[robatsu]

Top 10% - gets the coding job done, timely, high quality code. Knows his craft well, but is still essentially a craftsman.

Top 1% - Designs systems that will get the job done for the business/larger org goals, or even designing systems that generate business props. Also, the systems designed are capable of being implemented/maintained by lesser skilled developers. Of course, knows his craft, but is something of the artist, can see the Buddha in the gearbox.

The 1% guy should easily, as in no conceptual difficulties, be able to write a compiler/interpreter, that clearly demonstrates whether one is simpatico with the machine, capable of thinking like one. Actually should love doing this sort of thing, probably, and might have to resist the the tendency to incorporate custom command processing components/metadata/languages in every project.

[eli_gottlieb]

The 1% guy should easily, as in no conceptual difficulties, be able to write a compiler/interpreter, that clearly demonstrates whether one is simpatico with the machine, capable of thinking like one.

You do realize that compilers and programming languages are an actual realm of expertise, right? I can, and do, write a compiler, but I couldn't construct you a web framework from scratch.

Acknowledge others' expertise rather than trying to construct a mountain to stand on.

[robatsu]

Sorry, I think I poorly worded my comment. The point I was trying to make was that (imo) the better devs have an extremely deterministic viewpoint towards their code and are always paying attention to all the contextual layers in which they are coding and how their code may affect that context.

The example of compilers/languages sort of follows from this, as this (constant) awareness leads to pattern recognition which then leads to surmising about leveraging languages, code generation, etc, to take advantage of these patterns & reduce effort/errors.

There are other examples, just that I picked this one.