Scientific typographic characters

From: Michael Everson (
Date: Mon Nov 11 2002 - 12:17:29 EST

  • Next message: Michael Everson: "RE: Is long s a presentation form?"

     From the NY Times


    The Noah's Ark of the Web, 7,000 Characters at a Time

    IT'S one of the most frustrating problems encountered when passing
    documents back and forth electronically: the little square boxes that
    mean a font someone else used to create the file cannot be rendered
    on your computer. While Portable Document Format, or PDF, files,
    which essentially are copies of printed pages, have helped mitigate
    the problem for most computer users, that solution has not satisfied
    scientists and mathematicians, whose formulas and equations contain
    many symbols.

    Using those symbols on the Web has been particularly inconvenient.
    Most publishers use the symbol-friendly PDF format, but then
    researchers cannot easily embed links to other files or background
    information within those documents as they can with HTML files. But
    HTML documents have their own drawbacks. For instance, they often
    display equations as separate graphic images that cannot be resized
    or searched and greatly increase the size of the file.

    Now a new set of fonts being developed by six publishers of
    scientific, technical and medical journals promises to contain every
    character - more than 7,000 in all - that might be needed in a
    technical article published in any scientific discipline. When
    complete, sometime next fall, the fonts will be shared freely with
    publishers, software manufacturers and scholars, under the condition
    that they not be altered.

    "This work is a breakthrough for publishers and scientists," said Tim
    Ingoldsby, director of business development at the American Institute
    of Physics, one of the publishers working on the project, called the
    Scientific and Technical Information Exchange, or STIX
    ( "The display of math symbols in publishing has
    always been difficult, but those problems have only become worse with
    the Web."

    The set of STIX fonts will work very much like the Symbol or Zapf
    Dingbats fonts in most applications, where users choose from a grid
    of dozens of characters. The STIX font will have the appearance of a
    Times font, but the characters will not look any different if a user
    switches to a different font, like Courier or Helvetica, Mr.
    Ingoldsby said. "The symbols will work with pretty much any font," he

    Mr. Ingoldsby said most scientific characters lack "flavor" - they
    are quite plain to look at - so adding one of those symbols to a
    document composed using, for instance, a serif font, which has fine
    lines projecting from the main strokes of the letter, will not make
    the scientific character stand out. Designers are also adding the
    alphabet, numbers and other common characters to the STIX font, so,
    Mr. Ingoldsby said, there will be no need to switch between fonts.

    "This is meant to replace the font which people use today called New
    Times Roman," he said.

    About 200 characters of the STIX fonts are being finished each month,
    Mr. Ingoldsby said. So far, about half of the 7,000 characters have
    been completed.

    With so many symbols, however, the STIX fonts could be cumbersome to
    use. The developers are working to come up with a method that will
    make it relatively easy for users to find the symbols they want.
    Symbols will probably be organized by type or subject, with the user
    selecting a category (and possibly a subcategory) from drop-down
    menus. A grid of symbols in that category will then appear, from
    which the user can choose the appropriate one.

    Creating a new font set is a complicated process. First, developers
    must correctly copy the shape of each character. Then they must
    adjust its metrics, or how the character is positioned in the space
    in which it is supposed to fit. And finally, they must make another
    set of adjustments to be sure the character looks good on a computer

    William H. Mischo, head of the Grainger Engineering Library
    Information Center at the University of Illinois at Urbana-Champaign,
    said that the STIX project had the potential to solve a problem that
    dates back to the 1400's, when Gutenberg first conceived of movable

    "The two biggest problems since then for properly rendering
    intellectual works have been tables and mathematics," Mr. Mischo
    said. "Here we are in the digital age and we're still having these

    Because math equations have been included in Web pages mostly as
    static images, as either a PDF or a graphics file, scholars have not
    been able to take advantage of many of the Web's distinctive research
    capabilities, Mr. Mischo said. For example, a mathematician cannot
    just plug a particular equation into Google and expect to find other
    scholars working on a similar problem, since the symbols in a graphic
    will probably not turn up in a search.

    "For someone trying to read a scholarly publication, the current way
    of doing things presents difficulties," Mr. Mischo said. "You can't
    enlarge, you can't pull it apart and you can't search it."

    The lack of a comprehensive font for math symbols presents aesthetic
    problems as well. The text in math publications is usually
    unattractive because publishers are often forced to cobble together a
    variety of fonts to create complex equations.

    "Courier may have one set of math characters and Bookman may have
    another set of characters, but they are not going to look good
    together," said Paul Topping, president of Design Science, a company
    in Long Beach, Calif., that makes an equation editor for Microsoft
    Word. "STIX will be a coordinated set of fonts that are meant to work

    Of course, new ideas are always being developed in math and science,
    and some require new symbols. Mr. Ingoldsby, of the American
    Institute of Physics, said STIX will be updated when new characters
    are created.

    "We're trying harder to work with authors so they come up with
    something new only when there absolutely has to be something new," he

    Michael Everson * * Everson Typography *  *

    This archive was generated by hypermail 2.1.5 : Mon Nov 11 2002 - 12:54:08 EST