More explanations: why nano is Gentoo’s “default editor”

I say “default editor” in quotes because Gentoo really does not have a default editor, but rather a fallback-editor for whenever the users hadn’t chosen one themselves. And this is why stuff like the stage3, sudo, openrc and so on default to that when $EDITOR is unreliable (that is, until we can move to a full-fledged wrapper script).

Paul, in the previous post explained it quite well, albeit briefly, but since some other people asked me about that again, I’ll quote him and then go in a bit more of details:

… Third nano was kept as the default because it is small, straightforward, and easy to get out of when you don’t know it (let’s face it neither vi(m) nor (x)emacs is that easy to get out of). Fourth, by having the EDITOR/VISUAL variables point to nano (which is installed by stage2 as well as stage3) things would be working (although probably not optimal) out of the box. …

Indeed Paul speaks well: neither vi (or vim) nor emacs (or xemacs) is going to work. Why is that? Some people even goes as far to say that “it’s called *vi*sudo“ and “vi is the default Unix editor”. At the same time, I wouldn’t be surprised to hear “Emacs is the default GNU editor”. Well, let’s decompose it in two different problems.

Technical problem: the system set is messy already (and this is also “thanks” to my own pambase creation — which is, though, a solution to a previous problem). The packages that do enter the system set should be, whenever possible, small and with little dependencies. This includes, obviously, USE-optional dependencies, since when they are enabled, the dependencies also become part of the system set (like it happens with gnome and pambase). The dependency tree of nano is near invisible, while those of both vim and emacs are quite long (sure there are things like elvis, nvi, zile and jove, but this will bring us to the next point…).

Reality problem (or flame-retardant method): there is no way on Earth that any average number of people will be accepting to use the same editor as default, when taken from the list of “most popular” editors. Hey we cannot even find consensus among the three developers of lscube (I’m using Emacs – yeah I’m an Emacs user, burn me – and my colleagues using VIM and Eclipse). If we were to choose vim, all Emacsen users will complain, if we were to use GNU Emacs, both vim and XEmacs users will complain, and so on so forth. So we use nano, which is likely to make unhappy the most people, but all in the equal amount.

There’s a corollary to the reality problem above: if we were to choose any of the lightweight variants that I named above, or more of them, most likely the problem would be more or less the same, since almost all people would be pretty unhappy with the choice and would then unmerge that and install their editor of choice. But it would most likely upset the two “main factions” unequally, which would make one feel discriminated against.

Newbie problem: finally, there is another note: both Emacs and VI (and respective clones) aren’t exactly the most user-friendly editors. A newbie user who has no clue how to work in Gentoo is unlikely to guess at first sight how to use either of them, while nano is pretty much the easiest thing you might find around. You can fight as much as you want about powerfulness (Emacs and VI are obviously much more powerful than nano) and you can fight about relative easiness of use (:w versus C-x C-w) but nano is going to win over both of them in that regard.

So no, we’re not going to change the default anytime soon.