Memextogether

..

DRAFT

In Build Your Own Little Memex with Babashka, I called for a community effort to build memexes together, and build tools for building memexes together. I put a start into code at https://github.com/teodorlu/lemex.experimental/. Thinking back, I don’t think I started this right.

I presumed to be able to create a cohesive namespace of memex operations from the start. Instead, I believe a good first step is to assemble a ball of mud and be able to share that ball of mud.

Another way to start:

1. Collect functions community members have already written in one place
2. Avoid breaking those functions

We could put these functions in teodorlu.memextogether.mud.

Package conventions:

1. Don’t break consumers.
2. Avoid lots of dependencies where possible.

But we should depend on some libraries. I propose:

• babashka/fs
• babashka/process
• babashka/shell

Candidates:

• An RDF database of some sort, eg datascript, datahike, xt or datomic.
  • Problem: a whole database is a lot of dependencies
    • Possible mitigation: memextogether depends on a databse interface, memextogether consumers bring their own database.
  • Problem: per 2023-10-21, I (Teodor) haven’t been able to make use of relational in building play.teod.eu. I have some relations, but I basically only use those relations to build the home page. There is no search, there is no querying, there is no structured data modeling.