[Devel] Re: [Fonts]Another approach to text in X

[Keith Packard]

Around 23 o'clock on Dec 24, Vadim Plessky wrote:

> You have mentioned that your solution is "Independent of output technology or 
> rasterization; it's just a configuration, customization and matching library"
> How you deal with kerning in it?

I don't.  This library gets you from a font name to a file name (+index) 
along with data needed to rasterize the font (size and various hints).  A 
simple convention (replace .PFB/.PFA with .AFM) can yield access to the 
metrics associated with the Type1 font.

I believe I can solve this configuration/customization part of the problem 
separately from the remaining rasterization and layout problems.  Given 
that others are actively working in those areas, I'd rather not attempt to 
standardize something not yet ready.  I believe we're long overdue for a 
standard mechanism for installing and using fonts; I think we can build 
concensus relatively quickly and dramatically simplify the use of fonts 
throughout the computing environment.

> |   The idea is to list all of the available faces in 'match' order -- the
> |   distance from the face to the requested pattern.  Now this list can be
> 
> How you are going to measure distance?
> It can be VERY difficult.

I have a relatively simplistic metric, the sophistication comes in 
generating useful patterns which match sensible fonts.  In particular, the 
pattern can contain multiple face names; those are matched in the order 
listed, allowing applications to suggest likely face names.  I'd like to 
avoid panose-style matching based on characteristics of the font and stick 
to something simple that applications can work with easily.

Keith Packard        XFree86 Core Team        Compaq Cambridge Research Lab