Re: [css-fonts-3] i18n-ISSUE-287: Figure 4 - combining diacritics

--On Thursday, September 12, 2013 18:07 -0700 John Daggett
<> wrote:

> Richard Ishida wrote:
>> -background Typography Background
>> Figure 4 & "The use of combining diacritic marks creates many
>> variations  for an underlying letterform"
>> Not sure what is the intent of fig 4. If it's to show how a
>> font has to  cater for variation in placement and shape of
>> glyphs, it might be better  to use examples of tone marks in
>> Thai and medial consonants in Burmese,  etc. I can supply
>> examples if needed.
>> This would also introduce the idea of complex scripts, which
>> I thought  was lacking in this section.
> It's simply there to illustrate that basic letterforms are
> often combined with diacritics.  I do think it would be
> interesting to go into this in a little more depth and add
> examples for complex scripts but, given the late date and the
> informative nature of this section, I think this should wait
> until the next level.  Any examples and/or illustrations you
> can provide in that regard would be most appreciated.


Based on too many painful experiences (mostly outside W3C),
there are big problems with people who think they understand
decorated Latin characters to assume they can extrapolate from
that understanding to all i18n issues.  Because the way most
common languages that use Latin scripts do so, those people tend
to see strings consisting of large percentages of undecorated
characters with a few decorated ones mixed in.   They also tend
to believe, based on extrapolation from their limited
experience, that the maximum number of decorations per character
form is one.   Another problematic inference is that one
decoration is just about as good as another because the reader
can and will deduce what is needed from context.

While the ligature example may be ok, I suggest that any "how to
decorate a character form" example that confines itself to
diacritic marks in Latin script isn't just "informative" but
actively harmful because it encourages the "other things that
use built-up character forms are just like Latin" assumption.

Suggestions, which I hope are consistent with "late date and the
informative nature...":

(1) Add at least one Latin character with more than one
decoration on it.  

(2) Add a second line of examples that draw from at least one
non-European script.  I think it would be good to choose some
examples, that do something other than modify a phoneme or
provide a phonemic break, but perhaps that is a future version
issue.  Other that that, I don't think it makes a lot of
difference what examples are used, but Richard or others may
have better experience and examples than I do.

(3) Change the title of Figure 4 from "Variations with diacritic
marks" to "Variations with diacritic and other marks" or
something similar.  I'd be tempted to use "decorations", but I
don't know if your readership would be familiar with that term
and its strictly non-pejorative and non-dismissive uses.

I agree that more explanation, or even explanation of those
examples, can wait for "the next level" or future revisions.
But I think it is very important to at least provide strong
hints that this isn't just about Latin and that extrapolation
from Latin might cause issues.


Received on Friday, 13 September 2013 12:30:00 UTC