Computer Corner by Bill Kibler

Forth Day 1995

Well Forth Day came and went again, pretty much un-announced and under
attended. Each year on a Saturday in November, FIG sponsors a full day
event discussing Forth. This year's event happened on November 18 at Dr.
Ting's place of work in San Mateo California. I drove down with three
others from the Sacramento Forth group.

Our trip down was filled with discussions ranging from politics to Forth
and embedded projects. The day started with donuts and social
discussions. About 10 AM the official meeting started and short
introductions from John Hall, president of FIG, and several chapter
leaders, brought attendees up to date on Forth happenings. One visitor
from Taiwan said their FIG chapter has over 300 members and is
developing a commercial Forth (he can be reached at
cchin@simon.pu.edu.tw).

After introductions, the first of several speakers started talking about
their projects. Jeff Fox gave an update on F21, the mpu21 version for
handling Fiber communications. There are a number of bugs and errors in
this prototype run, but Chuck Moore says he has been able to work around
them for testing purposes. Speeds are looking good and sometime early in
'96 a chip with less bugs will be available. A European company is
apparently funding and pushing the project on a fast track of
development.

John Hall then talked about a RF signal tracking system using Forth
Inc's ChipForth on a Innovative Integration SBC. They use VME style
boards with TMS320C31 as the controller and signal processor. This is
for UHF signal tracking with high speed switching between horn antennas
mounted on a dome.

Dwight Elvey talked about DSP filtering in Forth. He indicated you can
do 36Khz filtering on 486/33Mhz system, all using FPC and integer math.
He provided sample code and extensive explanation of how it works.

After our lunch break, Alred Tang presented ideas on doing Exact
Rational Math using forth. Leonard Morgenstern presented his latest
Red-Tress: a way to achieve balanced binary trees. Al Mitchel from AMR
(see back page) talked about using and selecting microprocessors for
embedded control. Al compared price to performance and pointed out some
of the good and bad features of those he uses and sells.

Mosaic's Patrick Campbell talked about the QED industrial controller
which has been getting some good exposure lately. Bob Nash related his
experience with using Express and Forth Inc's polyFORTH at SMUD
(Sacramento Municipal Utility District). Bob has done things others gave
up on and in times most thought impossible, all because of Forth's tools
and his learned philosophy. Those concepts he uses are: keeping it
small; modular projects; bottom up design; having a vision about the
project; and making sure your design and concept is sound. In two years
they have put on 6 systems using Express (the industrial control version
of polyFORTH, a high level PLC type program) and leveraging their
knowledge from the first successful project onto the next one.

John Baumgarner talked about a Forth project he is doing to augment
pilots ability to navigate. The idea is to have visual images projected
into the pilots view from navigation devices, a form of virtual reality
for pilots. Dave Jaffe, our last year's Programmer of the year
recipient, gave an update on the Finger Spelling hand. Three companies
are investigating development: one in West Virginia to do the hand
itself; a college in Israel on using the hand as proteus or hand
replacement; as a device to do hand therapy - to move a hand and thus
exercise the muscles; possible use in remote or dangerous situations
needing tactile operations.

Chuck Moore gave his normal Fireside chat in which he updated and
explained more about his progress with the F21. He covered his bad
errors and how testing hadn't uncovered them. He discussed some design
considerations and his options in dealing with them. Chuck also
commented on a recent product release shboom, which may be an illegal
copy of his previous work. As usual it was a very enlightening last talk
of the day and got us already for the evening dinner.

As one of the day's events, the dinner speaker, Skip Carter, was able to
draw over 25 members to a local Chinese restaurant to hear him speak.
Skip talked about where he considers Forth is going and some
considerations of changes needed. From there a round table discussion
continued and eventually spilled out onto the street as we out ran our
welcome in the restaurant.

Skip's feelings are based on his personal experience and his two ways of
using Forth. He has used and is continuing to use it in embedded
oceanographic projects. He came with a working six legged bug like robot
that spawned many jokes. Skip also found he still uses Forth to perform
many scientific data projects and as tools to solve special problems. He
see's forth making it's way into the new operating systems and is
currently looking for Forth for his Linux system.

Skip Carter's WEB site currently holds most of the FIG's library of
material, and he is also the librarian for the FIG Forth Scientific
Library project. This is a project to collect library routines that are
used in the Forth community and are all ANSI Forth compatible. It is the
first of many ANSI standard projects.

I asked the question as to what is going on with ANSI and if any
commercial vendors are selling one. Skip said he had talked with three
vendors at Rochester and two of the three were interested, but later
decided not to change from their current direction. That means there is
a Forth ANSI standard, but no commercial versions available for use. A
public domain version was suggested but nothing more than that.

Lunch Program Contest

I skipped over lunch in order to explain it better here. For many years
FIG meetings have had lunch time contests or seminars. This year a
programming contest was presented and three groups tried their hands.
Bob Nash, Charlie Shattuck and myself comprised one team attempting to
program traffic signals. I have included the circuit since I think it is
an easy way of learning how to do real life problems using simple tools
and devices.

As you can see the diagram is rather simple and shows just how a
parallel port can do real I/O. The contest handout says this comes from
Dr. Ting's “The Second Course” which is the second book of learning
Forth (can be bought from FIG). The parallel or printer port has 12
output bits and 5 input bits.

This simple design limits the crossing signal operations to having both
sets of light work the same. By that I mean the north south set of
intersection lights will operate the same and force the left turn
signals to cycle for both directions before turning on the green.

The contest went something like this, turn on all stop lights at power
up. With no traffic turn on E-W green and N-S Stop. Then the fun starts,
when traffic approaches you must use various amounts of delay with
caution, stop, left turn, then go and so on. Basically you make it work
as traffic signals normally do, all programmed in less than an hour.

Well our Sacramento group did fairly well against the other two groups.
No group got it fully working, but we all got some amount of operation
to happen. We had a couple of on-lookers who where very much interested
in seeing how we went through the design and programming stages. We
basically did a bottom up approach. Create a basic set of words that
control the numerous possible light combinations. If you analyze the
possible conditions it quickly becomes apparent that you have only 8
output commands to the port.

Once the basic commands are set, the next step is tying them together
with timing loops and then integrating it all into the possible traffic
switch conditions. This last one is where we had the most problems and
ran out of time doing. Overall I felt it is an excellent learning
project and recommend you try this on your own. We have done similar
things at local meetings and find it a great way to get others into
Forth and just doing things for fun.

Overall I felt the Forth day this year a bit weak, definitely under
advertised, and if nothing else shows a small lack of direction in the
FIG organization. A number of new board members were to meet after
Thanksgiving and discuss making changes. I have placed a request to hear
what went on, but as yet haven't heard a word.

So till next time, keep hacking. Bill.