Re: [Devel] OpenType futures

[George Williams]

On Wed, 2004-08-11 at 02:48, Werner LEMBERG wrote:
> Anyway, I don't object to moving the validation code completely to the
> otlayout library if this has advantages, but somehow I have the
> feeling that validation is really such a low-level operation that
> separation from the otlayout library makes sense.
My basic feeling is that the OT tables can be extremely complicated, and
that validating them (checking that all offsets are ok? how many layers
of subtables get checked? checking that glyph indeces are valid?)
essentially involves parsing them in their entirety.

I feel rather as Antoine does that a separate validation pass will be
time consuming. And that validation and parsing should be done
simultaneously while interpretation can be moved to a different layer...

If such a layer is provided for OT, would it make sense to do something
similar for GX?