home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Kosovo Orphans' Appeal Charity CD
/
KosovoOrphansAppeal.iso
/
applications
/
_linreg
/
docs
/
history
< prev
next >
Wrap
Text File
|
1997-11-25
|
4KB
|
83 lines
Linear regression utility
=========================
By Paul F. Johnson, Sept 1997
=============================
History file
============
As normal, I've included a history file, just for completeness really, but
also so that you can appreciate how labourious this sort of thing can be!.
Without any sort of break from tradition, the list goes from the most
recent to the least recent!
Versions
========
Vsn 1.33 - Cosmetic alterations to a few bits on the windows. Inserted
the icon bar menu onto the types of regression and data entry
windows.
Vsn 1.32 - Found a fun couple of bugs which gave an r^2 value which was
negative and if the gradient was over 9999, then the program
would substitute in !!!!.!!!!!!. This has been sorted out.
Vsn 1.31 - Sorted out the standard deviation problem. I should give tonnes
of thanks to Judah Arotsky at Uni of Salford for helping me out
with this one.
Vsn 1.30 - Finally managed to get the r2 value fixed. This and the standard
deviations have caused me more problems than anything else in
the entire program!. I have fiddled with this more times than I
wish to remember!. Phew!.
Vsn 1.28 - Inserted the help/history icons onto the info window. Wrote the
help/history files. Managed to get the standard deviation to
do what it is supposed to. Tidied a bit of code up.
Vsn 1.25 - Adjusted the data entry window so that the cursor and return keys
can be used. The only problem was that I followed the WimpBasic
manual which is incorrect!. After three attempts, it worked.
Vsn 1.24 - Fixed a small bug which meant that a "standard regression" gave
the same answer as a "standard deviation".
Vsn 1.20 - Sorted out a number of small bugs which would cause the program
to crash. Played with the data entry window to give 15 instead of
10 entry points.
Vsn 1.17 - Deleted from the answer window the value boxes inserted in Vsn 1.10
as they weren't needed any more.
Vsn 1.15 - Sorted out the intercept bugs (x and y gave very strange results
on a 100% straight line!)
Vsn 1.13 - All code now entered was mysteriously wiped from the code. This has
never happened before, and never since. Wiped the file and started
again. Luckily, I had 2 copies of virtually the same file.
Vsn 1.10 - Added value boxes to the answer window as the answers thrown back
were rubbish. Added code. Deleted the menu button from below.
Vsn 1.06 - (This encompasses all version from V1.03 to V1.09 - I never got
around to renumbering them). Fixed problems with the window
handler, positions, sizes and the such. Fixed a loop hole in the
code which caused the program to hangup. Played with a the icon
and inserted a menu into the choices box. Wrote the Impression &
Text versions of the formulae from the TechWriter original. After
about an hour using Formulix and Impression, I started to wonder
why anyone ever used Impression for technical documents - it's
terrible, but that is only my opinion and probably doesn't belong
here!
Vsn 1.03 - Created menu, re-wrote the sigma and intercept code.
Vsn 1.01 - Compiled and 100% crashed.
Vsn 1.00 - On paper, it looked easy. On the computer, it wasn't.
--------------
P.F.J. 27/9/97