Some thoughts I've gathered over the years on what I think about GraphQL. All of this is subject to change of course, and some of it may be "hot-take"-ish, but at the end of the day, I've made decisions regarding GraphQL with my customers, users, and fellow developers in mind and with the mantra that if it ultimately doesn't make a big difference for any of those people and justify the work that it requires, it's not the best investment of time. It's more important to please your users, ship products in a timely manner, and use tools that keep processes simple and familiar.
- A majority of the world still runs on REST and probably will for a while.
- The challenges of larger companies that originally benefitted from GQL are not everyones challenges and they likely never will be.
- I don't want to require my API users to have knowledge of GQL.
- Strongly typed APIs are good, but I don't particularly enjoy the tools in the ecosystem right now to use them via GQL
- GraphQL seems to have been born out of the req