Version 0.1.2 is a bug-fix follow-up to 0.1.1, and is
binary-compatible with it:

    * The symmetry- and spiral-generators' weights where accidentally
      coupled in mkvisualid, such that "--symmetry" set weights for
      *both* the symmetry generator *and* the spiral generator, and
      "--spiral" did nothing.

    * Decreased the probability of child-glyphs being used in the
      along-a-path generator from 100% to 80%, in accord with the
      table on page 10 of the VisualIDs essay.



Version 0.1.1 is a minor (bug-fix) follow-up to the initial 0.1
release, correcting various minor errors in both design and
implementation and adding (or enabling) minor features that were
missing (or disabled) in 0.1, for example:

    * The `spiral' generator has been enabled in the `mkvisualid'
      utility, with a "--spiral" ("-i") option to control its weight
      in the randomiser.

    * In general, more care has been taken to ensure that, by default,
      the glyph-generation algorithms used by the mkvisualid utility
      are identical to those used by the library routines.

    * The "-s" abbreviated option was mistakenly aliased to both
      "--symmetry" and "--shape"; "-s" has been granted to "--shape",
      with "-y" being the new shorthand for "--symmetry".

    * There is a new "--outline-width" ("-O") option that allows
      specification of the thickness of the white outline that
      provides contrast when glyphs are rendered over dark
      backgrounds; this was previously fixed at 50% of the foreground
      stroke-width.

    * The `perfect fit' logic that automatically scales glyphs to fit
      perfectly on their canvases now works significantly better, and
      even supports non-round line-joint/-cap types.

libvisualid 0.1.1 is binary-incompatible with 0.1, but should be
source-compatible in every meaningful way; the libtool library-version
has been updated so that 0.1 and 0.1.1 shared libraries can coexist
for the benefit of any applications that have already been built with
and linked against 0.1.
