Comparison
Cognitive loadvsMicro-frontend
Cognitive load
the team owns nine services and can explain four of them.
The total amount a team must hold in its head to do its work — domains, systems, tools, on-call surface. Treating it as a budget is the most useful team-sizing idea of the last decade, because it explains why adding one more service to a team can halve its throughput. It is also the strongest argument available against a proposed split or acquisition of ownership: not that the work is too much, but that the understanding required is.
Full entry →Micro-frontend
three teams each own a section of one page and each deploys their section without the others.
Splitting a single interface into independently owned and deployed pieces, composed at build time, at the edge, or in the browser. It buys team autonomy on a large product and is usually adopted for that organisational reason rather than a technical one. It costs duplicated dependencies, an inconsistent interface unless design system discipline holds, and a debugging story that spans several repositories.
Full entry →