I ran into a bit of a glich or maybe one would called it a “bug” the other day on my 2068, that I haven’t heard tell of before.
!——–Background——–!
I was entering a program from Jan/Feb 88 Time Design Magazine, 1987 Federal Tax Calculator, by Herb Bowers, Sr., and after entering the program, and running out the typos, I ran into a error report of Nonsense in Basic, 1350:1. That particular line converted information held in a string (D$) and converted the string’s value to a numerical variable called PLUS. On the checking of the variables up to that line all seemed correct. After running the program in my Filing Status(married, joint) several times, with the same Nonsense in Basic report, I hit the single status, and lo and behold the program ran to completion. So to work right all I need do is…..no, no debug the program.
Here’s part of the code
1320 GO SUB 1500 1510 REM-SINGLE STATUS-
1330 FOR F = 1 TO 60 STEP 12 1515 IF FS = 1 THEN LET D$ ="0180000000
1340 LET HI = VAL D$ (F TO F + 2 111680019800152244800285400530400359E9
1350 LET PLUS = VAL D$ (F + 3 TO F + 7) 1475438.5"
1360 LET PCT = VAL D$ (F + 8 TO F + 11) 1520 REM-MARRIED, JOINT, Q/W-
1370 MORE CODE FOLLOWS BUT NOT NECESSARY 1525 IF FS = 2 OR FS = 5 THEN LET D$ =FOR THIS "030000000011280003300015450040800028 " " 9000884000359E92459038.5" " " AND SIMILIAR LINES FOR MARRIED FILING " " SEP. AND HEAD OF HOUSEHOLD.
Note the data in the sample may not be correct!
*——-Searching for the problem———*
Ok, the program worked in the single status fine, but not in the other 4 filing status. I did a PRINT D$, and in each line where D$ equalled a line of data, the code was correct, letter for letter. I went to line 1340, a line earlier than 1350 and started the procedure in the immediate mode calling for the val of D$(f TO f + 3), and stepped thru the whole D$ with no problem while in the Single Filing status. But when I got to running the procedure in Married, joint status, the bug showed up, in that NO letters appeared after position 16 in D$. Again I did a PRINT D$, and all the letters and numbers in D$ appeared as typed in and as in the magazine.
FS = 2 (Filing status = 2 to get Line 1525)
“0300000000112800033000154500408000289000884000359E92459038.5
1340 LET HI = VAL D$ (F TO F + 2)In the first loop F will = 1 and the D$ will be picked from the first character to the third, and the value of 030 would be assigned to HI, and it checked out HI = 30.
I did a edit on Line 1525 which held the D$ in question, and looking for a O instead of a 0, noticed the cursor arrow hung up on the 16th character. I made a new line number and entered a new line with the same data. EDITting the new line, I put the cursor at the end of the line, and started deleting the line. When the cursor got to the 16th character, it stuck, a question mark appeared, then the erase procedure erased the question mark and continued deleteing the rest of the string. I reENTERed the data and ran the program and it WORKED. The remaining 2 lines of D$ data for different filing status acted in exactly the same way.
?———-Cause———–?
I recalled the same type of problem a couple years earlier, I had entered Wes Brzozowski’s Proportional Print Program, letter for letter and it didn’t work. I used the same method of entering lines with similiar data; I ENTER a line of programming, then with a new line, simply EDIT the line, give it a NEW Line number, change the data and go. But in lines of many deletions and additions a counter of errors must kick in a extra bit or something and throw out the CPU’s ability to read a statement. This is simply my guess as to the cause of the problem, but other known bugs in the 2068 revolve around errors, and the computer keeping track of those. Maybe deletions are also counted, or trigger a glich. Anyone else have similiar problems? Or different ones?
Running this “bug” past others at a recent SINCUS meet, Dave S. felt it had something to do with lenght of the line data,(or the number of characters for a string data structure) which may remain the same even though the line number is changed.
By the way the program is handy for checking your returns, not for preparing them, but double checking your math. If you are interested in 1988 Federal Tax stuff write Herb Bowers, SR., 2588 Woodshire Circle, Chesapeake, VA 23323. for a list of his programs. Time Designs (see page 3)Vol 5,No.1 has 1988 Fed Tax Calculator. A wee bit late for most of us, though.
UPDATED INFO: According to Clyde Tackley, Herb Bowers is not issuing the 1988 TAX program as it cannot be debugged in time for this year’s taxes.