From: Paul Rohr (paul@abisource.com)
Date: Sat Apr 27 2002 - 04:26:16 EDT
At 11:07 PM 4/25/02 -0400, Havoc Pennington wrote:
>GNOME != Pango.
Sure. Yet any of the global-scale companies backing GNOME must surely have
a vested interest in the success of Pango.
http://foundation.gnome.org/
http://foundation.gnome.org/advisory_board.html
But hey, that's your crowbar. I'll let you wield it wherever and however
you'd like.
>I'm not sure Pango has many resources at all. ;-)
In skimming the Pango archives, I saw talented employees of little companies
like IBM and RHAT (among others) discussing the technical work they were
doing. By comparison, I'm only aware of approximately 3 person-years of
salaried effort being invested in all of AbiWord.
Who me? Jealous? Perhaps. :-)
>The thing is, this is a piece of infrastructure where you have a ton
>of special-case knowledge embedded in the specific implementation.
>That's the kind of infrastructure it bugs me to have several copies
>of. I don't care about sharing PNG readers or XML parsers, because
>they're all the same and all conform to an external spec, the
>specific implementation of those doesn't matter.
Yeah. The true value-add provided by Pango will be a robust set of shaping
modules to encapsulate all that special-case nastiness, right? If so, then
I suspect that all we're talking about now is a chicken and egg problem.
As it turns out, Tomas has already managed to use FriBidi to coerce our
cross-platform layout engine to produce editable, printable BiDi support on
his own. Now he's looking at FreeType for simplifying the rendering logic,
with more complex shaping still a future TODO.
By comparison, Pango's using the same library chain (FriBidi, *iconv,
FreeType, etc.) to much the same ends. It seems as though both efforts are
roughly comparable in their BiDi support, and Pango's definitely ahead on
recruiting folks for glyph shapers.
Indeed, if Pango's XP and printing support weren't still big question marks,
I doubt that there'd be much to discuss. You'd have all our work done for
us, and we'd just jump on the bandwagon like we did for libwv. :-)
However, I can see why Tomas might be hesitating to take on the
architectural cost of adopting and porting Pango while it's still so
X-display-centric. Gaining even incomplete shapers is a win, but losing the
ability to print is fatal for us.
>I'm a big fan of having as many users as possible for infrastructure
>bits, including users on other platforms. More users means fewer bugs,
>more developers, all the good stuff.
Amen, brother.
Paul
This archive was generated by hypermail 2.1.4 : Sat Apr 27 2002 - 04:27:31 EDT