Although there is general
understanding that ASCII code 2/1 prints as a vertical stroke (perhaps
of somewhat varying thickness) with a dot (perhaps not exactly circular)
a short distance below it, printing it as a vertical bar of uniform
thickness with no dot is *not* a violation of ANSI X3.4.
From the ISO's point of view, it does not matter at all if you
print 'A' with the glyph of 'B' and 'B' with the glyph of 'A', or even
print all of 'A', 'B', 'C',...,'Z' with the glyph of 'A'.
So, it is pointless to pursue low end of the quality.
The point of code character set is that, if the receiving side has a
proper equipments, and if enough profiling information is given,
the received data can be printed so that native users of the language
written can be satisfied with its quality.
Also in particular, I suspect attempts to write
a fairly narrow definition are going to end up being futile, and the
time is best spent on more substantive issues.
As Keld demonstrated with RFC1345, giving proper meaning to "charset"
so that no further profiling is necessary is quite doable.
Masataka Ohta