Random in DrawScript

In BASIC RND/10 gives me what I expect: One of the ten numbers from 0.1 to 1 inclusive with one decimal place. The same function in DrawScript often doesn’t give these exactly, for example, it has spewed out 0.899999999 instead of 0.9 or 0.600000002 instead of 0.6. How come?

Added: The formula isn’t being represented correctly in textile. To spell it out it is
RND open_bracket 10 close_bracket divided by 10.

Welcome to the world of floating point rounding errors!

1/10 can’t be expressed accurately in binary; it’s the infinitely recurring sequence 0.0001100110011001100110011…

So when you ask for RND(10)/10 you really do get numbers like 0.899999999 or 0.600000002 as a result. Depending on how many digits of precision are used when displaying the number (and on how accurate the binary to decimal conversion is!), you’ll either see the inaccuracy or it will get rounded to the expected (but technically incorrect) value like 0.9 or 0.6.

(PS – you can use <notextile> … </notextile> to (mostly?) stop textile messing with text)

Yes, thanks for the explanation, I thought it was something like that. I have got round the problem by doing everything with integer variables.
The original variables took values to only one decimal place and so I can input them into the program 10 times as large and declare them as integer varaibles. Then I don’t need to divide by 10, and the implicit units are tenths. The more tricky bit was finding a way to choose M integer values within a range 1 TO N (M<=N) (M<=N) randomly, but without repetitions.
By the way, the error checker for DrawScript programs isn’t very helpful as it gives silly line numbers often larger than the length of the program and on clicking on throwback the highlighted line is not related to where the error is.

Added: Eh! What a joke, I have used the “notextile” tags as you suggest and the inequality has come out twice immediately before me saying “but without repetitions”. :-((

Welcome to Textile. :-)