RE: Symbols for chemical bonds

From: Ernest Cline (
Date: Wed Mar 30 2005 - 10:16:30 CST

  • Next message: Marion Gunn: "Re: U+0023 in French"

    I goofed and originally sent this just to Jukka

    Ernest Cline

    > [Original Message]
    > > From: Jukka K. Korpela <>
    > >
    > > Which Unicode characters would be most suitable as symbols for chemical
    > > bonds? By visual appearance, one might think about using
    > > - em dash (or en dash) for single bond
    > > - equals sign for double bond
    > > - identical to for triple bond
    > > - strictly equivalent to for quadruple bond
    > > but these would differ quite a lot from the usual semantics of the
    > > characters (except for the em dash, which is rather general-purpose
    > > anyway). Besides, there would hardly be any guarantee that all the
    > > would be of the same width in a given font.
    > I would expect (and a casual inspection of a few fonts supports my
    > expectation) that minus sign, equals sign, identical to, and strictly
    identical to
    > would all have identical widths. However...
    > > I found the document "Elsevier Science Grid in Unicode",
    > >
    > > which discusses the mapping of a large number of symbols
    > > to Unicode characters, and it contains no Unicode equivalents
    > > for the bond symbols.
    > ... the ESGrid provides several complications.
    > First, for the double, triple, and quadruple bonds, it contains two
    > that vary only in size. One set is described as em dash in width (which
    > set of characters I gave above are usually narrower than) and the other is
    > described as being 6 point.
    > Second, it provides for three glyph variants for a partial double bond
    > and two for a partial triple bond. One of the partial double bond
    > is equated in the ESGrid to U+2393. This indicates that trying to
    provide a
    > one-to-one mapping from the ESGrid to Unicode might not be desirable.
    > Finally, the ESGrid has bond "characters" such as "lbond3" that are
    > clearly intended to serve as graphics primitives and not as part of linear
    > text as they serve to point to items that aren't in the same line of text.
    > > Since bond symbols are often used in chemical formulas inside texts
    > > it would seem natural to encode them as characters.
    > I would tend to disagree. Chemical bond indicators appear to be
    > graphical and not textual. In particular, the fact that they cannot be
    > confined to a line of text causes me to reach that opinion.
    > While there may be some things in the ESGrid worth encoding in Unicode,
    > the chemical bond glyphs are not among them in my opinion.

    This archive was generated by hypermail 2.1.5 : Wed Mar 30 2005 - 10:17:58 CST