That Google felt the need to invest millions of dollars into making a document delivery platform into something resembling an app platform is not really relevant here. Sure, it is robust. But the whole point of the thread is engineers are notoriously bad at asking what the customer _needs_ from a product vs what the engineer finds fun to write. You want to have fun as an engineer? Work on a side project.
The larger the dependency, the more it should have to do to justify itself. "It's fun to write in" is not a sufficient justification for a 70mb dependency. "I need a cross platform app" is better.
As an engineer, you're responsible for the transitive dependency graph that gets delivered to customer machines. If you want to take on gigantic dependencies because they are the current lovechild of ascendant Medium bloggers, be my guest. I prefer to go off of more than fun and social proof when making engineering decisions, however.
The larger the dependency, the more it should have to do to justify itself. "It's fun to write in" is not a sufficient justification for a 70mb dependency. "I need a cross platform app" is better.
As an engineer, you're responsible for the transitive dependency graph that gets delivered to customer machines. If you want to take on gigantic dependencies because they are the current lovechild of ascendant Medium bloggers, be my guest. I prefer to go off of more than fun and social proof when making engineering decisions, however.