Thanks to everyone. I'll get the latest grml this weekend and have a go at that, first. There are a lot of systems around that would probably talk and give good text console service that won't quite run orca and all the resources it needs either because of the speed of the CPU or amount of available ram. As one who is still using what amounts to a talking serial terminal to connect to a TTY on a Unix system for many applications, I am excited about the talking distributions out there but the resource requirements to get orca and speech running are pretty rigorous. I never thought I see or, in this case, not hear the day when 256 megs and a 1 GHZ processor wouldn't even boot the live CD, but all I have to do is try the ubuntu live CD on such a system and it starts, you hear the drums so it almost boots and then about ten minutes later, you realize that there is trouble in paradise. My wife reports the screen is mostly blank with random bits of light and color but that's it. It's too bad you can't do a talking ubunto install without orca. I think software synthesis is the only way to go but it must be able to run as if it was a hardware synthesizer which means no spelling because of interrupt service routines, etc. Part of my talking terminal is a screen reader I wrote in 8086 assembler and I can absolutely swear as to how vexing interrupt service routines can get because unless you have parallel hardware, nothing ever really happens simultaneously except for outside interrupts and then you must prioritize which one you don't do now. The only way software speech can work right is for the CPU that is doing the sound to not be the CPU that is handling everything else. That is essentially hardware speech but using the built-in sound hardware. Martin McCormick
