Post

One of the problems Discourse in particular has is that it’s using SPA tech for no especially good reason… and that particular SPA tech stack has had some serious difficulties getting serious investment in performance on the axes that really matter, because of the dynamics you allude to at the end.

LinkedIn also uses Ember.js, and the problem I and others had working on it there was the same sort of dynamics you ID’d in the post you linked at the end: funding sustained investment in foundations is not generally rewarded or even appreciated, especially if it is “boring” compared to “features”.

The other dynamic that gets overlooked a lot is the importance of both perf cliffs and long-term effects. We would get pushback because improving the page load speed by a full second (25% improvement!), which one of my projects did, didn’t directly move business metrics. “So why bother?”