Notes
Notes from the workbench.
Working notes on the patterns we use when we build. They fall into three domains: data and tenancy, constraint solvers, and regulated SaaS. Useful if you are evaluating us; useful if you are solving the same problem yourself.
- Data and tenancy
- 2 notes
- Constraint solvers
- 1 note
- Regulated SaaS
- 1 note
Data and tenancy
01 · 2 notes
Isolating tenants, reconciling records, keeping the schema honest under load.
- Read →
Fuzzy record linkage in Python: when rapidfuzz is enough
Most reconciliations need a normalisation pass and a string distance, not a machine learning system. Until they do.
Python · Fuzzy matching · Record linkage
- Read →
Multi tenant Postgres with row level security: a working pattern
RLS is one SQL clause and three operational disciplines. Skip a discipline and the clause stops protecting you.
Postgres · multi tenant · RLS
Constraint solvers
02 · 1 note
Modelling hard scheduling problems and choosing the floor to build them on.
Regulated SaaS
03 · 1 note
Where the law meets the schema, read for the people who design the tables.