Re: [Devel] beating the patents - an idea

[Werner LEMBERG]

[Cc'ing to the devel list.]

> One idea I had after I wrote the last email was a "quick and dirty"
> fix: that is, to use automatically generated "mask" files for the 10
> or so most common fonts at small sizes.
>
> These are simply the differences between the font rendered with and
> without hinting, stored as a bitmap and then super-imposed over the
> characters when rendered.  Because the whole image of the font is
> not stored, this might avoid any copyright issues, and it would
> render super-fast.

This is indeed a new idea noone has ever thought of until now!  Sounds
interesting to me.  The resulting bitmap strokes would be very small
(using bdat's format 4 -- provided we get more information -- or
adding a new format suitable for such fonts).  The only thing
necessary in FreeType would be a single function, say,
FT_Render_Superimpose(font1, font2), which takes two font handles and
superimposes the bitmaps.  Hmm, glyph metric differences are needed
also since autohinting can change the advance width, but this needs
more investigation.

Font editors like pfaedit could easily produce such a `difference
font' automatically, rendering the font with one time with native TT
hinting and another time with FreeType's autohinting (possibly
hand-tuned afterwards).

What do other people think?

> Here is a much more ambitious idea if you are interested: (one that
> might generate lots of commercial interest) It would be really great
> if freetype became THE world leader in a new type of font, -one that
> had native support for all the advanced features that a person could
> ever want in a font, eg. shading, colors, 3D, rotation of
> characters, slimming/fattening, animation, stretching, wire-frame
> etc.
> 
> Something that would put it so far above ttf etc. that the other
> companies would have to play catch-up...

Inventing a new font format is nothing for the faint-hearted.
Additionally, there are already way too much font formats outside.
Note that some of the features you've mentioned will never be part of
FreeType but should be implemented one level higher, and indeed,
there are already graphic libraries which have support for them.


    Werner