From: Asmus Freytag (
Date: Fri Aug 07 2009 - 00:16:59 CDT

  • Next message: Apostolos Syropoulos: "Re: Guidance wanted on implementing Greek context-dependent casing"

    On 8/6/2009 8:07 PM, Doug Ewell wrote:
    > Asmus Freytag <asmusf at ix dot netcom dot com> wrote:
    >> However, this loose talk about "glyph variants" spooks me. The
    >> *identity* of a logo is in its *appearance*, not in the organization
    >> it symbolizes. From a character encoding perspective, if the new logo
    >> looks different, then it's not a variant glyph of the logo encoded by
    >> the existing character, but a new, *unencoded* entity.
    > Does Unicode now freely admit to encoding logos?
    Hmm. Quoting out of context is just so much fun.

    Just before the passage you quoted, I wrote:

    "Peter is correct in that adding a character would have to surmount some
    really substantial hurdles, because of the fact that UTC and WG2 have
    become more firm over time in their rejection of logos as encodable
    entities. "

    I think that makes clear that nobody is "freely admitting" anything.

    Perhaps you were confused by my calling the new logo an "unencoded
    entity". Calling it that merely recognizes that is is *not* already
    encoded. By itself, it says nothing about whether it should be encoded
    or whether it even can be encoded.
    > I know, the old JIS logo was probably encoded as a compatibility
    > character.
    You can drop the "probably". It's a compatibility character.
    > The new one could not be, unless:
    > (a) there is a character set somewhere that encodes the two
    > separately, or
    > (b) all pretenses to "compatibility" are abandoned.
    I think we are all in agreement on this one.

    I'm going a step further by noting that for a logo it is really not
    appropriate to think of multiple logo designs for the same entity as
    "glyph variants" of each other. (*) In my view, that is true whether you
    are considering "implicit" glyph variants, such as a character might
    exhibit across different font designs, or explicitly defined glyph
    variations using variation selectors.

    Whenever you encode something (or try to decide whether something is
    already encoded) you need to look at *what* has been encoded, in other
    words, the definition of the *identity* of the encoded entity. For logos
    (and for many symbols) that identity is tied very closely to the
    appearance (more closely than is the case for ordinary letters or
    punctuation). For logos, the identity is their appearance. That, in
    addition to the IP problems with logos, is the reason why they encoding
    them as characters is awkward (and why neither UTC nor WG2 are keen to
    add any).


    (*) except if the designs really are basically only typographical glyph
    variants, of course.

    This archive was generated by hypermail 2.1.5 : Fri Aug 07 2009 - 00:20:36 CDT