1998-08-04 - Whom they will destroy they first make mad.

Header Data

From: bill payne <billp@nmol.com>
To: kalliste@aci.net
Message Hash: e9f63cd8b2675c9aeb07f6ffe52dedc47086b478b86e76ce32b8378f544a1b06
Message ID: <35C7145D.5D03@nmol.com>
Reply To: N/A
UTC Datetime: 1998-08-04 14:06:14 UTC
Raw Date: Tue, 4 Aug 1998 07:06:14 -0700 (PDT)

Raw message

From: bill payne <billp@nmol.com>
Date: Tue, 4 Aug 1998 07:06:14 -0700 (PDT)
To: kalliste@aci.net
Subject: Whom they will destroy they first make mad.
Message-ID: <35C7145D.5D03@nmol.com>
MIME-Version: 1.0
Content-Type: text/plain

Tuesday 8/4/98 7:45 AM

J Orlin Grabbe

I liked your article. http://www.aci.net/kalliste/apocalyp.htm

Reminds me of one of my favorite books.  The Great Crash by John Kenneth
Galbraith.

Since you are a Harvard grad, mathematician Garrett Birkhoff bought
me lunch at the Harvard faculty club in about 1976.

We were making plans to get a machine combinatoric and stochastic
algorithm subprogram library written.  Like LINPACK, etc.

But NSA killed that project and my NSF grant DCR75-08822.  See Bruce
Barns, the Puzzle Palace by James Bamford.

We returned from Durango on Monday afternoon.

I have some INSIDE news from the wife of a former AF base commander in
Turkey about

	1  Ron Brown's flight
	2  Khobar towers

1  The AF warned the White House not to fly Brown's flight because of
bad weather. Advice issued from Ramstein. Advice ignored.

White House removed Ramstein AF base commander.

2  AF base commander was removed for the Khobar incident although base
commander had nothing to do with lack of security.

Former Kirland AFB and new base commander are buddies with our friends
in Durango.

I think that we have made the government MAD.
http://jya.com/whpfiles.htm

Wayland told me yesterday that the Texas A&M profs went over the reasons
for
rejection of their SBIR grant.

One of the MAIN reasons, Wayland told me, was the grant did not
spell-out the
WAY the profs were going to IMPLEMENT their idea.

My attached method for ACCOMPLISHING the goal was included with the
prof's
REQUESTED revision of their SBIR grant, Wayland told me.

I am getting enmeshed in techie stuff.  

I need and WANT to finish the digital FX PRONTO.

Let's all hope for settlement of the UNFORTUNATE mess before it gets
WORSE.

bill


Tuesday 8/4/98 7:06 AM

tech@cyberpowersystems.com
www.cyberpowersystems.com
512/238-9905

Yesterday I was trying to send a e-mail QUICK.  

I did not close the PowerPanel Window before invoking IE and the
external
ZOOM Flex modem software.

Dial-Up Networking window in Windows 95B reads

	!  The modem is being used by another Dial-Up Networking connection or
	    another program.  Disconnect the other connection or close the
program, and
	    then try again

					OK

I disabled the decode on the Superhigh Speed Serial Card by NOT
JUMPERing any addresses.

Guys, it appears to me AT THE MOMENT that the problem is that the
CyberPower code is interfering with the Dial-Up Networking software
within Windows 95.

It would be nice to be able to use BOTH the CyberPower auto shutdown
feature and Internet at the same time.

I realize that the software problem is likely in Windows 95 but I am NOT
OPTIMISTIC that Microsoft
will fix the problem.  Therefore.

Got any fixes?  

I will phone tech support later today.  I am not reading e-mail.

Thanks in advance.

bill


         	WEED KILLER computer interface proposal section 7/27/98 10:11 AM

Solution to controlling and collecting data from the WEED KILLER involves interfacing a personal computer
running a version of the Windows operating system to the WEED KILLER analog/digital hardware.

Windows is not a real-time operating system, therefore  microcontroller controller/collector hardware 
interface must be installed between a Wintel PC and the WEED KILLER hardware.

