The development of Atom shows a common anti pattern of software development: choosing the technology first. The goal was to develop an editor with web technologies, and not to develop the best editor in the world. The result is as sad as foreseeable: a buggy and slow memory hog named Atom. When I tried it last time, a couple of days ago, it was impossible to enter an @ sign into the text area.
The problem is that "best" is subjective. I think one could more accurately say that they're trying to develop an editor that is easily extensible with technologies familiar to many developers.
i.e. your perception of what they want Atom to be and the reality of what they want Atom to be are probably not the same.
My take? Revisit this in five years. I bet it'll be one of the richest ecosystems in software development. As one who has written Atom plugins (live unit testing w/ real-time feedback), I've never encountered such a developer experience until Atom.
Lets say if one of the main goals was to build a very easily extensible editor, one might still choose to build it on web technologies.
While I don't use atom, I can see the reasoning. Since I'm comfortable with webdev, the barrier to entry for extending atom is much lower compared to other newish editors.
I assume their goal is to someday integrate Atom into GitHub proper (running in a normal web browser) in order to have developers living in a 100% GitHub ecosystem. The chosen stack makes a lot of sense given this objective.
