This is a typical example of floating point error.  Calculation errors in floating-point math are not "bugs" in the language but a fundamental limitation of storing infinite decimals in finite space.

The issue here is that 0.1 cannot be represented exactly in binary. Every time you press a key, a tiny rounding error is added to a. Eventually, you’ll see values like 0.3000000000000001 or 0.6999999999999999.In BASIC dialects (like MMBasic), the most reliable fix is to calculate with integers and scale the output for the display.

Some examples of programming around it often use a small value - search on epsilon (or floating point epsilon) as that's what it tends to be called.

John

Thanks for beautify, Peter,
great help for hobbyists with older brains and eyes
along with fred777 and toml,


I do think that leaving blank lines intact for readability would be better

(more blank lines, for emphasis)


thanks, Dave

@matherp

Peter!

Just a small idea. Can be good if we can use string variable (not only unquoted constant string) for object names in STAR and ASTRO commands.

Better way to build up a kind of epheneris program for example.

G@bor

@Peter,

Beautify spoils labels ... in the attached code, after edit/^A/F2 there are spaces in front of the label b_loop: inserted and the code does fail with

[11] If in_str <> "" Then GoTo b_loop
Error : Cannot find label

Interestingly, there are seemingly 2 character positions inserted but if there is only one char to remove. The code runs again if the inserted chars (or char?) are removed.

Version is:
PicoMite MMBasic RP2350B V6.03.00RC1

failing code is: