.@ccorcos builds up the concept of a relational database from first principles: ccorcos.github.io/filing-cabinet…
I've been working with databases for ~30 years, but this piece had me re-examining some things I take for granted.
Wish I had something like this when I was first grappling with indexes and query planners!
Wish I had something like this when I was first grappling with indexes and query planners!
Generally speaking I love the ELI5 approach: start from a beginner's mind and pull in concepts one at a time.
(Another example is this piece wyag.thb.lt on building Git from scratch by @ThbPlg)
(Another example is this piece wyag.thb.lt on building Git from scratch by @ThbPlg)
A brilliant move in Chet's article is using badge numbers as a way to understand unique identifiers. 
Unique IDs are a critical concept in computing, yet they are counterintuitive for non-computer folks. We're used to ambiguous identifiers like names of people ("Jane Smith") or streets ("Green Ave").
Long UUIDs can be intimidating and seem unnecessary when you're used to names.
Long UUIDs can be intimidating and seem unnecessary when you're used to names.
If I have one complaint about the article, it's that I'd like to see the discussion of "end-user databases" (Airtable, Notion, ...) explored further!
It brings to mind 1980s-era databases like dBase, where end-user and programmer were not so sharply delineated as today.
It brings to mind 1980s-era databases like dBase, where end-user and programmer were not so sharply delineated as today.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
