There are a couple of reasons but I can cite a few, I tried to push for PaaS and failed the hard way:
1. No widespread practice of environment provisioning automation. Today Docker and k8s are ubiquitous. Having a list of requirements your application needs well kept were hard to find - which Heroku needs.
2. In the same sense as above, automated deployment was not a common thing.
3. Barriers to cloud adoption. Development tools were not widespread. Paying for software development and deployment is common now, it wasn't.
I agree. If we can start a culture of relying on SQLite instead of PostgreSQL/MySQL, a whole server-side application can be a simple standalone binary.
Also, having a binary makes it easy to bundle in an Electron app.
They do something which is "common" in Apple-land (Dash for MacOS does this, a Twitter client did this as well, over and over): They get the same app, add some features, call it v2, launch as a new app and remove v1 from the store.
They don't "give" you the new version. They take away the app you paid "once", and provide you with a version with an expire date. So you have no choice. You either pay them, again, or lose access to the v2 (subscription based) app.
I don't mind paying for good software, I even think Hallide is worth $60. But I won't make the same mistake again. So best of luck Lux! I really wish you all the success. If you treat your customers right this time.
1. No widespread practice of environment provisioning automation. Today Docker and k8s are ubiquitous. Having a list of requirements your application needs well kept were hard to find - which Heroku needs. 2. In the same sense as above, automated deployment was not a common thing. 3. Barriers to cloud adoption. Development tools were not widespread. Paying for software development and deployment is common now, it wasn't.