Plan for improvments in FMT:

1.	make macro case distinguishable 
	done.
2.	alternate syntax for command operand invocation
	!/command-operands/
	does not require "super fmt" mode.
	done.
3.	modify various parts of FMT to use 2 bytes for each output
	character.
	plan to extend the under line command /u/ to have a following
	digit which indicates the type of underline.
	e.g. /u1/ word /-u/ will underline it.
	byte 0	the 8 bit charcter code (ascii)
	byte 1	4 font bits (selecting 1 of 16 current fonts)
	within each font, there are 7 types of emphasis:
	0	normal
	1	underlined /u1/
	2	bold	/u2/
	4	italic	/u3/
	+ combinations; some devices will not support all of these.
	some may be mapped into each other.
	e.g. printer has 0 + 1 only.
	laser printer has 0, 1, 2, 4, 3, and 5
	bit 0 indicates "special", with the following bits having special
	meaning. e.g. "special" + space ==> space of required length.
4.	specify various quantities in more flexible units (inches, ems,
	dots, etc.

5.	dynamically load the font tables for each device.
	device 9700
	device	terminal
	device	p300
	device multiwriter
	device	pn
	device	tn
	etc. etc.
	each device has atributes specified in the device table files.
	some of these will replace the current control phrases:
		[no] tn
		[no] pn
		[no] terminal

	There will be an optional output filter for each device specified.
	e.g. popen("/usr/lib/fmt/p300.filter","w")

6.	font 1 = 'roman' after the device is specified, specify the
	device depent font type.
	p300.roman would contain the basic information about the 
	device (e.g. the unit size) e.g.
	device	unit	units/inch
	------	----
	p300	char	10
	9700	dot	300
	e.g. device 9700; font 1 = 'roman'
	would use the file 9700.roman.n
	where "n" is the name of the emphasis 
	a missing emphasis file defaults to "0".
	each character has a value, which is the size of the character
	in the basic units of the device.
	a dash indicates no character is defined at that position
	P300:
	# 0	1	2	3	4	5	6	7
	-	-	-	-	-	-	-	-
	-	-	-	-	-	-	-	-
	-	-	-	-	-	-	-	-
	-	-	-	-	-	-	-	-
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	1	1	1	1	1	1	1	1
	alternate:
	have each character named, and follow it with the size it
	takes, it's octal code or hex or character name
	e.g.
	a 'a' 1
	b 'b' 1
	c 'c' 1
	d 'd' 1
	alpha 0314 1
	...
Another possibility:
the device xx command acts like a:
	include /usr/lib/fmt/device
this would also allow one to do: --9700 as well as a: -device=9700
then all one would need would be new control phrases that define the
characteristics of the new device:
e.g. units = 300 dots = 1 inch = 10 spaces

a new control phrase is used to introduce a new font: font n = 'name'

device multiwriter
format 'titan12/ps'

Plan of action
------------------
change fmt.c and related pieces of code to output 2 bytes at a time:
curfont, followed by the data byte
then change word_end, to calculate the length of the word in appropriate
way:	unit	if using non proportional spacing
	table[curfont&FONTMASK]]	otherwise
Plus consideration for overprints etc.
also need to change justify & centre to calculate in the appropriate way.
