Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The manager was admitting to doing something that he knew was inefficient. He ultimately knew that multitasking was the worst practice, but he chose to do it anyway.

There's ignorance, and then there's willfully choosing to do wrong.

Edit: This was at an investment bank, in the proprietary trading group. There are no customers or deadlines. So we (should) only work on something we believe will make money.



In such a situation I think all tasks should be globally sorted (according to business-determined weighting) by the metrics:

- how much do we expect to gain if we do it

- how much do we expect to lose if we don't

- how much (black swan worst case) might we lose if we don't

- how easy is it

and then programmers should just pick off the top of list, updating the tasks metrics if they subjectively change.

That way there are multiple tasks queued but only one in play, and context switching happens when a task changes importance.


but there are circumstances too sometimes... like meeting a deadline for a customer deliverable etc.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: