Close-ish to do done with that work stuff. Anyway.
Right now, my focus is kind of ping-ponging back to conlanging, and my feeling on that is, okay, sure, do what feels good.
Right now, I'm reviewing the kind of stuff I put together for one of my attempts at conlanging, and trying to figure out what kind of tooling I want. Here's how I'm looking at that:
- For the early work on a language, I don't know what works for me, and I need something that obviously-sort-of-works, so I can temporarily avoid thinking about tooling in order to have content.
- As I get more done, I want to use that as a guideline for refining tooling, but make sure that whatever tooling I come up with is appropriate for the earlier stages of completion.
My basic hunch is that I should be able to get by in the early stages with a mixture of text files and spreadsheets. The reference grammars I want to put together will look like:
Introduction: Plain text.
Phonology: Some tables, some text explanations.
Morphology: Some tables, some text explanations.
Derivational morphology: Some tables, some text explanations.
Syntax: Text, not sure if tables will come up; I'll need to investigate that.
Semantic fields and pragmatics: Text, with maybe some minor tables
Writing system: Text, with maybe some minor tables
Examples: Maybe some text, focused on tables, and requiring facilities for:
- Converting text from a simple-to-type romanization to a more expressive one
- Converting text from a simple-to-type romanization to the native writing system
- Converting text from a simple-to-type romanization to a phonetic representation
- Validating that a gloss is associated with a given root
- Validating that a root is realized correctly?
- Linking a root to the lexicon
- Linking a morphological component to the proper section
- If I wanted to make things very "straightforward" for me, I could try to set this up as "provide the intended root along with the intended gloss notation, and derive the other sections"
- Have a free translation below all of that.
- Could there be facilities for selecting different writing systems?
Lexicon: Tables, possibly using some of the above facilities. They should include:
- part of speech
In addition, my ideal wishlist is to track multiple stages of development and somehow maintain derivational links between roots at different points in the history of the language.
I'm... not sure how much of that can be properly done in Sphinx, like I'm hoping for as an endpoint. Like, at some point, do the requirements basically force me to be writing a Python file that essentially generates ReST nodes which somehow get awkwardly crowbared in. I think I wouldn't like that.
In any case, I'm going to start off with, plain text, plus either a spreadsheet, Python code, Lua code, or maybe try to make the tables in SQLite? I have to admit I feel some sort of weird interest in the idea of throwing away all formatting concerns and just typing stuff in DB Browser, buuuut some of these tables are going to need merged cells, so I can't use just SQLite. (Although, to be fair, the merged cells are for stuff that wouldn't really fit in a relational database anyway, so...)
Anyway, it's late, as usual, so I should wrap up and get to bed.