Again, it's not a black or white issue. There is no universal constant threshold where it suddenly becomes viable. If you ask my personal opinion, I would say:
- If it's not even 10% effective, don't waste any resources. Replace it with another method.
- If it's 10% - 50% effective, improve it.
- If it's >50% effective, it's probably fine, leave it.
My point it that there can be many other shades of grey here. It's not fair to lump all opponents of the current implementation into the same basket.
alt Hacker News
Again, it's not a black or white issue. There is no universal constant threshold where it suddenly becomes viable. If you ask my personal opinion, I would say:
- If it's not even 10% effective, don't waste any resources. Replace it with another method.
- If it's 10% - 50% effective, improve it.
- If it's >50% effective, it's probably fine, leave it.
My point it that there can be many other shades of grey here. It's not fair to lump all opponents of the current implementation into the same basket.