Re: eepitch with prefixes

[Eduardo Ochs]

On Sat, 9 Apr 2022 at 17:54, Tomas Hlavaty <tom@logand.com> wrote:
> (...)

Hi Tomas,

in the last years I realized that many of the design decisions that I
used in eev were not obvious, and I started to write about them and to
record videos about them. In some cases I also had to rewrite old code
to make it follow these design decisions more closely, and I was much
happier with the new code.

Some of the things that you are proposing are against the design
decisions that I've been following. Let me try to explain.

I have very little contact with "users" and in most cases in which I
needed feedback from "users" I only got that feedback after several
months or several years. This was very frustrating - but I realized
that I was already writing eev for "non-users", in this sense:

  (find-1stclassvideo-links "2021orgfornonusers")
  http://angg.twu.net/2021-org-for-non-users.html

  (find-1stclassvideo-links "2022findelispintro")
  http://angg.twu.net/find-elisp-intro.html

The feedback that I was getting from "non-users" was very good, so it
would be good karma to continue writing in the same style, i.e.,
mostly for them.

Let me ask you something before going ahead. When you talk about
"users", who are the "users" that you are talking about? My guess is
that these "users" can be divided into:

  1) you, when you see yourself as a "user" of eev who found something
     hard to learn,

  2) "concrete", "specific" users that you interact with, and with
     whom you would like to exchange e-scripts (mostly by e-mail, I
     guess?),

  3) "abstract", "generic" users that you imagine but that you don't
     interact with...

Can you say more about them? Can you quantify?

> It is not "hard", it unnecessarily complicates the user interface.

The user interface of eev is all designed around the idea that Emacs
can be used in a way that blurs the distinction between users and
programmers, and that makes the source code very easy to understand
and to test. A setq is much simpler than using file-local variables.

> eepitch-preprocess-line is internal implementation detail.

It is not - it is THE interface to a big world of simple hacks...

> I am just trying to say that it would be more convenient if the
> default covered the regexp use-case instead of being identity.

The default can be changed by putting something like this in the init
file:

  ;; See: (find-eev "eepitch.el" "eepitch")
  ;;      (find-eev "eepitch.el" "eepitch" "defvar eepitch-regexp")
  ;; and: https://lists.gnu.org/archive/html/eev/2022-04/msg00006.html
  (setq eepitch-regexp "^[:blank:]*[•]\\(.*\\)")

Do you read the Emacs News?

  https://sachachua.com/blog/all/
  https://lists.gnu.org/archive/html/emacs-tangents/

It always comes with links to lots of blog posts that are like "here
is how I configure the package foo-blah.el to make it super useful and
super easyful"... there were never any mentions of eev there except
for things that I wrote myself and submitted the link to the EWN
myself, though... anyway, would you like to write a something-that-
looks-like-a-blog-post explaining your preferred configs for eepitch,
and post it either to the eev mailing list or to a place that is more
blog-like? If you do that I will put a link to it in at least one of
these two places:

  (find-eev "eepitch.el" "eepitch")
  (find-eev "eepitch.el" "eepitch-this-line")

More soon!
  [[]], Eduardo Ochs
    http://angg.twu.net/#eev-blogposts