1. I started work in interpreted languages. The ethos/culture at this level is that there's always "something that will make things faster without losing the GC feature". Hence Java, Go, Rust etc. are in this category of sweets.
2. Now I'm just forcing myself to do C to throw out all the mysterious grey optimisation trade-off questionmarks lock-stock.
3. PLT : in lang(or any)-design, the front-end design matters most ... back-end design is boring : we already know how things work. But if someone sells/ships a poorly DESIGNED product, then it handcuffs the back-end implementation options. This continues to be interesting on a daily basis when I think about langdesign.
4. In general, the design of things is "80% can it be built"; "20% can it be sold". Mistakes in marketing are when people sell3 but can't ship what was promised, or can't maintain it past the gotomarketstage.
No comments :
Post a Comment