seen in our case. If there wasn't already a partial support
    >> for this in unicode, this discussion would not be here. So let's fix
    >> that. Either way.
    > I'm sorry, but you're really missing the fundamental point. Unicode doesn't
    > want people to use "the current 'crippled' support for superscripts",
    > because the "superscript" letters aren't there to provide support for
    > superscripts.
    > The existing super-/subscripts are in Unicode for a reason, and that reason
    > is *NOT* providing superscripting per se. They're available either because
    > of the need for round-trip compatibility with older standards or because of
    > some specific needs of some writing systems such as IPA, which use
    > characters like "ʰ" as something other than "a superscripted h." The fact
    > that ʰ is typographically derived from and still looks like a superscripted
    > h is irrelevant. None of these requirements happen to include anything that
    > looks like a superscripted q.
    > If you really want a superscripted h, use rich text. Using ʰ for a
    > superscripted h is, from Unicode's perspective, as bogus as using 𝑎
    > (U+1D44E) for an italicized a.
    > (And for the record, I don't find things like H₂O violently objectionable
    > except insofar as people infer from them that Unicode has deliberate but
    > half-baked support for super-/subscripts. In the long run, it's bad
    > practice to do things like that, but it's on the same order as using three
    > periods for an ellipsis or two hyphens for an em-dash.)

    Thanks a lot John for taking part in the dicussion (I hope you won't
    regret:). Your arguments are very good ones and I am starting to
    understand why the current situation is as it is. It is mainly for IPA
    and compatibility reasons, but that's it. No subscripts or
    superscripts. Thanks also to David and all others to their point of
    view as well.

    Especially thanks a lot for your opinion why you think the proposal
    would not succeed. Nevertheless, I can provide answers to your points:

    1) That is true.

    2) All distinct latin letters and greek. The reason is that those are
    used in mathematics and physics. Noone uses accented letters, hebrew,
    or Cyrillic (in sub/superscripts in mathematical formulas).

    3) The intent is to allow latin and greek sub/superscripts, as those
    are the ones used in physics/mathematics. One should not put
    subsubscripts, as there is no such support (half baked) in unicode.

    Ok, anyway, my problem is to have sub/superscripts of latin and greek
    in a terminal. Do you think I have a chance of succeding in extending
    the escape sequences for terminals?

    I know it's offtopic for this list, but nevertheless, that is my real
    problem. I thought a solution to this could be by extending unicode. I
    learned that it is probably not a good solution. So I am trying to
    find a different solution to my problem. I am sure someone on this
    list has a feeling about how such processes go and if it has a chance.

    Thanks a lot again to everyone who took part in the discussion,

    This archive was generated by hypermail 2.1.5 : Tue Jul 01 2008 - 14:43:57 CDT