| 15:02 | <jschoi> | I’ll have to do a more detailed read of this later. On first read: Looks like there were a bunch of pent-up feelings at TC39 ad a whole, with some people against any further new syntax in general. Which the ones expressing them eventually acknowledging them as off-topic to your question. Looks like the most salient concern there is about namespace collisions, which are a bigger deal for standalone functions (whether ordinary or this-based) than for prototype inherited prototype methods. Looks like most people there simply do not care much about DCE and webpage weight…or at least they say things like, “3D assets will dominate webpage weight, so code size doesn’t matter, and preserving no-build workflows are still important or more important.” |
| 15:19 | <jschoi> | It is funny to see someone claiming that shrinking client-side code weight is a mere “toolchain niche”. Surely they’ve encountered, in their everyday life, the truth of the 10+-year-old meme that the Web is too heavyweight now (e.g., https://thatshubham.com/blog/news-audit / https://news.ycombinator.com/item?id=47390945 or, further back, https://idlewords.com/talks/website_obesity.htm). And when most people on the planet, especially in the global south, have capped ISP data…one could even call claiming that “client-side code size concerns are a niche” is a privileged position. Webpages that use Three.js are not somehow exempt from this. At least the “assets will be bigger” argument makes some sense…but certainly poo-pooing all client-side optimization into a niche does not. |
| 15:27 | <jschoi> | Anyways, namespace collisions seemed to be the most important concern. Though there are plenty of other large APIs that use deal with them? Like the standard libraries of Go, Haskell, or Clojure. Or any large C API that uses prefix conventions—though I suspect that approach might be distasteful to many in that thread. |