There are many reasons to switch to abstract schema beside support for other RDBMS systems. For example, if the table is deployed in production, then we can do drift tracking to make sure the schema stays the same as paper. Or producing documentation automatically.

In T270882#10809793, @Ladsgroup wrote:

There are many reasons to switch to abstract schema beside support for other RDBMS systems. For example, if the table is deployed in production, then we can do drift tracking to make sure the schema stays the same as paper. Or producing documentation automatically.

I see that, but this extension will not be deployed in production, and there are a lot of tables and stored procedures https://github.com/wikimedia/mediawiki-extensions-MathSearch/tree/master/db

If it's not deployed in production, then makes sense