[anon1385]

Use of SharedArrayBuffer in timing attacks was widely predicted and there were working examples showing how it could be used to extract information from browsers long before vendors enabled it by default. It's hard to have faith that the people who ignored all those risks won't do so again.

It would be an existential crisis for several of the browser vendors if they can't add things like threads with shared memory to javascript/webassembly. The existence of those organisations is based on the idea of the browser being a replacement for native platforms, and if they can't include basic and vital features like threads then they will always be at a huge disadvantage to native platforms.

Do you think Mozilla will admit the entire browser-as-platform project is flawed and disband the organisation? Or will they push ahead with these new features despite the risks?

[geofft]

Mozilla is the same organization behind Rust. "Unsteuctured shared-memory concurrency is inherently unsafe, and also not necessary for performance; here's how some structure permits a compiler to make it safe" would be a huge ideological win for certain parts of the company.

(Also, meanwhile, none of the native platforms that Mozilla is trying to replace have disabled threads.... so by that standard, Mozilla is already miles more credible here.)