3 ms·
There have been multiple submissions here on HN about how somebody figured to replace Hadoop with a well-crafted bash script which also calls several UNIX tools
by pdimitar 7y ago
There have been multiple submissions here on HN about how somebody figured to replace Hadoop with a well-crafted bash script which also calls several UNIX tools, and achieved 235x speed improvement.
I think the posters you reply to meant exactly such cases.
For the record, I have entirely replaced `grep` with `rg` (ripgrep) in all my workflows. `rg` is newer, written in Rust, and utilises all CPU cores -- unlike `grep`. So I am not against objective and demonstrable progress.
But as another poster replied to you, many of us are against wasting huge amounts of time and effort on very marginal improvements just because the tech is new.
Tell me your tech offers 2x or more speed (or less defects)and I'll use it. But tell me "hey, let's use Kubernetes because.. well.. well... a lot of people use it", and I'll pay zero attention.