I don't ever want Speakup to be user space for one reason. Having it in the kernel is the only logical way to get speech from bootup to shutdown, which makes it possible to install a Linux distribution from CD with speech. I hope the future support of software speech synthesizers will not break this feature.
