GraphQL Abandoned: That Didn't Take Long

Basically, after my first stab at it, I read some articles. And confirmation bias kicked in.

Well, possibly.

Basically, there was a page arguing the pros and cons. And it pretty much just summed up my initial impressions. GraphQL seems like a better answer to the types of services I'm not presently making. But, it might make increasingly more sense with a Microservices architecture in an API Gateway. Though, that is certainly dependent upon a lot of other things.

The direction I've decided to go in is to build out an OData API, but one which only returns DTOs containing the information I want. It is small, but not a Microservice. So, I don't have a ton of things to worry about, and I would have ways of automating the heavy lifting anyway.

So, I won't have much more to add on GraphQL. Though, I get to play with the latest release of AspNet Core OData. I may convert the thing to OpenAPI at the end. Though, I may choose not to as well. If this goes well, this project will replace a piece of software I actually have running out in the real world.

The biggest goals are to modernize an older piece of software I have and make the database layer redundant. Messing with new service layers that don't offer any tangible benefits just make the project riskier in my opinion. "Modernizing" already means a whole lot of change as I'm going from WPF to Vue.js and from a .Net 4.5 desktop app to an AspNet Core 5.0 web service.