3 ms·
In many circumstances, it’s better to optimize for developer time over compute time. I’m glad that I can choose from zillions of (slower than necessary) web/ele
by remolacha 6y ago
In many circumstances, it’s better to optimize for developer time over compute time. I’m glad that I can choose from zillions of (slower than necessary) web/electron apps instead of a smaller number of “optimized” apps written by “real engineers”.
- tgv 6y agoAs hinted at in the article, that becomes technical debt that can be really difficult to get rid of later. Of course, optimizing may create another kind of debt, but that Firefox has lost the race from a very convenient lead is telling. Ideally (IMO), if you want to optimize development time, also consider future development time.
- arvinsim 6y agoTechnical debt means guaranteed work in the future for software developers. I only say this half-jokingly.
- astrange 6y agoThe problem is that it gets exponentially harder to fix something the later you do it. Even if you wait a few years it could be easier to rewrite from scratch than fix a large bug. Btw, people bring up a features vs. speed tradeoff here and I don't think it's true - in my experience perf issues come from stupid random bugs more often than they're fundamental parts of new features. There is definitely a development time vs speed tradeoff, though; the best way to save time here is to have really good regression tests.
- tonyedgecombe 6y agoI don't know if we have optimised developer time either. I'm certainly slower when doing anything on the web than I was with VB6.
- jbverschoor 6y agoPeople like to think they're smart. Being able to use complex things makes you look smart. Creating a complex system makes you look even smarter.