Here we describe the naming system inside emacs would you describe it as verb-subject?
C-h f kill-TAB shows quite a bit of stuff
As to our own code, use the package prefix, and then subject-verb. According to MeatBall:HumaneInterface, subject-verb is better suited to how humans approach action. We often know the subject and change our opinion about the verb; the other way around is strange to us. Example: You know you want to work on a buffer. Switch to it? Bury it? Kill it? The other way around is unnatural: You want to kill something, but was it the word, the buffer, or the process?
Thinking about it, you’ll see that currently Emacs is not following this strategy. On the contrary, M-x apropros RET \(kill\|remove\|delete\) RET shows many functions with optional package prefix, verb, subject. This is also true for \(create\|make\), with some exceptions like buffer-create (vs. kill-buffer).
I agree that Emacs does not follow this strategy. I do not know what solutions are available. My only thought is a MeaningfulNames package that provides aliases that do follow the Humane Interface patterns.
> Isn't this <em>object</em>-verb?
Subject-verb would be buffer-die rather than buffer-kill.
> Yes this is <em>object</em>-verb which makes a lot of difference. The noun-verb approach referred to by MeatBall is discussed in the context of user interfaces. One of the main arguments for the noun-verb order rather than the verb-noun is that choosing a verb then a noun introduces a mode (the verb selected), which does not happens if the sequence is noun-verb, because selecting a noun usually does not introduce a mode. When programming a convention like verb-object matches the usual order of words in a sentance and is probably fine.