Actually, I think the web audio API was a beta feature for chrome, and everyone ended up using it, but in his final instance, it ended up breaking the former instance, the bug was actually corrected in late 2013 by scirra in C2, everything before that used a former way of handlind the chrome web audio API that worked at the time, even though it should not have been included innan engine in the first place (it had been because it was the only way at the time).
If anything, it proves that one code base for all browsers is possible as long as you are carreful with not fully implemented and validated features.
In short, rather than making things work ASAP, it is better to support what is validated and standardised, as they will work long term wise.
If it is just a dream anyway, then I will more be doubtful of the future of multiplatform rather than everything else, as that would mean 2 things:
Either standards can not be achieved, which just means that everyone should just stop, and no more professional should do anything in web design (how many people are actually saying the w3c is useless? Some are)
Or that people just do not want that, which would be surprising but that is how the market works.
Before that, what did we have as multiplatform?
a language that got compiled for every individual platform? That is silly really
or a language that is compiled and read in a virtual machine? People will abuse the problem of said machine
Here, we have specifications, and so anyone can support them to make its own interpreter, however, harmony is required, if a platform does not do its job correctly, either we should force it to do it, or drop support for it completelly.
Also, we should not have any browser specific beta features not yet fully implemented, due to that exact problem that happened, this is real life, we cannot rely on cheap tricks until it breaks, we need to make everyone agree on the standards.