Artifact 4f3c811750a878d78ec1c88a0d876bb0e41696ace161cd9eddac099dc1c809e1:

• Executable file r38/lisp/csl/cslbase/fonts/README — part of check-in [f2fda60abd] at 2011-09-02 18:13:33 on branch master — Some historical releases purely for archival purposes

  git-svn-id: https://svn.code.sf.net/p/reduce-algebra/code/trunk/historical@1375 2bfe0521-f11c-4a00-b80e-6202646ff360 (user: arthurcnorman@users.sourceforge.net, size: 2700) [annotate] [blame] [check-ins using] [more...]

The versions of the fonts here have been MODIFIED by A C Norman
(September 2004) in an attempt to circumvent what appears at the
time of writing to be an X server anomaly that caused some characters
to fail to appear on the screen. The modification was performed as

Also the names specified for the fonts seemed to make it ghard for me
to use fontconfig/Xft to select them, so I changed names.

(a) Use t1disasm to convert the fonts from binary (*.pfb) to human-
    readable form.
(b) Remove the AMS copyright notices as required by their terms-of-use
    document. Observe that this is not just permitted but required!
    [required by AMS]
(c) Change so that the "family name" fully identifies that font, and all
    names for the font are in lower case.
    [seems to be required so that fontconfig/Xft/freetype can match the
    font in an unambiguous way]
(d) Duplicate the definition of the first glyph defined by textually
    copying its definition, but giving that Postscript procedure
    a new name (eg /acnminus).
    [The x.org server on Cygwin and several Unix systems appears not to
    be able to render ?the first glyph defined in a type1 font?. Even if
    this is an X bug and it gets fixed there are liable to be broken X
    server configurations in use for some time so a work around by using
    a custom font seems safer to me]
(e) Altering the setting for symbol 161 to be the new not the old
    version of the definition, but NOT changing the version for
    character code 0.
    [There are copies of glyphs with raw code 0-31 and 127 mapped at
    161...]
(f) Use t1asm -b to convert the font back to binary packed form, and
    t1asm to nake an ascii packed form.
    [The *.pfa forms are for embedding in documents for printing.]
(g) Use type1inst to create a fonts.scale (etc) index file. Review
    these by hand.
(h) Use fontforge to create Truetype versions of the fonts, ensuring tha
    the truetype styles (sub-family) give the name I want (eg cmr10).
    See fontforge.sf.net.
    [On Windows I explictly use the Truetype versions, in part because I
    expect that to support old versions of Windows best. I am aware that
    my truetype versions of the fonts will not be as well hinted as would
    be perfect. But anybody with a better truetype version of these fonts
    could just drop the truetype file in to my fonts directory. I keep all
    the various formats present across all platforms to make it easier for
    me to manage things, but the type1 fonts are used on Unix/Linux/Darwin
    and the Truetype ones on Windows.]

This process should not change the shape, matrics or encoding of any
character.



             A C Norman.  October 2004