Statistical analysis galore



#2

Ok, the roll never stopped. I finally caught up with my HP-41 back-log.

A few months ago I posted a rather mammoth STAT-program. It is mainly the excellent program from William Kolb, but with added basic statistics, momentums, kurtosis and skewness of data and with a proper user interface via menus.

Some, very understandably, asked for documentation of the STAT program (reading through the 1122 program steps to figure it out was hard enough for me tonight). I finally got around to documenting it just now.

The brother of the STAT program is a program called DATA. It simplifies working with data sets on the 41. With DATA you can create single (one value) data sets or double (X and Y) data sets, save it to XM, retrieve it from XM, add more data points, save it back to XM, work with fixed size data sets (e.g. or 12-months rolling data sets), remove data points, print the data sets as a graph on a printer and more.

Along with the file management system posted earlier, you will have the mother of all data set and analysis programs on you HP-41.

Here is the STAT and DATA programs .

Testing for bugs and enhancements are very welcome.


#3

I started entering the STAT program, it didn't fit on my CV :-(
I started again on a 42s but don't remember what state it was left in.


- Pauli

#4

Hi Geir,



just tried to compile DATA into my RAMBOX.

There, an error report came up with "NO Lbl 10", and indeed, if you look at your program there is no correspondend label for the GTO 10 directive in Line 158.

A bug?



#5

:) Yes. A bug.

Somehow, parts of the program caught the infamous bit-rot. It's gone, and I am well in the process of reconstructing that part of the program (the LBL e: Printing of data set).

Seeing what parts are left under LBL e leads me to conclude that I was not on the right track in the first place and so I have almost written a whole new routine for this. It is now one of the most challenging programming tasks I have had on my 41 (including the challenges I have had with MCODE).

I will release the new version (0.4) soon. Perhaps later today.


#6

Quote:
Somehow, parts of the program caught the infamous bit-rot.

What is bit rot?

#7

From Wikipedia article:

"A program may run correctly for years with no problem, then malfunction for no apparent reason. Programmers[who?] often jokingly attribute the failure to bit rot".


#8

I worked on a project once where we found the cause of many of "bit rot" errors. A more accurate description of the problem is "working by accident."

It turns out that programs often contain a bug, but because of some other property of the program or the running environment, the bug never manifests. A simple example is an uninitialized variable: the program allocates space for a variable, but doesn't initialize its value to something meaningful before using it. This is a bug. But if the location of that variable gets written by some OTHER part of the program just before the variable is allocated, then the code may work because the value turns out to be legal.

So the code works, but not because it was designed to work that way. It works by an accident of fate. The bug is never detected because the software works.

Fast forward 10 years and now there's an Operating System update, or a faster processor causes the program to execute slightly differently, or more memory does the same. The point is, *something* happens that causes the memory to be initialized differently from before. Now the program breaks, for seemingly no reason.

It's a wonder that this stuff works at all :)

Dave


#9

Extremely interesting to me as a psychiatrist. I think the brain may be a victim of these kind of errors as well. As one memory gets "reorganized", others get "overwritten".... I know that I'll be using this term, now, in reference to neurophysiology. (Or at least to my own errors!)


#10

Perhaps this is the reason a husband and wife (who have shared many of the same experiences for years) will remember certain details differently, and each be equally sure their recollection is the correct one.

#11

Quote:
It's a wonder that this stuff works at all :)

Back in the dark ages when folks were just starting to talk about computers in the home one of the superprogrammers at Honeywell said of the idea

O.K. FOR AEROSPACE BUT NOT SAFE AROUND THE HOME.


Possibly Related Threads...
Thread Author Replies Views Last Post
  Best statistical fit Richard Berler 8 398 10-30-2013, 11:25 PM
Last Post: Walter B
  Entering,Saving,and Analysis /Fitting X Y Data on the Prime Harold A Climer 6 350 10-26-2013, 01:54 PM
Last Post: Tim Wessman
  HP85 Waveform Analysis inaki 1 158 04-23-2013, 01:20 PM
Last Post: Paul Berger (Canada)
  Statistical Question Namir 3 186 12-22-2012, 03:39 AM
Last Post: Bruce Larrabee
  Hp-41 structural analysis pac porting to hp49g+ giancarlo 7 333 05-14-2012, 04:14 PM
Last Post: Luiz C. Vieira (Brazil)
  Project Euler Problem 39: Statistical Mode on the HP 48GX? Peter Murphy (Livermore) 3 160 07-29-2011, 09:44 PM
Last Post: Peter Murphy (Livermore)
  Direct current circuit analysis program ferroburak 3 189 11-10-2009, 10:46 AM
Last Post: Vieira, Luiz C. (Brazil)
  OT: Tutorials for R statistical language Namir 8 356 10-29-2009, 01:37 PM
Last Post: Tim Wessman
  HP35s 2-D frame analysis John Wasilewski 7 269 11-03-2007, 10:08 AM
Last Post: Rodger Rosenbaum
  My analysis of Joel's HP35s' bug Valentin Albillo 5 227 10-16-2007, 06:16 PM
Last Post: Valentin Albillo

Forum Jump: