4 ms·

Very much agree with this. That said given the quality of what they have put out, I'd surmise the issue here is morelikely they don't feel the same pain on loa

Very much agree with this. That said given the quality of what they have put out, I'd surmise the issue here is morelikely they don't feel the same pain on load times? Bandwidth too good, rigs too powerful, dev envs overpowered, etc. to enable developer velocity can often also mask real world problems. Hell, even hot reload can mean you don't see load as often and are able to easily rationalize as "not that big a deal" - I wouldn't necessarily take it as a sign they are not deeply invested in their craft.

s0l1dsnak3123 4y ago
I very much agree with you as well actually. One of the things that I've instituted at my work is our methodology for acquiring test phones for our product (WhatsApp for hospitals, basically). We go to a supermarket and buy the most eye-catching box of the budget phones available in the phone aisle. We've squashed so many bugs and performance problems this way. It's cheaper, it instils empathy in product decisions, and our customers love the user experience as a result. Similarly, one of my first big decisions after joining was to wean ourselves from our AWS addiction using Kubernetes. This also allowed us to design a dev env that fits on a modern machine comfortably, but can scale to entire Hospital Trusts in production. Once again: focusing on portability now means we can sell our product on-prem and hybrid, because we designed our developer UX this way.
foobarian 4y ago
I don't know, this seems like it was a pure CPU problem on good hardware. But anyway a developer who has 10 hour days and backlog stretching beyond the horizon will not have idle time to go try things out. It needs to be someone's explicit job to go hunt down things like this. And given it was data-size dependent, I wouldn't be surprised this was not an issue at the start while someone who's job it was to look for this kind of issue was looking. And the issue appeared after they got retasked to something else. In fact maybe this is the only kind of issue that you end up with after all is said and done. Just like in WW2 they reinforced the parts of the fuselage on returning fighter craft without bullet holes, because if a bullet hit there, the craft would not return.

s0l1dsnak3123 4y ago

I very much agree with you as well actually. One of the things that I've instituted at my work is our methodology for acquiring test phones for our product (WhatsApp for hospitals, basically). We go to a supermarket and buy the most eye-catching box of the budget phones available in the phone aisle. We've squashed so many bugs and performance problems this way. It's cheaper, it instils empathy in product decisions, and our customers love the user experience as a result. Similarly, one of my first big decisions after joining was to wean ourselves from our AWS addiction using Kubernetes. This also allowed us to design a dev env that fits on a modern machine comfortably, but can scale to entire Hospital Trusts in production. Once again: focusing on portability now means we can sell our product on-prem and hybrid, because we designed our developer UX this way.

foobarian 4y ago

I don't know, this seems like it was a pure CPU problem on good hardware. But anyway a developer who has 10 hour days and backlog stretching beyond the horizon will not have idle time to go try things out. It needs to be someone's explicit job to go hunt down things like this. And given it was data-size dependent, I wouldn't be surprised this was not an issue at the start while someone who's job it was to look for this kind of issue was looking. And the issue appeared after they got retasked to something else. In fact maybe this is the only kind of issue that you end up with after all is said and done. Just like in WW2 they reinforced the parts of the fuselage on returning fighter craft without bullet holes, because if a bullet hit there, the craft would not return.