I’ve just started the Phoenix part of the utterly brilliant online course by @pragdave. On generating the Phoenix app he uses the --no-ecto
flag and explains why:
I don’t think anybody needs to add a database layer into their Phoenix application. Because Phoenix is purely concerned with communicating on the web, and anything you use a database for should be in its own separate service. So I don’t see any point at all in putting Ecto into any Phoenix application that you generate from scratch, you always want to put your logic into separate applications in the way we’ve done so far.
Do you agree? Is this how you build your Elixir/Phoenix apps? Can you think of pros/cons?
If you haven’t got Dave’s course I highly recommend it. There’s much, much more like this and I’ve learned such a massive amount from it already - particularly on HOW to approach modern development using Elixir, rather than ABOUT Elixir itself. It’s definitely one of the best online course I have ever done… and I really don’t say that lightly!