Re: Why people still want to encode precomposed letters

From: Andreas Stötzner (
Date: Tue Nov 18 2008 - 13:04:54 CST

  • Next message: "Re: Why people still want to encode precomposed letters"

    Am 16.11.2008 um 10:04 schrieb Andrew Cunningham:
    > Actually it is that simple. For a font to work, there needs to be
    > anchor points and attachments for the base character and  combining
    > diacritic. …

    Not neccessarily. Andron Mega has no anchor points at all and works
    well with almost any combination in TextEdit under Mac OS X as some
    other fonts do as well. It works even on the web in the highly
    specialised diplomatic online edition of medieval Old Norse texts
    (google MENOTA) which is loaded with combinant characters.

    > Microsoft typography team did something similar when they developed
    > the repertoire of base character / combining diacritic support.

    The problem is, that support of such kind is not regularized. As
    Michael Everson pointed out, there are no easily accessible guilelines
    for font producers, something, which really *would* be recognized by
    many designers. All this Opentype and -table and -anchor stuff is
    *arcanum* to normal people and does not help to make the world better.
    It is just far too byzantine.
    Maybe this situation is policy, too. However, it is not supportive for
    all font providers except MS or SIL (e.g., 99,99% of them).
    > The combinations of base character and combining diacritic(s)
    > supported is a decision of the font developers.
    A font developer will never be able to trace every possible combination
    eventually needed by an end user.

    > No font will support all theoretical combinations, maybe ot even most.
    > What they will try to do is add the combinations that fullfill the
    > purpose fo the font.
    Dead theory, again. There is no (single) “purpose of the font”. Fonts
    have to serve multiple purposes, as encoding does.

    It is highly unrealistic to assume that at least some of the principal
    fonts will come with sufficient anchor point programming ever. Few
    well-sponsored specialists like John Hudson may be so lucky to labour
    on this for months, all others (including MS fonters, I fear) can not,
    because it is merely impossible to get paid that effort.

    Even if one *would* set up some well-defined recommendation to deal
    with combinations-to-be-expected, there will always be unexpected ones
    around the corner. It will never work sufficiently. NEVER.
    The only way I see is that the big implementors realize that officiant
    encoding business and professional typography belong together and that
    rendering is the gap to be filled.
    Just insisting on policies and leaving the trouble to others has lasted
    too long.

    with regards,
    A. Stötzner.

    Neu: Die Andron gibt es jetzt in Griechisch!
    Rundbrief – Signographischer Anzeiger

    Andreas Stötzner Signographie
    Signographisches Institut Andreas Stötzner i.A. (Pegau/Sa.) Tel. +49-34296-74849 Fax +49-34296-74815
    Willkommen auf

    This archive was generated by hypermail 2.1.5 : Tue Nov 18 2008 - 13:05:22 CST