Re: Imports / inclusion of s.el into Emacs

[Dmitry Gutov]

On 01.05.2020 21:05, Philippe Vaucher wrote:

> I'll try to explain my side first so you see where I'm coming from. So, 
> I guess a lot of s.el users think it's a massive improvement over what 
> emacs has to offer reguarding string manipulations. This is visible in 
> issues like https://github.com/magnars/s.el/issues/88.

I imagine this impression might come purely from seeing a set of these 
functions collected in one place. And not necessarily from including all 
the helpers as well.

> I never coded 
> with clojure, but my argument for these helpers is that in code we do 
> string manipulations all the time, they are trivial to write, so why not 
> have trivial helpers.

The criticism would be exactly the same: if they're trivial, why have 
them in the library at all.

> I understand there's a limit to such helpers 
> because you could write thousands of them, so good taste should be used 
> to introduce "the minimal set".

Exactly. We do have some of them in subr-x already. We can add some 
more, though I'd only pick the non-trivial ones, among those that are 
likely to be used.

> I usually don't care about these extra 
> function calls, what matters more to me is how beautiful the code looks 
> and how readable it is.

Often, I'd agree. But the tradeoffs when creating the standard library 
of the language are somewhat different, IMHO.

> Because string manipulation is so trivial, I'd like them to be 
> grouped together so they are easy to find. In other languages I can 
> usually go to one page and have the list of all strings related 
> functions. In Emacs the manual at 
> https://www.gnu.org/software/emacs/manual/html_node/elisp/Strings-and-Characters.html while 
> helpful makes it hard to "read it all" and find what you are looking 
> for. For example I couldn't find concat on that page.

I fully support your criticism here. The old-timers will tell you to use 
the manual (from Emacs), but we both know that a lot of user are still 
much more likely to resort to a web search, and there the manual is not 
so great (no index, at the very least).

>  From your side, I understand that your definition of the minimal set is 
> much thiner, and maybe what would help is to give the complete list of 
> the functions you find superfluous and why (list at 
> https://github.com/magnars/s.el so far you mention `s-prepend` and 
> `s-append`). For example I'd disagree if you said `s-left` and `s-right` 
> are not useful to import because it can be done with `substring`.
I will respectfully refuse this invitation: I have other work to do.

As for 'concat' being hard to find: adding a 'string-concat' alias 
(without any plans to remove the original) could satisfy both sides, maybe.