Converting SL documentation to speech

Having SL documentation available in audio form would be useful to some of us whose eyes aren't enjoying the task of reading small text anymore, as well as for people who don't read too well, or for those for whom English is not a language they are fluent in reading, or for those who find that hearing things helps them remember things better than reading them.

Second Life documentation, such as the LSL Guide, has a number of characteristics that challenge text to speech programs. The way that function names are done is one example: "llSetText" doesn't look like English to most text to speech programs.

What would the best approach for making a large number of conversions such as converting "llSetText" to "ll Set Text" and a number of other sorts of conversions to yield a correctly and sensibly pronoucable file for a text to speech program?


---------------------------------------------------------------------------------------------------
Note: this is not a scripting question, this is not something one would want to do in LSL .

My Mac read "llSetText" just fine. But I have a text editor (BBEdit) that can find-and-replace across files. It uses GREP pattern matching which I'm not familiar with, but I'm pretty sure can do things like recognize upper-case letters and replace with a space and the letter.

Opera, with it's shiny new voice capabilities, read it as "ull-SetText".

I'll read any parts over skype that ya need

I'd really like to be able to have all the documentation available for in spoken form, and the documentation changes.

Blue, you can start by including attachments of the LSL wiki, LSL guide, and Support wiki in mp3 format for anyone who might find it useful.

Opera tries its best but I read such things as llTan as "el el Tan" not as "ulTan", and Opera's voice system reads llSetPrimitiveParams as if it were one word, it's not quite right.

Opera's text to voice might actually work adequately for normal english, although I think it, like some other tts systems I have tried, chokes when you give it extremely large amounts to read at once, like the entire functions definitions sections It goes along fine for awhile till at some point Opera crashes.

Opera's tts if a mighty good deal considering it's free with the free browser.

I picked the LSL guide (not the wiki) as an example because it because it has large sections which are written in a linear manner. Something like the LSL wiki, however, written in hypertext form, would need to be extracted from it's home and worked on a lot by hand.

Even something as relatively linear in form as the LSL guide function definition section needs work beyond just pronouncing the run-together scripting function names correctly.

Each function name in the LSL guide reads like "A.1 llGetPos() vector llGetPos() returns ..."
the "A.1" is a section number for the outline stucture, and is not something one would of necessity want to reproduce.

For teaching LSL one might find that having the function definitions available in audio form separately from each other would have didactic value. An additional input mode might mean greater content comprehension and retention.

What if I read it all like Louie Armstrong?

Or like Bob Dylan?