This will fall in deaf ears, and maybe rightly so, if computing history is any guide. In 1954, the biggest issue John Backus had to overcome for FORTRAN was pe

This will fall in deaf ears, and maybe rightly so, if computing history is any guide. In 1954, the biggest issue John Backus had to overcome for FORTRAN was performance: "it is difficult to convey to a reader in the late seventies the strength of the skepticism about "automatic programming" in general and about its ability to produce efficient programs in particular, as it existed in 1954." But the world went on to use higher-level languages with worse and worse performance characteristics. We can't blame developers trying to use better tools (we could, but that wouldn't change a thing). DOM, CSS, and the imperative mutation model was a curse on front-end development, ensuring that only companies as big as Google could write something like GMail (for which they wrote their own Java-to-Javascript compiler first). If something needs fixing it is the underlying browser model. Maybe it will get fixed sometime if this trend continues to its terrible but logical extreme, and something finally gives. This is also purely a matter of cost - if businesses do care about performance then they'll have to put more developers per team, spend money on training, ask for fewer features slower. If performance _really_ mattered, then they'd see a loss in revenue, and as a natural response start talking about performance in their agile meetings, and JIRA cards.