[Accessibility] Fwd: X11 libraries requested for future LSB spec
Bill.Haneman at Sun.COM
Wed Jun 21 02:23:59 PDT 2006
On Tue, 2006-06-20 at 19:18, Jonathan Blandford wrote:
> On Tue, 2006-06-20 at 18:44 +0100, Bill Haneman wrote:
> > Standardizing on a library like cspi wouldn't work with our current
> > generation of ATs (notably orca).
> > I still think it's very much the wrong thing to do.
> Sure. That could be the case too. In which case the conclusion might
> be that standardising it in the LSB isn't interesting yet.
I don't see a problem with the maturity of the IDL; it's remained binary
compatible for over 5 years now, and the CORBA spec defines a
validation-quality C binding to it. The "many-worlds" plan going
forward will allow us to modify the core technology, migrating it away
from CORBA or supporting other technologies as appropriate. For
anything with an IDL compiler this will be very straightforward.
You mention 'exposing CORBA details', but the IDL doesn't do this. The
CORBA C bindings do expose CORBA details but that's just a fact of life
in the CORBA C world; other language bindings are better at hiding the
implementation details. Thus by standardizing on the IDL we are only
exposing CORBA details to C language clients in a CORBA-implementation
environment, which seems not only appropriate but nearly unavoidable in
C. Presumably a C binding generated by some other IDL compiler for a
different IPC backend would similarly expose implementation details
appropriate to the object/IPC technology in use.
On the other hand, cspi is not something I relish supporting at the same
level going forward. I think python is probably going to be the AT
client development language of choice going forward, but as it uses the
C CORBA bindings at present, standardizing on them makes sense from that
perspective as well.
> Accessibility mailing list
> Accessibility at lists.freestandards.org
More information about the Accessibility