-*- mode:bolio; package: steve; readtable:cl; -*-

This is a list of things I will try to do before I leave. You probably
want to test these later since some of them may be done hastily. Others
may not be done at all, and they are reasonable things to work on later.
Also look at the file [nil.steve]bugs.lsp which is a text file 
which I use to keep the current list of known bugs/misfeatures/schedule
of what to do. It is not really in english, it is just shorthand to 
remind myself, but you will probably be able to decode most of it.

(1) List of files to look at for further pointers to problems and
unfinished projects.
 Bugs.lsp
 docvar.lsp ;M-X list-redefinitions, M-X List-Options [Now finished]
	These can be used as the basis for M-X Edit-Options
 syn.lsp ;M-X Syntax-Modification (edit syntax table)
	This can be used as another basis for M-X Edit-Options and
	also Dired.
 [editopt.lsp contains edit-options.]

(2) List of things to do.
 	(1) Document extentions better.
		Major modes.
			The file modes.lsp defines the major
			modes, using the functions defined in mode.lsp.
			modes.lsp is very easy to understand.
		Minor modes.
			Barely implemented. Not pretty to use yet.
			See autofill.
		Variable Binding in a buffer. (Is this documented yet?)
			See functions local-bind and bind-in-buffer 
			(In mxcom.lsp)
		Key binding in a buffer
			Not implemented yet. (!) I will work on this.
			[See (3) below.] [Now implemented]
		Recursive editing. [O.k. now. See recurse.lsp]
			Recursive editing is in bad shape. I have not
			yet done enough with it to see what functionality
			has to be implemented to make it reasonable.
			Look at syn.lsp & info.lsp for examples of
			of recursive subsystems. 
		???

	(2) Make minor-modes easier to use.

	(3) Define a function to bind keys locally to one buffer.
		While working on this I redid the entire key binding scheme
		The new one should be easy to interface to the VMS ECHOIN
		protocol. (GJC) It is also faster and handles inheritance
		of control-x sub-commands correctly. (A new feature.)
		Because it memo-izes the key bindings there is no longer
		a penalty for many levels deep of modes built on each other,
		so binding in a buffer will simply add one more level of
		indirection.
		Debugging note: Because I modified ECHOIN (now editor-echoin
		and echoin-primitive) there may be bugs in keyboard macros.
		Look at (functions): editor-echoin, string-read&save and
		ed-command-interpret if you find bugs. [Now tested.]

		The function (bind-key-in-buffer buffer key-seq binding)
		should now work. (It is not yet tested.)

	(4) Perhaps separate M-X names and functions more.

	(5) Perhaps look at the Zwei perfect backward parsing algorithm
	    again.

	(6) Try to integrate the standout feature of searches with the
	    redisplay so it is done faster. This would make searches even
	    more pleasant than they are now. This probably has high
	    priority since it is a major user frienlieness feature, and
	    I am the best person to try to hack the redisplay. Unfortunately
	    the redisplay was designed to be very fast, not to be hacked so
	    this may not be possible. I will at least comment the redisplay
	    where I think it needs to be hacked.

	(7) Think about, document and eventually implement some better scheme
	    for initializing the many variables dynamically bound in the
	    function ED (from file ed.lsp). The problem is that many of these
	    should be cusomization variables, but there are many computational
	    dependancies so just setting a variable may not work right. On
	    top of that they must be set at run time, because many of them
	    depend upon the terminal size, which is not known until then.
		Bletch.

	(8) Perhaps change the redisplay so that just-expand-line is
	    a line method. Implement suppressed text. Implement fonts.