1) Tradeoffs that seem reasonable at the moment may appear less reasonable later, when the environment changes. For example, when web pages were mostly texts, it didn’t matter if there was some bloat in the web browser. But then the web pages themselves exploded in size, and now opening five new tabs of Less Wrong caused my Firefox to slow down and sometimes crash, which makes me angry at both Less Wrong and Firefox—I see no reason why inactive tabs should tax the computer so much. So today, making the browsers more efficient would help all Less Wrong readers, all Substack readers, all Electron app users, etc.
2) The usual problem that creating value is not the same as capturing value, and the companies only care about the latter. Making each part of the system more efficient makes the entire system more efficient as a whole, but people are only willing to pay for some parts of the system.
For the browser issue your anger should be directed entirely at LessWrong[1]. Browsers are one of the few types of software which are still hyper-optimized, but there’s only so much you can do when people decide to render a complex application client-side using 6 MB of JavaScript:
Inactive tabs use memory because the alternative is losing your state when you switch tabs and having visible lag and reseting the state when you switch back. It’s possible browser makers are taking the wrong tradeoff and users don’t want tabs to remain in memory, but they’re not being inefficient, they’re optimizing for UX over memory.
Similarly, Electron is highly optimized for what it is; there’s just limits to how efficient you can make a browser while still presenting a fully standards-compliant JavaScript and DOM implementation. Also Electron can’t stop people from writing inefficient JavaScript.
Assuming you think the LessWrong team should stop doing some of the other things they’re doing and optimize performance instead. I disagree but this is a value judgement.
Two possible objections:
1) Tradeoffs that seem reasonable at the moment may appear less reasonable later, when the environment changes. For example, when web pages were mostly texts, it didn’t matter if there was some bloat in the web browser. But then the web pages themselves exploded in size, and now opening five new tabs of Less Wrong caused my Firefox to slow down and sometimes crash, which makes me angry at both Less Wrong and Firefox—I see no reason why inactive tabs should tax the computer so much. So today, making the browsers more efficient would help all Less Wrong readers, all Substack readers, all Electron app users, etc.
2) The usual problem that creating value is not the same as capturing value, and the companies only care about the latter. Making each part of the system more efficient makes the entire system more efficient as a whole, but people are only willing to pay for some parts of the system.
For the browser issue your anger should be directed entirely at LessWrong[1]. Browsers are one of the few types of software which are still hyper-optimized, but there’s only so much you can do when people decide to render a complex application client-side using 6 MB of JavaScript:
Inactive tabs use memory because the alternative is losing your state when you switch tabs and having visible lag and reseting the state when you switch back. It’s possible browser makers are taking the wrong tradeoff and users don’t want tabs to remain in memory, but they’re not being inefficient, they’re optimizing for UX over memory.
Similarly, Electron is highly optimized for what it is; there’s just limits to how efficient you can make a browser while still presenting a fully standards-compliant JavaScript and DOM implementation. Also Electron can’t stop people from writing inefficient JavaScript.
Assuming you think the LessWrong team should stop doing some of the other things they’re doing and optimize performance instead. I disagree but this is a value judgement.