Oldstyle numerals are a stylistic variant that isn't needed to preserve
legibility in plain text, nor is it needed to prevent false positives in
plain-text searches, so I don't see any reason to encode it in Unicode.
It's a nice example of higher-level formatting.
Whether to include multiple sizes is a trickier question that is still under
debate. Unicode has (or will have) n-ary operator characters for all
standard math n-ary operators, e.g., summation, integral, product, union,
etc. The size of braces and n-ary operators can be derived to some degree
by algorithms. The user could overrule the algorithms if desired. Plain
text using a single size for the various different size braces (parens,
brackets, etc.) is legible; the size variations just make things prettier
and easier to read. So most probably we won't encode them, although they
are requested in the STIX proposal.
On the other hand, the brace, integral, and bracket pieces will most
probably be encoded, even though they aren't part of the STIX proposal. The
primary reason is their appearance in a number of legacy character sets,
including TeX, HP Math8 character set, and Windows' symbol font. Another
reason for their inclusion is the ease of use in communicating with printer
drivers. Note that TeX and other math typesetting systems use such
characters to create large integrals, brackets, braces, and parentheses, not
for ease of implementation but rather for correct color. The composite
large characters look better than the corresponding scaled characters, since
the latter look too bold.
Murray
> -----Original Message-----
> From: Scott Horne [SMTP:shorne@metaphasetech.com]
> Sent: Friday, July 02, 1999 12:20 PM
> To: Unicode List
> Cc: Unicode List
> Subject: Re: Superscript asterisk
>
> > Because the AMS, MathSci TEX, STIX, and math fonts in general have
> consistently
> > made distinctions between such entities as phi and straightphi,
> > epsilon and straightepsilon, etc.
>
> TeX fonts also encode several sizes of grouping characters (such as
> parentheses), radical signs, and other characters, and components
> from which to construct arbitrarily large ones. They distinguish
> text figures (so-called \oldstyle numerals) from lining figures.
> Will all these system-specific entities get space in Unicode, too?
>
> Scott Horne
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:48 EDT