Building my own metadata library
Inspired by some of the insights in “The Magic of Small Databases” 202301292232, “Build your metadata library” 202301291444, and “Why We Need More Boutique Search Engines” 202301292238, I’ve decided to build my own metadata library using Datasette.
Note: Why can’t I use Airtable for this?
Airtable would work great for my purposes! Except for one, glaring issue: it has a complete and utter lack of search, and I really want to see what workflows I could unlock once I have the power of actually functional search at my fingertips. (See 202301292246)
#update 02-20-23: It’s funny that I claimed that I wouldn’t do this on Airtable because it has bad search, and ended up coming to Airtable for this anyway. (See https://airtable.com/appeI4Fto3INw0Je2/tbliavWk8DfgNiCoL/viwbDNQJS6hax069i?blocks=hide). Some of the reasons:
- Airtable’s search is good enough once you learn how to use linked records to connect things that you actually care about. This isn’t even that hard to backfill with scripts, if necessary.
- I’m pretty committed to Airtable as a product anyway given that I work there, so I don’t feel too bad about having my data ensconced within Airtable’s system. Worst case, I’ll export my data to PostgresSQL or something using https://uibakery.io/move-airtable-to-postgresql
- Airtable is very easy to set up, has great automation and API, and already has a nice user interface already. It can even serve as the backend of a website if I really wanted it to: https://mzrn.sh/2022/04/29/using-airtable-as-a-jekyll-website-database/ Some of the reasons why I might not want to do this in the future:
- What we look for in a resume — Having my own Datasette setup looks really good as a resume project. Airtable is slightly less impressive.
Things I want to use Datasette for:
- Things 3 tasks
- Books I’ve read
- Workouts I’ve done
- Notes I’ve written each day
- Spotify playlists
- NYT spelling bee
- Beeminder
- Pinboard
uid: 202301292209 tags: #starred #relational-thinking #projects