[supernes]

If your spec is entirely up in the air, you can start by replicating the examples. Lack of meaningful examples and/or in-depth documentation is the first red flag. You can also scan the reported issues to see if others are having trouble with core features, which is a sign of project immaturity.

Either way, it always helps to have some basic use case in mind. Bigger than "Hello, World" and smaller than a MVP. The outcome of any evaluation will heavily depend on whether it's a good fit for what you're trying to achieve.

[varorav]

Great point on being able to use motivating examples and testing out how user friendly the documentation and "getting started" is.