5 ms·
> read and re-read the paper until it makes sense and implement the algorithms from scratch. More often than not to realize that some nasty detail got omitted
by tetha 13y ago
> read and re-read the paper until it makes sense and implement the algorithms from scratch.
More often than not to realize that some nasty detail got omitted because it was nasty.
- dasmoth 13y agoTo me, this is one of the reasons why it's good for people to do independent re-implementations. Academia doesn't value these highly enough. Recomputation and reproduction are both valuable
- lifebeyondfife 13y agoI agree completely. From my experience, because of things like the nasty omitted details mentioned in the parent comment, techniques are rarely reimplemented unless they claim pretty substantial improvements. I think having a starting point of reproduceability would lead to more reimplementations.