Re: Margin kerning

[Michael Piotrowski]

Hi,

address@hidden (Ludovic Courtès) writes:

>> I agree.  In my test document it actually happens twice in one
>> paragraph that the line ends with an ellipsis, and the ellipses in the
>> margin look strange and might even be misleading.
>
> Or maybe people should avoid using ellipses?  ;-)

Well, that's probably asking too much, unless you've got a better
alternative ;-)

However, ellipses at the end of a line may be misleading in general;
the following is an excerpt from my test document:

  [...] by modern painting can be synthesized in the new, un-
  fettered institution of advertising: "All these findings
  came to focus in the practical tasks of contemporary ad-
  vertising art. Advertising could utilize them because it
  was not handicapped by traditional forms... it belonged
  to its very nature to be contemporary and forceful...
  the most heterogeneous elements--verbal message,
  drawing, photography, and abstract shapes--were
  employed... Posters on the streets, picture magazines,
  picture books, container labels, window displays...
  could disseminate socially useful messages, and they
  could train the eye, and thus the mind, with the neces-
  sary discipline of seeing beyond the surface of visible
  things, to recognize values necessary for an integrated
  life..." (Kepes, Language of Vision, 221)

The ellipses after "forceful" and "displays" make the passage
especially hard to read, I think.

>> It might be interesting how this is handled in InDesign, but I don't
>> have easy access to a Windows PC or Mac to install the demo version.
>
> I don't either.

Too bad ;-)

>> A per-document user-editable table (i.e., not per-font) might already
>> provide more than enough customizability.
>
> I would even say "more than needed".  ;-)  I'm not sure many people would
> be willing to specify this sort of parameter by hand, for each and every
> character.

I was unclear: I was thinking of a table specifying *only* the glyphs
to be margin-kerned, *maybe* with the desired widths (for tuning in
case of unusual character designs).  All other characters would not be
subject to margin kerning.

> Another solution might be to add heuristics to avoid margin-kerning
> glyphs that are "unreasonably wide".  "Unreasonably wide" could be
> defined as "wider than `x'" for instance.  This same heuristic could be
> applied to subsequent characters also.  For instance, a quote followed
> by a comma may be narrower than `x' and therefore eligible as a whole
> for margin kerning.  On the contrary, the ellipsis glyph may be too wide
> and will never be margin-kerned.
>
> I'll give it a try unless you think it doesn't make sense.

Sounds good, too.  However, I'm a bit wary of fully automatic
formatting based on heuristics: If you're using a font with unusual
character designs or if you've got certain language-specific
combinations where the heuristics fail, you'd have to turn off margin
kerning for *all* characters.  That's a bit like automatic use of
ligatures: It works fine for English, but in German, where you don't
want ligatures to span syllable boundaries, you either have to check
every ligature manually or turn off ligatures in general.

But these are just some idle thoughts ;-)

Greetings

-- 
Michael Piotrowski, M.A.                               <address@hidden>
Public key at <http://www.dynalabs.de/mxp/pubkey.txt>