3 ms·
Honestly, as an ex-Googler (10 years), I have to say it's never that simple to explain why some tech is adopted or pushed at Google and it's never so top-down a
by cmrdporcupine 5y ago
Honestly, as an ex-Googler (10 years), I have to say it's never that simple to explain why some tech is adopted or pushed at Google and it's never so top-down and unicausal.
Often it comes down to certain engineers having outsized voice and influence combined with them having a preferred hammer and then hunting around the organization looking for nails. IMHO that's what happened with Dart & Flutter.
I saw it happen in two PAs, and the second time it involved literally rewriting an entire shipped product on a platform I worked on. Slowly and tortuously, rewritten from a JS/HTML application to Dart/Flutter on "native" (i.e. not web or android or ios target), and only a few months after said product launched. To me this was classic Joel Spolsky "Things You Should Never Do, Part I" aka "rewrites considered harmful" alarm bells, but who listens to me, I'm just a grouchy old engineer.
One promise was that it would be faster (because "native"), but this was frankly based on a lot of untested assumptions and biases and was never correct. V8 and the Chromium rendering engine have (low estimate) hundreds of thousands of hu-man hours of optimization put into them. It's actually really hard to beat. Chromium has a pretty decent accelerated rendering stack, and boatloads of work went into getting it optimized for smaller SoC type devices (work often done by friends / coworkers of mine, actually).
And switching to Flutter meant having to jerry-rig (or worse, rewrite from scratch) basic things like accessibility (screen reader, magnification, etc.) or virtual keyboard that Chrome (or at least ChromeOS) had solutions for and we had just spent over a year making work for our product. Flutter delegates down to Android or iOS's implementation of these features, and as we were neither, we had no such thing. So it had to be built.
So the net result was either rewriting things that had already shipped, or, more commonly building awkward translation layers and bridges between the two platforms so that both things could run at once. a) waste of engineering hours b) source of bugs c) was happening instead of dealing with tech debt and improving or evolving the existing codebase.
The one upside was portability between Android and our product, so the same feature/app could be written for both... which I would accept as the only compelling reason to justify what was done.
I could go on, but I'd probably give away internal secrets or something, and probably piss someone influential off or burn some bridges.
My overall point being: test your assumptions and don't use some tech just because you a) like it or b) wrote it.
Also I don't really understand why Dart even exists. In the 21st century, there are very rarely serious problems to which "we need a new language" or "we need a new operating system" are the correct answer.