Re: Fonts

From: Hans Aberg (haberg@math.su.se)
Date: Tue Nov 14 2006 - 06:26:46 CST

  • Next message: Werner LEMBERG: "Re: IE7 (off-topic)"

    On 13 Nov 2006, at 21:22, Murray Sargent wrote:

    > Hans Aberg wrote:<unk<unk> For example, if one wants to have
    > superscripts to the left, one has<unk> to write ${}^ba$, which is
    > semantically incorrect, as then the $b$<unk> will be a superscript
    > of nothing, and further, if say the<unk> typographical position of
    > $b$ should be adjusted relative to what it<unk> superscripts, this
    > cannot be done.<unk<unkActually one can define prescripts by the
    > notation<unk<unk_<presub>^<presup> <base><unk<unkwhere nothing
    > precedes the prescripts, but a legitimate base does follow. With
    > the linear format (Unicode Technical Note #28), "nothing" can be
    > ensured by preceding the prescript(s) with an ASCII space, although
    > this prescript notation isn't in the current TN28 paper. This
    > approach is similar to TeX's _1F_1 with no leading {}. On p. 129 of
    > The TeXbook, Knuth does recommend putting a {} before prescripts.
    > But the problem with this approach is, as Hans notes above, that it
    > implies a null postscript base, whereas semantically and
    > typographically one wants the base to be what follows the
    > prescripts. With Word 2007 it's particularly important that the
    > correct base is used, since that enables automatic cut-in kerning
    > to take place between the scripts and their base. Once the notation
    > is built up, the prescripts are associated with the correct base.
    > Subsequent output in MathML or OMML has the correct
    > semantics.<unk<unkMurray<unk<unk<unk<unk<unk<unk

    If iterated, one essentially arrives at math RTL notation, which is
    also used in other circumstances; for example, $f(g(x))$ can also be
    written $((x)g)f$. There are simultaneous sub-/sub-scripts, which
    imply text with no specific reading direction, just as in the case of
    diagrams. There are semantic differences between $a'$: whether it is
    a new glyph (or character) or a letter with an operator on it. Just
    some problems that immediately comes to my mind.

    The advantage of TeX when typesetting math, is that one knows it can
    do the job, even though one has to wrestle with it. Other software
    tend to be geared towards more applied math, so that stuff that is
    essential drops out. But for sure, it would be good with a TeX
    successor, that can do all it can do, without having to wrestle so
    much with programming.

       Hans Aberg



    This archive was generated by hypermail 2.1.5 : Tue Nov 14 2006 - 06:30:06 CST