Created: 2020-09-22T01:38:44+00:00
1. First impression was, "I have org-mode why do I care about this?" 1a. Annoyed that "it isn't org-mode." 1a1. Key bindings like C-j instead of Alt-Ret. 2. I do like that each "cell" has its own path. It is possible to say "go to cell 1.2.3" or "1b3," wheras in org you can only go to a named headline. 2a. In a sense org is designed to be used with yet more tools. Tools like counsel and helm make it easier to select nodes interactively when you go to refile or such. 2b. These tools are powerful but add more complexity to the emacs installation. 3. Easily broken, tries to be clever 3a. In org what you see is mostly all there is. Then formatting operators are used to make parts of that pretty. And code folding is used to hide things you don't want to see. There are fragments of ex. "property drawers" here and there. 3b. org is basically a helper that edits text that looks org-ish, but hides nothing from you unless it's been asked to. You can easily create and manipulate structure with non-org functions as long as they "look org-y" in the end. 3c. Kotl wants to take things over and misses a few spaces. Sometimes when creating a new node while in a node with multiple lines, you will get another node element in the wrong spot. The clever editing routines then prevent you from fixing the markup. 4. Uses up some of your magic 4a. Kotl consists of separate header, content, data, metadata pages which are hidden using narrow. This means you cannot use the narrowing for your own purposes, as one can do in org to focus on particular subtrees. 5. Based around the NLS/Augment design 5a. I think this is where kotl really starts to make more sense. Compared to org-mode it is not that special. Org had more utilities, more support, recognition, even some tools outside of emacs support it. But NLS/Augment is about paragraph-level "statements" and not big headlines with contents. 5b. For example in org you can link to a headline, and possibly use unique ids to track headlines as they get refiled around. But the contents of those headlines does not exist from a logistical standpoint. 5c. In kotl and in NLS/Augment, each "statement" or paragraph is it's own "headline." This does not strictly matter up front. 5d. Where each thing having its own identifiers starts to matter is that you can link between one paragraph to another paragraph and that link stays alive when the paragraph is moved around in the document. 5e. So strictly kotl is worse; it does not do TODOs like org does unless you bring your own scripting. It does not have all the extreme navigation aids like helm-org-rifle. It does not have extensive support to publish documents or outlines with. 5f. Kotl becomes interesting instead through the lens of NLS/Augment: when conjoined with the "buttons" support of hyperbole and other documents written with those buttons in mind. 5f1. You write documents in outlines such that each paragraph is numbered. Those documents then explain some particular thing and perhaps they cite other things. The citations are implicit or explicit "buttons" which link back to say, an e-mail message thread where some decision was made. The outline then ties all the relevant messages together with buttons and is itself a new document that can be referenced. 5f2. So think org-mode, but each paragraph silently has its own UUID that can be linked to from anywhere else in the project. 6. Uses serial numbers to identify all things, but then also has relative paths. 6a. These are written to the file but hidden through emacs magic. So each bullet point actually begins with some "0145" type number and then the contents of that cell. 6a1. The leading zero signifies an absolute ID. 6b. Unique serials identify the bullet point, content and structure (via indentation) but meta-data is kept on a separate hidden "page." 6b1. The meta-data is a list of s-expressions which hold the serial number for the cell they refer to.