Re: Custom fonts (was: Tolkien wanta-be)

From: Chris Jacobs (
Date: Tue Mar 18 2003 - 11:54:01 EST

  • Next message: John Hudson: "Re: U+00D0, U+01b7 -- variants or distinct chars?"

    ----- Original Message -----
    From: "Pim Blokland" <>
    To: "Unicode mailing list" <>
    Sent: Tuesday, March 18, 2003 12:43 PM
    Subject: Re: Custom fonts (was: Tolkien wanta-be)

    > Chris Jacobs schreef:
    > > If I interpret a B font declaration on a webpage as a private
    > > agreement
    > > that for data in that font on that webpage a PUA will be used were
    > > U+E000 is
    > > a banana that does not imply that I claim anything about which PUA
    > > I use for
    > > other purposes.
    > You keep making it more and more difficult for the rest of us to
    > follow you.
    > To start with, your use of "PUA" is not what is generally meant by
    > "PUA".
    > There is no such thing as "a" PUA; there is only one Private Use
    > Area, which consist of all the codepoints that will not be assigned
    > a specific use by the Unicode Consortium. You cannot say that the
    > range of codepoints U+E000..U+E0FF is a different PUA than
    > U+E100..U+E1FF.

    I don't say that the range of codepoints U+E000-U+E0FF is a different PUA
    than U+E100-U+E1FF.
    I do say that if a webpage has U+E000 defined as banana and I have it
    defined as apple, that then their range U+E000-U+F8FF is a different PUA,
    belonging to a different extension of unicode than my range U+E000-U+F8FF

    > Secondly, you must be aware there is not, and will not be, a rule
    > about what characters in that area should look like. Yet you insist
    > on trying to convince everybody it's a good idea to remap, for
    > example "banana" to U+E100, even if the font calls it U+E000. You
    > keep on about what a good idea it would be to be able to rearrange
    > code points such that no matter how many fonts you have in use,
    > there is always a banana at U+E100. This is a restriction, an
    > unwelcome intrusion on the PUA!

    Being able to do things is not a restriction.

    > You also can't seem to decide if this is just something you want to
    > do on your own computer, or if you also want to use this scheme for
    > information interchange with other users.

    That is not something that should be decided upon here once and for all. If
    I want to use this scheme for information interchange with some other users
    then I establish private agreements with those other users.

    > Now what you do in the privacy of your own home is none of our
    > concern, but when communicating with the outside world, there are
    > certain rules and guidelines you should abide by. And one of those
    > guidelines is a plaintext file should not have PUA characters in
    > them, unless its author also specifies it should be displayed using
    > a certain font.

    Nope. Specifying a font is just one method of specifying a private
    agreement, but certainly not the only possible one.

    Suppose I want in my PUA control characters, like a <MOVE> or a <COPY>
    system support operation, How do you want to define those characters in a

    > Now if the font it should use is known, the proper
    > codepoint to display this banana of yours is also known, because
    > this info is in the font. Ergo, no need to remap!
    > Since not all fonts have a banana, it really doesn't make much sense
    > to not specify a font. The computer wouldn't know what to do!
    > Lastly, I must say I think it's a pity that the suggestion I made
    > yesterday has been ignored so quietly. You know, in a HTML
    > environment, to retrieve names for characters from the font file
    > itself, to relieve the author from the task of having to enter
    > numerical values.
    > For an example, suppose you have a font named "Tengwar Quenya", with
    > a character named "hwesta" at U+E00B,

    Then you would no longer be able to specify the font by the font name, since
    you would have two fonts named "Tengwar Quenya".

    > you could use it in an XML
    > file by defining an entity, <!ENTITY hwesta "&#xE00B;">. Now my
    > suggestion was the browser program which displays this file should
    > be able to look at the font information in the XML file, open the
    > font file and retrieve the names of all characters in it, so it can
    > show the "&hwesta;" character (and all other characters) without
    > needing a long list of ENTITY entries in the XML.
    > Anyone else think this would be a good idea?
    > Pim Blokland

    This archive was generated by hypermail 2.1.5 : Tue Mar 18 2003 - 11:30:42 EST