#+TITLE: Lisping sideways
#+author: screwlisp
* tl;drae :don*t_read_anything_Else:
#+BEGIN_EXAMPLE
,* But knowledge representation :life:in:plain:text:
So org files.
Look like simple heirarchical markdown with some devices (names,
tags) displayed in pretty color and hyperlinks are smartly
clickable. Headings and source are foldable. Elisp source by
default.
#+END_EXAMPLE
* My description of what I normally do did not seem easy.
When I just want to jam a bit, it normally goes like this:
|-------+------------------+-------------------+---------|
| emacs | | | |
|-------+------------------+-------------------+---------|
| | new org file | | |
|-------+------------------+-------------------+---------|
| | | ensure (lisp . t) | |
| | | eshell sbcl | |
| | | (slime-connect) | |
| | | Title, heading | |
| | | demarcate source | |
|-------+------------------+-------------------+---------|
| | | | lisp |
| | | | writing |
|-------+------------------+-------------------+---------|
| | | save to my tilde | |
|-------+------------------+-------------------+---------|
| | update gophermap | | |
|-------+------------------+-------------------+---------|
| done | | | |
|-------+------------------+-------------------+---------|
this is what is happening right now.
Okay, I've now remembered
==M-x customize-variable <RET> org-babel-load-languages <RET>==
==Value-menu== Lisp -> toggle -> Accept-and-save.
So that's one resolved permanently by emacs-fu.
But my point remains. This doesn't seem like a great thing to suggest
other people try. Either everything I said is compeltely trivial, or
something isn't completely trivial.
* source management
Is it
- residential to the org-file
- lob-ingested from another org-file
- tangled
- To here
- To an ASDF system
- In ==~/common-lisp/==
- In ==~/.local/share/common-lisp/source/==
- In ==~/.emacs.d/my-elisp-package/==
- Something else
** About common lisp systems :aside:
This has reminded me of two things: How I get the vibe that Kent
basically wishes competing common lisp system facilities would compete
and then standardise, rather than ASDF having arguably kind of done
parts of what Kent thinks is a good idea, and everyone else just
agreeing to settle with ASDF forever. People being lisp compilers.
However having talked to lots of people who use any of the language's
various classic source loading management (and residential)
facilities, and remembering sometimes lisp source is held inside an
elisp package if it's closely associated with emacs, this isn't even
obviously true.
For example, some people just iteratively load a lisp file into their
current image, relying on the defvar / defparameter distinction to
preserve or clobber state.
It's not that any of the aboves are obviously the best; they have
different strengths and weaknesses.
** Everyone likes whichever one they started with
I see this line written about distinctions in programming or sysadmin
choices.
But I would argue two things:
1. These differences are distinctions
2. These distinctions have different strengths and weaknesses
Implying different hackers and different sysadmins, who appear to be
subtlely different are in fact exponentially different.
(If every small difference is a dimension; imagine how every weiqi
game turns out to be unique.)
** Emacs is the lisp editor
emacs has a tight relationship with the history of lisp programming in
the 60s, 70s and 80s, as well as 90s+ but less visibly.
So to have elisp emacs without a tenant lisp is obscure; and to have a
tenant lisp without an emacs is obscure.
In this case I am calling the big proprietary IDEs, and Interlisp
Medley, obscure though they really want for a special classication.
In particular, I think using lisp in the same manner one interfaces
with 8c is obscure.
* One hack at an answer
Well the topic wandered a bit; clearly it's not a simple situation
("what is the good?").
But I thought about it and I think I can distill my own essence like
this:
|------------+------------------------+-----------|
| Have emacs | | |
| Have lisp | | |
|------------+------------------------+-----------|
| | Configuration org-file | |
|------------+------------------------+-----------|
| | org-babel framework | New |
| | org-file | Local |
| | | org-babel |
|------------+------------------------+-----------|
* Knowledge representation rears eir beautiful head
Organising previous source inside emacs reminds me of how I've seen
and misunderstood people asking for ideas about organising
high-dimensional collections. I was always akin to saying, "hey why
not links in an org file" or else - what about a heirarchical
directory structure (like gopher or unix).
I'm going to potentially ill-use my rhetorical skull of Erik
Sandewall. I really like a cute line I saw in one of his 80s
publications, that if your knowledge representation is trivial text
files on a heirarchical filesystem, protip, you've failed.*
** I think Erik actually walked this back, and his indemnifying of software :aside:
individuals against having their brains fiddled with by external
forces, settling on basically encouraging students to jiggle the
s-expression sources of their personal novice-5 clones in their own
directions (while turned off, as long as it preserved coherent
individuality).
*** By the way, there must be a whole bunch of Swedish people in their mid :aside:
30s who have first hand experience of Sandewall's cognitive
intelligence courses.
* But knowledge representation :life:in:plain:text:
So org files.
Look like simple heirarchical markdown with some devices (names, tags)
displayed in pretty color and hyperlinks are smartly
clickable. Headings and source are foldable. Elisp source by default.