Ah interesting. Well at the risk of being a little over-scrupulous, an issue I have with this flow is that you still need to commit to a version number to test successive versions of your schema while iterating. Then, when you finally arrive at a schema format you want, you need to change the schema’s self.version field back to 1-0-0 (or whatever) and publish it to your actual Iglu, and also update the code where you’ve inserted your tracker to point to the new version as well. This is the part that strikes me as a bit error prone. It seems like this is sorta the exact thing the schema draft service was created to avoid, right?
Basically I want to avoid the case where I’ve created a new_schema/1-0-12.json while devving against one Iglu instance (be it in SnowplowMini or my own staging Iglu instance), then publish that exact same schema as new_schema/1-0-0.json in a different Iglu instance.