after more testing, the text that will be "hals response"(text in the response window) is around 64k limit. the actual spoken response is around what you said.
it might be that the text response is sent to the speech engine, and it can only parse so much at once.
the answer would be to parse one line, then wait, then parse another line.
the problem is i don't know how to tell it (the plugin) how long to wait. the "wait" is based on time (for 1 sentence...ect.) , but that can vary greatly.
once the text is sent to the speech engine, Hal doesn't have any way (i think) of knowing when the speech engine is finished saying it.
problems problems problems....
believe me, i am working on the issue. i have looked at many possibilities.
i know there is an answer to everything, but i don't have this one YET.
-the dude