Essence of the Wintel data collector problem is that Windows 3.x or 9x responds to a hardware interrupt
usually between 70 to 150 microseconds.  In rare occasions the interrupt latency may extend to 1.5 
milliseconds or even longer.  A microcontroller responds to an interrupt in several microseconds.

Wintel hardware controller interface is even more difficult than collection for the reason that the Windows 
operating system only gives control to an application when Windows decides.  In the collection, mode at
least a hardware interrupt signals Windows that the application wants control.  However, the microcontroller
can send the Wintel an interrupt asking the applications code whether there is any message it needs to
send the microcontroller.

Microcontrollers have specialized timers, serial expansion ports and are, therefore, designed to be interfaced to 
analog and digital hardware. 

An 80C32 family microcontroller is proposed for the WEED KILLER application for reasons.

1  The 80C32 will do the job.

2  Multiple vendors of 80C32 guarantee future supply at a competitive price.  Current suppliers include
 Intel, AMD, Winbond, Dallas, Philips, Siemens, OKI,  ATMEL, ... 

3  High-speed parallel port bi-directional IEEE 1284 enhanced capability port 9 (ecp) communications 
hardware between an 80C32 and PC is in the final stage of development.

4  IEEE 1284 hardware drivers are supplied with Windows NT.  Custom assembler dll drivers are available
for 9x and 3.x.

5  A public-domain Forth 8051 operating system hosting a high-level language and interactive assembler with 
complete source code documentation is available on Internet. http://jya.com/f86/whpf86.htm

Hardcover book further documenting 5 is available from Academic Press.
http://www.apcatalog.com/cgi-bin/AP?ISBN=0125475705&LOCATION=US&FORM=FORM2

Only a Wintel machine is required for both hardware and software for the WEED KILLER project.  

Usually a Forth hardware/software development probject on requres a voltmeter, logic probe, and, 
infrequently, an oscilloscope.  Reason  is the INTERACTIVE control of the hardware and software
from a PC keyboard and diagnostic information easily printed to a PC monitor.

Justification for assertion made in the above paragraph comes from Internet.

NASA uses Forth extensively for its space programs. http://groucho.gsfc.nasa.gov/forth/

Ballard used polyForth http://av.yahoo.com/bin/query?p=polyforth&z=2&hc=0&hs=0
to locate wrecks of the Titanic, Bismarck, and Yorktown.
http://www.nationalgeographic.com/features/98/midway

Sun Microsystems workstation boot into Forth then invokes Solaris. http://playground.sun.com/pub/1275/

Adobe Postscript is a version of Forth. 
http://www.mactech.com/articles/mactech/language/forth-postscript.html

Video game software are written mostly in Forth.

The Wintel side of the WEED KILLER project will be most-likely written in a small
amount of assembler interface code and Visual Basic.

While Forth threaded code software technology is extremely valuable in some settings,
it is not in others.

Java is a variation of Forth.  http://www.mrl.nyu.edu/meyer/jvmref/

Future of Java on Wintel machines is unclear at this time.

For example,

  The hottest items among techies is a browser called Opera.   
  This is a $40 shareware program that in speed and compact   
  size buts both IE and Communicator to shame.  It has a slightly   
  different interface from either of the majors - an interface some find    
  refreshing while other find less than useful.  As it's shareware, you can   
  try and then buy if you like it.  One reason for its speed is that it ignores   
  Java - the Internet's Bandwidth Pig (IBP).   
 
  The Rumor Mill by Paul Cassel
  ComputerScene Magazine  July 1998   
   
Forth executes code High-level at about 10% the speed of a compiled high-level
language.

Speed of execution of small applications is not effected by Forth's slow execution.

Reason is that initial code is written in high-level Forth.  

Inner loops are then translated into Forth assembler.

Speed is maintained with the advantage that data structures are created an maintained in
high-level language while the interactive operating system is retained for trouble shooting
both hardware and software problem.

Hardware cost of building the 80C32 the WEED KILLER boards is estimated at $10k.

Hardware design is estimated at 1 month labor at $50/hr for a total of $8k.

Software development on the 80C32 side in Forth and Forth assembler, software on the
Wintel side in Visual Basic and assembler, documentation, and training is estimated to
be 4 months for a total of $32k
   
 /\/\/\
end




Thread