15.I.82



Roger:

As you can tell from the directory included with the tape, I have put a copy
of KERNEL.MAR, SS.MAR, and FLOAT.STO on the tape. In KERNEL you will find
the following code:

	1. To permit floating point numbers under the decimal radix (only)
	2. To permit octal integers
	3. To allow local numerical qualifiers in the manner of MACRO
	   numerical literals (i.e. ^X, ^D, ^O).
	4. To signal an empty parameter stack when an "=" is executed
	   on an empty P_stack.
	5. To signal an empty vocabulary stack when an ">" is executed
	   on an empty V_stack (does not permit user to flush the STOIC<
	   system).
	6. To convert single precision floating point to ASCII strings
	   under Fortran G format.

In SS.MAR I have added two calls: $FAO and FOR$CVN_OUT_G to add a bit more
power to conversion routines.

The file FLOAT.STO defines operators necessary to the implementation of
floating point arithmetic. Nothing unusual here.

With regard to the .RNO file: some portions are as yet incomplete and
others are inadequate. You may wish to expand on some of them. Perhaps I
should remark that this document arose as part of a learning exercise on
my part with respect to STOIC. I had intended to submit it as an internal
report at GM Research Labs.; however, it may be useful to a broader segment
of the VAX community. I think that by co-authoring it I can get it out of
GM and into the open literature. I hope that I have removed all the pieces
of text that would make it sound like an internal report (in which, for
instance, I acknowledged you as author of STOIC etc.).

There are some areas with which I am still having trouble getting a clear
understanding of the action of the STOIC words. These would benefit by
some clarification on your part, particularly the use of "WORD" and some
examples of it. I note that user defined stacks lacks a balancing operator
called "BPOP" to recover pushed bytes. I have supplied such an operator
and you will find its definition in the detailed description section
of the text. Crude, but it works. I am not attached to the text, so it
can be cut, pasted, spliced, abridged, spindled, torn and mutilated if
such actions would improve it. Besides, I am sure that typos and
misconceptions abound!

Let me remind you that you agreed to send back to me a new version of
STOIC when it is ready, a copy of the new RED, and other utilities which
may be of use.



				Regards,



				Wayne Wiitanen
				Computer Science Department
				GM Research Laboratories
				Warren, MI  48090

				(313) 575-3180