Every few months a new framework promises to change everything. My selection criteria haven't moved: can the client hire for it, can a small team maintain it, and will it still be supported when the engagement ends.
Hiring pool first. A stack the local market can't staff is a maintenance trap, the system becomes dependent on whoever built it. Mainstream choices aren't boring; they're sustainable.
Maintenance burden second. Every dependency is a future upgrade, every service is a future outage. The stack with fewer moving parts wins unless the requirements genuinely demand more.
Longevity last. I look at release cadence, governance, and who's betting on it, not stars or hype cycles. The stack should be a boring choice in five years, not a cautionary tale.
The test I apply: if I disappeared after handover, could a competent local developer pick this up? If the answer needs qualifiers, the stack is wrong for the project.
