And then suddenly I got 30Go of cache on my hard drive because of all the sites that are so smart. And limiting the size of it will just render it useless as everybody will just tried to add to it, erasing the existing cache.
But even without that, then chances that you have exactly the same build than somebody else in cache is very weak unless it's very popular. And to get it popular means a lot of people has to download it, taking the several Mo hit download on first load. And most users will just quit the page before that, thinking it doesn't work.
The solution would be for browsers to stop the madness and decide by community to adopt a new standard with a decent language for the web. Be it Ruby, Lua, Python, at this point I don't care. I'm partial to Python, but I won't fight for it if it means I can get anything with a real stdlib, good builtins, namespaces and a readable syntax.
The same thing happens now with regular browser caching, it's not that big of a deal. People will design run-times targeting webassembly to be lightweight, but if they end up being too heavy, developers will use CDNs the same way they do now with other large front-end dependencies.
> The solution would be for browsers to stop the madness and decide by community to adopt a new standard with a decent language for the web
A subjective and impossible consensus to achieve. The community will never be able to agree, that much is obvious. The answer isn't "pick a language that works today and hope for the best", that is how we got where we are today, rather, we need a solution that gives the community the power to experiment with better solutions that aren't constrained to a single language.
I agree with your conclusion but I don't think caching will save us. Currently it doesn't work. The magic of cdn didn't happen. I still download every single bloated monster on the web every time I want to read it.
But even without that, then chances that you have exactly the same build than somebody else in cache is very weak unless it's very popular. And to get it popular means a lot of people has to download it, taking the several Mo hit download on first load. And most users will just quit the page before that, thinking it doesn't work.
The solution would be for browsers to stop the madness and decide by community to adopt a new standard with a decent language for the web. Be it Ruby, Lua, Python, at this point I don't care. I'm partial to Python, but I won't fight for it if it means I can get anything with a real stdlib, good builtins, namespaces and a readable syntax.