Re: Features needed for high-end DTP apps

[Joaquín Cuenca Abela <cuenca pacaterie u-psud fr>]

On Sat, 2002-10-19 at 23:25, Patrick wrote:
> Hi Joaquin.
> 
> > If your goal it's top quality DTP, then you should also try to do
> > optical kerning.  The literature about how to do acceptable optical
> > kerning is almost inexistant, but if you're looking for something that
> > can compete with InDesign, you need it.
> 
> I'm not clear on what optical kerning is. The example shown in this
> article: http://www.adobe.com.au/print/features/indesign/main.html
> suggests optical kerning shows the kerning value between each glyph and
> allows you to visually drag the line representing the kerning amount to
> visually change the spacing. There must be more to it than this ...

It has nothing to do with that.  Optical kerning is to kern a pair of
letters based in the shapes of the letters, and not in a hand crafted
table.  That's specially useful when such table doesn't exist (for
instance, kerning between letters of different fonts).

Everybody that has used TeX knows about the command "\/".  You use it to
add a bit of space between an italic letter and a roman one... except
when the roman one is a ".", a ",", etc.  That's because TeX doesn't
know how to kern letters of differents fonts (almost nobody knows how to
do it).

After searching for weeks, the best that I've got about how to do an
algorithm to do optical kerning that works are rumors.  First, the naive
approaches are:

1) equalize the minimal distance between the sides of each letter.  It
sucks.
2) equalize the area between the letters.  It also sucks, but a bit
less.

The only reference that I've found of a book that explains several
approaches is "Optical letter spacing for new printing systems", from
"David Kindersley", but it's too expensive to me.

It seems that the book explain that using the forth momentum of the
letters to calculate the kerning gives acceptable results.  I've also
heard that people don't started to get really good results until they
started to play with the hole between the letters, instead of playing
with the shapes of the letters.

Again, all that is nothing more than rumors.

> > >  o Tracking - customizable for fonts based on their various sizes,
> > >    and also for particular paragraphs. Again, some scripts may prefer
> > >    to extend characters rather than add space.
> 
> As is probably evident by now, I'm not entirely familiar with DTP
> requirements. Can somebody elaborate on tracking? I'm not clear on the
> specifics of what it accomplishes.

it's the general letter spacing.

> > If you're looking for top-of-the-crown quality, I have some ideas to add
> > to that list, but imo almost all of that stuff should be out of pango. 
> > You should also be aware that many people has tried before to make a
> > high level, high typographic quality library, without success (I guess
> > that big projects are too attached to their implementation, to throw it
> > out and pick another one that any cheap application can have).
> 
> Can you elaborate on your top-of-the-crown quality list?

There is too much stuff.  To start with, pick current pdfTeX, and add:

* Non linear positioning of floating figures.  There has been a bit of
research here, and it doesn't seems too hard to implement.  Basically
you should put your figures in the right position in order to minimize
the number of pages that the reader has to pass between the reference to
a figure and the figure itself (taking in account double-sided
documents, etc.).  TeX didn't did it in the first place because the
memory requeriments of that were too high...

* An extrapolation of the TeX algorithm to break lines into paragraphs
to break paragraphs into pages.

I was going to add another point, but I don't think it serves any
purpose... the current list is enough to keep anybody busy several
months (at least).

Cheers,

-- 
Joaquín Cuenca Abela
cuenca pacaterie u-psud fr