Why Did Data Management Move from Files to Tables?
Data management started with files like Excel spreadsheets. When you keep separate customer, order, and employee files, four problems pile up: the same person's information gets duplicated across files, inventory numbers differ from file to file with no way to tell which is correct, searching means scanning entire files, and when two people save at the same time, one person's edits disappear.
In 1970, Edgar Codd, a mathematician at IBM, proposed the solution: store data in tables of rows and columns, connect tables with joins, and eliminate duplication through normalization, keeping each piece of information in exactly one place. This relational model and SQL dominated the next 50 years, and Oracle in 1977, MySQL in 1995, and PostgreSQL all grew on this foundation.
What Did NoSQL and Cloud Databases Trade Away?
In the mid-2000s, as social media emerged, vertical scaling, replacing one server with a better machine, hit its limits. Splitting tables across multiple servers makes joins slow, so NoSQL appeared, giving up some accuracy in exchange for speed and horizontal scalability. That is why banks, where accuracy is everything, still use relational databases while social media runs on NoSQL.
In the 2010s, managed databases became the norm, with the cloud handling installation, backups, and capacity. Databases shifted from something you install to something you plug in like electricity.
Why Was Storage Separated from the Server?
File storage followed the same path. Early web services piled user files into an uploads folder on the server, but when the server died the files vanished with it, adding servers made it unclear which file lived where, and file transfers alone could choke traffic. In 2006, Amazon S3 detached file storage from the server: upload a file and it gets replicated across multiple servers, and you get back a single access URL. This is object storage. CDNs then pre-copied files to edge servers around the world, solving the delivery speed problem as well.
What Did Supabase Bundle into One?
Supabase, launched in 2020, is the confluence of these two streams. Firebase gained popularity by offering auth, database, and storage in one package, but being NoSQL-based, it struggled with complex relational data and precise queries. Supabase put the same convenience on top of a proper PostgreSQL. Create a table and APIs are generated automatically; file access permissions are managed in sync with user data in the database. We moved from an era of bolting services together one by one to an era where what you need comes connected from the start.
What Changes When Non-Developers Know This Lineage?
The quality of your instructions changes. You can say that putting images in the database will slow it down, so upload them to storage and save only the URL, and you can distinguish between data where accuracy comes first and data where speed comes first. This is why SH Consulting starts its vibe coding training for non-developers with this lineage rather than tool usage.