0.2	Initial Release		1/13/97

	Includes spo256-al2 server.  Maintained by David Sugar 
	<dyfet@tycho.com>.

0.2pl1	Patch Update 		1/16/97

	Updated build/dist and added config.dist.  This should make for
	better control of distribution sets off a live production system.
	This also solves 'second level' production problems off a 
	distribution set.

	Updated build/instinit to allow automatic linkage of the speak.init
	service script to a free 'S' slot in rc3.d.  Instinit is now smarter
	about how it handles startup script installation, especially for
	Linux.  Care with speak.init should still be taken for non-Linux 
	systems.

	Replaced missing config.c in sdk/std.

	Fixed vstat bug where it would fall-through into vmon code.

	Updated vmon to use more portable utent() routines instead of
	directly accessing the utmp file.  This was to improve portability.

	Updated vmon to use utmp routines to determine uptime if
	/proc/uptime does not exist.  This should help with porting
	to non-Linux systems, or with Linux systems when 'proc' is 
	not installed.

	Made statfs() dependent code conditional, and active only
	for linux, to prevent problems when compiling on non-Linux
	operating systems.  Code will need to be added to drive
	external "df" image if no statfs() service is supported.

0.2pl2	Patch Update		1/27/97

	Found problem in tcpsocket.c, which would not compile correctly
	on some sites.  

	Added diag build script "Make diag" target to re-mail system
	diagnostic reports.

	Added "config.diag" for "Make diag".

0.3	Second Release		2/14/97

	Changed sdk 'dev' stty support to improve stty setup operations.
	This involved minor code changes to getspo.c.

	Changed sdk 'net' binding code to standardize handling of network
	addresses passed as strings using getnetaddr().

	The [interface] bind = option may now be used to restrict the
	speak port to a specific subnet interface address.  This
	should keep the rest of the internet out of speak ports!

	Changed spo256 server to use a "character-state" parser in place
	of the prior simplified token scanning routines.  This is closer
	to the original WorldVU server, and is implimented much more
	cleanly.

	The server now also recognizes special "server" specific commands
	that may be embedded in the output stream.  These server commands
	are prefixed with an <ESC> and appear inside <>'s.  Server
	commands allow some direct interaction with the server, such as
	being able to force specific pronouniation rules.  More advanced
	devices (beyond the SPO) may someday make use of server commands
	to force change of tone, inflection, or volume.

	The [vmon] settings in speak.conf have changes to support control
	of timer intervals for monitoring system status.  The new interval
	option is meant for when we add mailbox and user login tracking.
	This controls the interval for such short duration events, in
	minutes.  The 'frequency' option indicates how many times per
	hour you want vmon to announce.  Once is usually quite enough.
	If you are upgrading, you will want to check the speak.conf
	file in the distribution.

	Vmon can now watch for and notify when new mail arrives in
	specified user accounts.  This behavior is controlled with
	the new [vmon] entries in speak.conf for 'maildir=' and
	'mailbox='.  Maildir should point the mail spool directory
	(typically /var/spool/mail) and one or more mailbox lines
	can be used to list those user accounts you want to have
	verbal mail notification for.

1.0

	The entire build and sdk process has been redone and unified for
	improved portability.  No functional changes have been made in the
	application.  The major difference, initially, is that, instead of
	performing a "make config", one now executes ./config to start the
	build process going for the package.

	We have also standardized 'config' around a hybrid file structure
	that is fully consistent with the Linux file naming standard, can
	comply with BSD file layouts, is compatible with most UNIX's,
	and automatically recognizes secure TSW "Linux" release 1 and gnu/
	glib 2.x introduced file naming and ownership conventions.  
	Support for TSW-LR1 html documentation and software package
	redistribution conventions are implied in the new 'config'.  This
	release will be available in source and binary rpm format for any
	rpm managed Linux system.