HN Explorer

Ask HN: How to Pick the Right Tools for the Job?

2 points by jyu ↗ HN
There's a trope "Use the right tool for the job." At the start though, it's hard to know what the hard bits of the job will be when the job inevitably shifts. It's also hard for me to evaluate new tech compared to tradeoffs of familiar tech.

How have you successfully / unsuccessfully tackled this?

1 comment

[ 4.9 ms ] story [ 10.2 ms ] thread
topicseed
Decide on a Cloud provider. Then, what services within it? Then, what client-side languages? Then, what compute system(s). Then, what architecture to use. Then, what libraries. Then what access patterns. And it goes on, and on, and on.

For too long, I got paralyzed by having too many choices – from NLP frameworks, to databases, and even CMS. The time it took me to read about all, overthink, wait, finally decide, was so long, it just is better to try a few key choices, and not overoptimize.

Obviously, it depends on whether or not you are building a side hustle or a regulated product... But generally, read enough to be aware of key options, pick a subset that makes sense (and perhaps a wildcard), implement PoCs, check them against your key metrics (latency, readability, maintainability, code to deploy times, vendor lock-in, or whatever else matters to you today and tomorrow, not three years down the line when you may perhaps hypothetically need to scale).