↗312 and ↗812 are closest. It’s the same algorithm, same execution speed for me (after the initial delay getting my head round the idea). As far as I can introspect, anyway. But also if we wrote a simple program it would take the same number of steps in either writing direction, since the two problems are mirrors of each other.
It’s not a pure mirror with numbers of different length. For example, if the question is “what number is biggest?” then ↗numbers have a faster algorithm in left-justified text because the reader can scan the right side only.
↗101
↗749,09
↗99
↗486,98
Which is of course why in the real world columns of numbers are right-justified. Still, this would solve the concrete problem of making my spreadsheets look nicer when I mix text and numbers in the same column.
Good example. This leads me to wonder, if we were starting from scratch, whether the relations between numbers (as you’ve demonstrated here), or the positional notation, would make for a better optimization target for numeral systems.
In a number, there are generally two things we care about: its magnitude, and its value modulo something (e.g. is it even? is it round? for large modulo, we just get the exact number out). A good number system would somehow communicate both.
Here is a list of numbers. Which two of these numbers are closest together?
815
187
733
812
142
312
↗312 and ↗812 are closest. It’s the same algorithm, same execution speed for me (after the initial delay getting my head round the idea). As far as I can introspect, anyway. But also if we wrote a simple program it would take the same number of steps in either writing direction, since the two problems are mirrors of each other.
It’s not a pure mirror with numbers of different length. For example, if the question is “what number is biggest?” then ↗numbers have a faster algorithm in left-justified text because the reader can scan the right side only.
↗101
↗749,09
↗99
↗486,98
Which is of course why in the real world columns of numbers are right-justified. Still, this would solve the concrete problem of making my spreadsheets look nicer when I mix text and numbers in the same column.
I don’t understand the argument. This seems just as easy in both systems.
I find comparing the first number easier than the last number (just because that’s where I start reading).
But it seems very likely that you would find the backward method easier if you grew up using that method.
Good example. This leads me to wonder, if we were starting from scratch, whether the relations between numbers (as you’ve demonstrated here), or the positional notation, would make for a better optimization target for numeral systems.
In a number, there are generally two things we care about: its magnitude, and its value modulo something (e.g. is it even? is it round? for large modulo, we just get the exact number out). A good number system would somehow communicate both.