xLoop

My favorite approach, design DB use DB to generate data classes and interfaces, design API spec use that to generate server studs and API clients, now you can start the business logic and UI in parallel... schema driven development

This is because of the [legacy ness] of core banking systems, think about it if the date of the txn is been used how will reconciliation be smooth. Also to the core when the txn moves from pending to posted is can see that as a new activity and trigger a notification

Java or .Net been used in the banking space has very little to do with the technical details(types and others). It's more of ecosystem, reliability (battle tested), security and most importantly legacy (most banking systems are old so consistency and easy integration)

You got me on this one 🤣🤣

My take: if you are able to convince yourself you will need microservices in the future, start with modular monolith. In that future that might never come, just add proper devOps and you can get your dreamed microservices. But always ask yourself if you need it 🙃

Microservices on the table again, sad when people always copy architecture with proper context. Burning company's revenue when one digital ocean droplet could have done the job. In the end I think we fall in love with the tech so much that we forget why we are building it

Ohh sweet, they fixed the thumb button

Just everything about it is incredible ✨️

2. The data in the Jwt is public(base64), so if you keep vital info there and the FE saves it (because he must use it in other requests until expired) you are exposing too much.. And more. Just read the OAuth 2.0 guidelines or explore JWE or PASETO

I think it's part of the state And the jwt is not meant for the client to decode, because of a couple of reasons: 1. Normally the FE won't verify the jwt with the necessary key, so how do you confirm the jwt is coming from your actual BE. So a hacker the exploit the FE

🫣

But efficient is module federation against native federation