Use integer weights #1

New issue

Open

opened 2025-12-28 18:15:58 +00:00 by sami · 2 comments

sami commented

2025-12-28 18:15:58 +00:00

Owner

Originally created by @fyrchik on GitHub (Sep 9, 2020).

float weights are cumbersome. We can use integer ones (or some kind of decimals) as they are easier to work with in multiple languages.

Originally created by @fyrchik on GitHub (Sep 9, 2020). `float` weights are cumbersome. We can use integer ones (or some kind of decimals) as they are easier to work with in multiple languages.

sami added the

discussion

labels

2025-12-28 18:15:58 +00:00

sami commented

2025-12-28 18:15:59 +00:00

Author

Owner

@alexvanin commented on GitHub (Sep 9, 2020):

I think it is possible to normalize values, considering 1.0 as math.MaxUint64 (32, 16?) and 0.0 as 0.

@alexvanin commented on GitHub (Sep 9, 2020): I think it is possible to normalize values, considering `1.0` as `math.MaxUint64` (32, 16?) and `0.0` as `0`.

sami commented

2025-12-28 18:15:59 +00:00

Author

Owner

@fyrchik commented on GitHub (Sep 11, 2020):

Actually we can use both and corresponding suffix to function names (U64, F64).

@fyrchik commented on GitHub (Sep 11, 2020): Actually we can use both and corresponding suffix to function names (`U64`, `F64`).