VassarStats/Siegel book
Mann-Whitney U
computation discrepancies

First posted
Wednesday December 9, 2009 15:27
Updated
Thursday December 17, 2009 08:57

Test Faces The Knowledge Gap.

Please note that csd attempts rudimentary configuration management with First posted and Updated.

No revision history which was required in Boeing 767 documents is posted.

When was it updated?

What was the problem?
Have you ever been to a configuration managment facility?

I have at the Boeing plant in Everett, WA in about 1979/80.

What it consists of is a large warehouse consisting of stacked metal platforms.

On each flatform are boxes of documents.

Most boxes will never be looked at.

But if a plane goes down and suspicion is that a part failed, then there is a rush to the warehouse to retrieve the box and its documentation.

Configuration management is as critical in software as in hardware.

VassarStats perhaps should address the problem of configuration management?

Test Faces The Knowledge Gap.


Friday December 11, 2009 17:31

Hello professor Lowry,

Thank you for your help.

This problem appears to be solved.

But I've not located the UA = 64 on pages 121-124 of Siegel yet - later.

I now feel much more confident in my decision to use VassarStats for Sandia National Laboratories Mixed Waste Landfill cadmium and nickel computations.

Reason is we are trying to be super-careful with this work is that it may cost Lockheed-Martin
substantial money if the landfill requires excavation and removal to solve any possible problems.


Google News/Business Friday December 11, 2009 06:53
I don't know if I am getting paid, other than two lunches, for this dangerous work or not.
The true Indian sets no price upon either his property or his labor. His generosity is limited only by his strength and ability. He regards it as an honor to be selected for difficult or dangerous service and would think it shameful to ask for any reward, saying rather: "Let the person I serve express his thanks according to his own bringing up and his sense of honor."
cheers
bill



There may be some significant money helping to prove or disprove pollution using the Mann-Whitney-Wilcoxon U, VassarStats, Excel, and CSS.
Asarco exits bankruptcy, will pay U.S. $1.79 billion
Dec. 11, 2009 12:00 AM

And as part of the reorganization plan, the U.S. government will receive $1.79 billion, the largest environmental settlement in U.S. history, according to federal officials.

The money will be used for the environmental cleanup and restoration of more than 80 sites in 19 states, including $32.7 million earmarked for about 10 sites in Arizona.

Arsenic pollution.


DO-178B, Software Considerations in Airborne Systems and Equipment Certification.

Some believe that software should be built like hardware.

Microsoft assigns product and revision symbols to some of its products.



Sandia National Laboratories built a radiation hardened version of the 8051. A license to modify Intel's 8051 layouts was apparently obtained.

As a result of our 8051 Forth work at Sandia labs, we got information about the revision history of the part. At the time Sandia got the layouts, if memory serves correct, it was on the 581st revision.
Intel 8051 family microcontrollers, we found at Sandia labs, had problems with weak output ports subject to burn out easily.

We could detect failed output ports easily and interactively with the 8051 Forth words PC@ and PC!.

PC@ and PC! [page 298] are remarkable for the reason that code is executed off the stack.



I wrote PC@ [PC fetch - port character fetch] and PC! [PC character store - port store]. Character is 8 bits.

The interactive incremental syntax checking assembler which assembled them runs on the 8051 family microcontrollers.

Payne, William (December 19, 1990) [1990] (in English) (hardcover). Embedded Controller FORTH for the 8051 Family. Boston: Academic Press. pp. 528. ISBN 978-0125475709.

There are between 50 to 60 person years of work in that book Jerry Boutelle and I estimated.

Charles H. Moore.

All 8051 parts, even from the same manufacturer, die may not be the same.

Here's an 80C32 40 pin dip.



While the top of the part may look the same, the revision code is written on the bottom side of the part.



Here's the above part revision symbols magnified 20X.



Below is advice taken from a Boeing 767 software standards document.




The Wednesday December 9, 2009 VassarStats Mann-Whitney U computation signaled no error even though the two sample sizes probably could have been computed from the number of data entered and checked to the sample size entered.

Number 2 and 3 above looks to have been violated.

This page is formatted with CSS by the author

CSS and xhtml appear to be an 'outer interpreter' which means it parses text, looks up tags, then calls a routine to implement the tag command. These outer interpreters may be written in scripting languges which usually execute slowly. But this is good for the fast microcomputer market.

If you make a mistake in either 'language', usually no error message is returned. Either nothing happens or unintended changes happen to a page.

Liberal arts approach to software? :-)

This apparent flaw in inplementing CSS and xhtml caused Go Daddy to implement syntax analyzers in its Go Dadday Quick Shopping Cart ® product.

Hello professor Lowry,

We encountered a problem with VassarStats disagreeing with Siegel computations.

Problem, on first glance, appears to be in the rankings.

I'll post both Siegel computation and VassarStats.

Any help in resolving this issue would be appreciated.

I checked Siegel's small sample example and it looks to agree with VassarStats.

Thanks in advance.

bill

First Siegel













Here is the data entered in column A and B.

Column C is later adding ranks from VassarStats.







Poster photographed in the Isaak Walton bar/high speed Internet room in the basement.

Vassar is a liberal arts college, too.