The format of this file is getting organized to help me keep track of
where we are going.  I am organizing it into three sections.
Immediate things which need to be done I.E. code re-organization.  New
features we are currently working on or at least in the near future.
Long term features which will move to the new features list as the new
features list empties.

Immediate:

Screen cursoring needs to become more reliable and consistant.

New-Features:

convert the individual synth drivers into modules (involves moving buffering
	into speakup proper).

synth_file fops need to be filled in and expanded such as some type of
ioctl function and a real synth_file_read() function.


Long-Term-Features:

keyboard macros (ala ASAP/qedit)

finding calling process, so per-program configurations can be loaded

Check park status and store it before doing a console switch. (non-trivial)

make numlock turn review pad to number pad.

user defined scroll-back ala shift-pgup.

insert as 'capslock-style'/'next-style' key -- have internalized speakup
	variable, which tracks state of the insert key; each speakup bare-key
	function checks the variable, and possibly calls the alternate
	function and returns

Expand phonetic spelling to toggle on and off and work with previous
	and next and spell_word functions.  Also shift pitch on phonetic words
	for capitalized letters. [partially complete]

Chuck wants a bleep if the shift key is hit and the capslock is on.
Here is his suggested test sequence:
1. check if the caps lock is set; if it is,
2. check if the char is alphabetic; if it is,
3. check if the shift key is pressed; if it is,
4. send a bell char to the console.

Frank would like a fast way to go to the end of the word in addition
to moving to the beginning.

I'd like a function to move the reading cursor to the last char on a
line in addition to the right edge of the screen.

Gene thinks we should have a /proc function to store and show the rom
version of any synth currently operating.

The items in each list have no connection with the order of
implementation.  I just jotted them down as I thought of them.

Bill Acker WB2FLW: When the final result is "none", then, no trace of
Speakup. and use default keymap otherwise business as usual plus use
"built-in" speakupmap.map

Look at lilo routines which init the serial device and compare to speakup drivers.

