Hacker News Reader

Comment by sonicrocketman | Hacker News Reader

For anyone who's able to, I'd love to hear how my awful method compares with other real methods. (i.e. not accounting for some mathematical nuance, performance, etc)

View on HN

In addition to the obvious issues, this method ignores python's __abs__:

https://docs.python.org/3/reference/datamodel.html#object.__...

You can make you own "numeric" types in python by implementing the proper methods in your numeric class.

I don't know why you are getting downvoted. I jumped into the comments, eager to see an explanation of exactly why this is horrible, and was disappointed.

I mean, most of the reason why this works is because Python doesn't care about type safety in the slightest. There's a reason why more strongly-typed languages would outright disallow a program like this: it's not only unsafe, it's outright incorrect. The ASCII symbol for '-' is not a suitable analog for our concept of 'negative'

You could certainly do this in languages with stronger type safety.

All it relies on is checking the type and toString/fromString conversations.

> The ASCII symbol for '-' is not a suitable analog for our concept of 'negative'

Why not?

Conventional computing has no concept of positive or negative, but rather uses a signing bit to discern their polarity (for example, an 8-bit value could represent either -127-128 or 0-255). That concept of "type correctness" is lost whenever you haphazardly cast unlike values between one another (in this case ints and strings), so we're trusting the Python interpreter to magically perform this illegal, slow operation without supervision.

I've seen similar code to calculate the last digit of a number. It's much slower than %.

Reply on news.ycombinator.com