boring technology

has been around
just works
has a future
makes simple easy (avoids accidental complexity)

boring technology is a shark, not a dinosour.

this page has an approach for evaluating the boringness of technologies.

bts

bts is the boring technology score. you calculate bts by multiplying the following factors:

longevity (<5 years = 0, 5 years = 0.1, 6 years = 0.15, 10 years = 0.5, 20 years = 1.0)
accidental complexity (0.1 = extremly high - typically more than 10 deployed components, 0.5 = medium - typically more than 5 deployed components, 0.7 = low - typically more than 3 deployed components, 0.9 - very low - typically less than 3 deployed components)
works and has community (1 = >50k, 0.5 >10k, 0.2 >1k, 0.1 <1k github stars)
activity (1 = last merge within a week, 0.5 = last merge within a month, 0.1 = last merge within a year)

extremely boring technology stack

extremely boring technology stack is an opinionated technology stack that works for a surprisingly large number of use cases. the accidental complexity is minimized by just using postgres for everything, using multi-page app for almost everything, and when that fails, using hypermedia-driven architecture.

technology	name	replaces	comments
backend	django	fastapi, flask	you will quickly outgrow the simplistic solutions, just use a basic django template
frontend	django	REST + react/vue	use admin where possible, mpa for all user views, and htmx/alpine when that does not work
database	postgres	sqlite, mysql	while postgres is more complex than sqlite, it has features allowing you to avoid other components
background tasks	django-tasks	celery	celery established, but it is too complex
caching	django	redis	use plain django caching for basic needs. for ‘webscale’ caching, use django-distill + s3 + cdn
full-text search	postgres	elastic
queue	postgres	rabbitmq, kafka	postgres queues can be used to get ACID properties
periodic running	superchronic	cron	version for cron that runs in docker container
multi-process	honcho	systemd, k8s	running multiple things inside one paas contaoiner
dependencies	pip	poetry, uv	uv looks promising, but so did poetry
deployments	docker + paas	k8s	just use a paas, such as fly.io

about

i do django consulting,

contact me at mikko at boringtechnology dot org