The Breakfast Post 1 made an excellent laborous study, poring over about 500,000 lines of code across nearly 2000 source files of SML code, on the anarchy of Standard ML’s guidelines on naming and separating with the capitalization rules the different aspects of the language, namely variables, structures,
signatures, functors and exceptions. They drew from these conclusions of a standardized convention through argumentum ad popularum, by apppealing to the wide adoption of each guideline between the most prominent projects programmed in SML. This argumentation could be compared to Common Lisp’s still-born standardization process, which had the more popular implementations blackmailing their concerns with threats of backing down from the project. I think this fails to see the variance on ergonomics between text input from layout to layout. My concerns apply to Ocaml’s general guidelines as well.
To start with, I had to take a closer look at a US/GB QWERTY -layout, since I
have never had to use one in my entire life, and the very first thing I notice is
that I can already see the non-issue of reaching towards the hyphen-underscore (- / _)
key, which is located in the number row, to the right of 0. It is a lot easier
than in SV QWERTY which has the - / _-key where the ? / /
would be located in a standard US QWERTY; and where the ! / § sits at in the
FR AZERTY.
I did some purely subjective experiments to compare if it’s as easy as it looks and I quickly came to the conclusion, that having the underscore, where it’s easily pushed by one of the more dominant fingers, is a lot better than a location, than a key reached with with a pinky finger. After all, throghout this I have the other (left) hand’s pinky holding the Shift down.
This is a simple fact of ergonomics being a huge matter for the rest of the
world, at least for the DE QWERTZ and the SV QWERTY (which is used all
throghout the Fennoscandia!) users. Both of these layouts have the
- / _ located right next to the right-hand Shift.
I understand why these things wouldn’t matter for the native English-speaking, or the French-speaking programming world, since you have your underscore accesible, with a simple keychord of Shift+8. A character, that is fairly simple to reach from the homerow, with having the ability to use your eg. middle-finger extended forward, is definetly not something a programmer would find themselves straining to type. Now compare this to the concept of Emacs pinky, a repetitive strain injury said to be caused by having to curl your pinkies repeatedly to hold Ctrl/Meta/Alt/Shift while chording the keys. 2
I can’t emphasize how much I’ve come to abhor snake_case and it’s variants.
Underscores in general too. For most, Lisp’s parenthesis-leaden syntax is an
abomination, while some might accept it as a neccesary evil for the efficient
metaprogramming facilities. For me, though, the fact that there are no other
delimiters or special characters to worry about than a parenthesis and a hyphen 3
acts as a leveler, along with it’s quite unusual and alien syntax for most
programmers nowadays.
(This is what specifically attracted me to Scheme the first time
I actually tried typing it out, rathen than moan about being Lost In a Sea of
Parens. Even the naming convention, kebab-case, is pure bliss!
Not having to keep a finger on guard duty, ie. sitting on top of Shift, can k
quite liberating.)
While SML and Ocaml don’t allow hyphen-delimited names due to the lexer, snake_case is still avoidable
I simply resist the practice of delimiting words with underscores in names, so I’ll resort to
camelCase and DromedaryCase (or more derogatorily PascalCase) when needed. I know it’s against
the Standard-library’s conventions on SML and the suggested guidelines for Ocaml, but mostly
I’m programming on my own. If I had to collaborate and add to someone else’s project, I would of course
follow the convention set forth in it. Otherwise others will have to conform to my convention and
I’m highly opionated, as evidenced by this, on the matter.
-p-suffix.