potiuk commented on issue #67515: URL: https://github.com/apache/airflow/issues/67515#issuecomment-4559861385
I 100% agree and support @kaxil's decision here. The initial arguments were flawed and essentially show lack of understanding how Dag File Procesing and task execution works. This whole issue seem to be hallucinated. And yes - we don't care if comments, issues or PRs are AI generated or not, when there is a real human behind it (as is in @kaxil's case). What matters is the merit. And this issue simply lack any merit - so whether AI is used to generate the comments that @kaxil wrote or not, does not matter, because the proposal is simply worth closing it. However if you think @dwreeves that your proposal has merit, there is a way you can still open similar issue - just verify your claims. Make actual benchmarking - take one of those big wins, example dags using them and benchmark time to start, memory used before and after. Share our findings. General rule of "premature optimisation" fallacy that you are falling in in this case is that you based your claims on some assumptions (which turn out to be wrong) and you have not measured how much such optimisation can bring. This is rule number 0 for any kind of optimization. Measure the possible gains that you can get by the optimisation before attempting it otherwise it's a completely wasted effort and a lot of "lost opportunities" of not doing something that can actually be beneficials. That's basic 1-0-1 of any optimisation efforts. So ... By all means. Do the homework first, verify your assumptions, measure them and show how much gain we can get. This is basic prerequisite of any kind of effort like that that anyone proposing it MUST do. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
