Hi Florian, I also reviewed your concepts and there's one minor point: You're doing gesture recognition _after_ transforming. This may prove to be a source of problems, but TBH, I don't see no general solution.
I think you make be better off by doing transforms, but retaining 'raw' data so you have at least the option to use raw instead to recognize gestures. E.g. clipping could get in the way. > Please allow me to steer the conversation back to the original topic. As > mentioned before, the question whether the format is XML or not is > basically just one of which iostream class you use.. Two things to consider: * good parsers _can_ parse well-formed fragments (as opposed to full docs) * xml namespaces are by far easier to extend than designing forward-backwards-compatible languages and associated parsers. Cheers, Simon _______________________________________________ xorg mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/xorg
