So in my typescript monorepo I split up code into core and services. Core contains all business logic and services simply call into core. I've been thinking about this more - does this actually make sense or is this a legacy pattern from the non serverless world? If you're all in on serverless, sometimes your business logic is implemented as a step function and splitting that creates a two way dependency. The historical benefit of splitting is so that core functions can be called in a variety of places. For example writing some kind of migration script. But in the serverless world, are local scripts even the right pattern? Most scripts like that are better modelled as step functions and you might as well run everything in the cloud

My business logic it’s on packages.

I guess another way to say it is I've optimized for running code in non AWS contexts. In practice that's just something local, tests or scripts

Imo the benefit of splitting is not reusability. it's Independence from your infrastructure layer so you can change things faster and write robust unit tests against your business logic. I struggle with the value of this too. It's a lot of work. I rarely put the effort in. I guess it is more valuable once we have a better understanding of the domain we are working in

Yep exactly. This is a pattern from DDD. But with serverless is there still that traditional split between infra and business logic?

Is that more because the step function tooling is awkward and doesn't enable good development practices?

I think I mention SFN because it's a good example. I think more broadly serverless blurs the line between infra and application so I'm starting to feel like the old patterns don't work as well

Yeah but wouldn't that stuff always be outside the domain?

Historically yes. But now it can literally be the implementation of a domain function

But we don't want those infrastructure details to bog down day to day stuff either. I don't really want my application code to know about some of the messed up single table dynamodb stuff

Domains can still encapsulate things from each other right

What do you mean by that?

You can still have domains that hide implementation details from other domains that don't need to know about it. I was mostly thinking about leaving infra code inside domains since they're doing more application like functionality these days

Ah okay thanks ☺️ It's worth saying I'd say like 60-70% of the code I have written in the past year using serverless is just like inline single file functions