Which reminds me of Brian Fairchild's signature:
#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."
Which reminds me of Brian Fairchild's signature:
#5 "If you think you need floating point to solve the problem then you don't understand the problem. If you really do need floating point then you have a problem you do not understand."
And another posted on The Embedded Muse today:
So how much was that fee one had to agree too?
I much prefer dividing by multiplying integers & tracking the imaginary decimal point.
Many years ago a guy asked for some advice...he said he had spent days trying to get some floating point to work (back then it was in asm) & was trying to scale an adc to divide by 7.38 or something. I told him to scrap it all & he should be multiplying by an integer instead...I thought he was gonna slug me. Once I explained, he went and tried it & had it running perfectly in a jiffy.
I much prefer dividing by multiplying integers & tracking the imaginary decimal point.
Absolutely.
I think I may have suggested it here just once or twice ...
EDIT
Didn't think it'd be long before another came along:
https://www.avrfreaks.net/commen...
That seems a common thread amongst us old ones... it's the realisation that there are no more 'values' available in (e.g) a 32-bit floating point number than there are in a 32-bit fixed point number. Suddenly it's obvious that if the range is so large, there must be gaps between the numbers.
I guess that modern CE teaching doesn't include theory of numbers, just how to find a suitable library and hope that it's bug-free.
Neil
In that case, you might as well make it a double!
Don't interrupt him--you don't want to be the type to cause an overflow!
(oops - wrong thread)