>A piece of software like text-editor only has a shelf-life of a few short years before you need to start from scratch and write a new one.

A piece of software like a programming editor, which until now leaves a lot to be desired, even in the most popular implementations (Vim, Emacs, ST, Atom, etc), sure needs constant reinvention and refinement until someone finally gets it right.

Atom would be close to that if it was not fucking based on web technologies. ST3 would be that if it was open source (so the APIs/internals could be improved faster). Emacs would be that if it wasn't based on a 30+ year old UI model and elisp. Vim would be that if it was more extensible and not 100% focused on the CLI experience.

No, it's much better to avoid posting sarcasm at all. It nearly never scans, and it doesn't really fit the HN culture in any case.

Also, your underlying point is practically offensive to the hacker mindset. Though there is nearly a cult-like devotion to the editing tools of yore, the devotion is not universal.

I know /s, and perhaps I'm getting old and idiosyncratic, but I really don't understand why invent new text editors (Sublime, Atom, ...) at this point.

With emacs and vim around, I'd argue that text editors is a solved problem, if there ever was one. It is very hard to match emacs in terms of useful features or adaptability. Granted, the learning curve has a reputation for being steep, and elisp could need a facelift, but I think addressing these problems would provide much more value than reinventing the wheel with another shiny editor.

I disagree. Vim and emacs are still here and powerful

/s grandparent, methinks.

Intel's top scientists are hard at work developing even slower and more inefficient programming languages and display technologies. I'm confident by the time we have 10 GHz processors, we'll also have a new generation of software that makes it perform like a 66Mhz 486DX.