When Parse announced their shutdown I (and several other developers) felt somewhat justified in having avoided (and heavily advised against) using services like Parse due to vendor lock-in and dependency on a single provider.

Over the years, I've been named as 'conservative', 'old fashioned', 'overly worried', 'paranoid', and more....as have others who have recommended not building software that is 100% reliant upon a 3rd party platform as a service, or upon a super-new technology.

As people begin their migrations away from Parse, we're seeing post after post of how to use Firebase, how to port to Firebase, how to run Parse's open source on your own server, or suggestions for other nearly identical services.

I fear that developers in the mobile space have not learned their lesson. Parse was owned by Facebook. As a result, many people claimed it to be 100% safe and not going anywhere. We see what happened. But Firebase is owned by Google, everyone says... Are people overlooking the products (and acquired companies) that Google regularly shuts down? Is that a gamble you and/or your companies are willing to take again? Are you going to rewrite your backend yearly as these services come and go?

Rather than jumping to the next Parse, Firebase, or other proprietary systems, developers should take a small amount of time to learn how to write their own backend service. The time it will take pails in comparison to the time your app took to write. Why are mobile developers so averse to writing their own backends? How long does everyone think that the open-sourced Parse server is going to stay maintained / secure? How long will that SDK stay working on iOS / Android after OS updates?

So, my question to everyone is this: Parse refugees have 1 year to move their code. Shouldn't they be spending it learning to write their own backend, rather than moving to another service where this is just going to happen again? Are we as an industry going to commit the same mistake again? Agree or disagree?

Disclaimer: Firebase and Parse were (and are) both great for prototyping. However, I firmly believe they should not be deployed to production as an integral part of any product.