“Maybe we should make a page of project ideas?”
Yes, we should. It would a crucial starting point for anybody thinking to contribute to the Crystal ecosystem.
IMHO, we should NOT invest time in developing tools - like IDEs, version managers and so on - because Crystal can already use many of the existing ones with good results. For example, I’m using Visual Studio Code as a IDE for Crystal, at the moment, and I’m totally satisfied. I would not use any other IDE for the same reasons I do not use Spyder when working with Python: I do not want to use two different IDEs.
Instead, we should focus on at least one “killer application”: something that could convince other programmers and other users to spend time with Crystal. In particular, we should focus on something that can attract C, C++ and Python programmers, not (only) Ruby ones.
This because what we really need for Crystal is to exploit its capability to re-use existing C and C++ libraries, in particular the libraries that made Python the most diffused programming language worldwide: OpenCV, TensorFlow, and so on (NOT Pandas because it is not easily “portable” to Crystal).
I do not actually know what can fit a 175 hr frame but I suspect that porting Crystal to Windows and/or developing new ML libraries could easily take longer. Focussing our attention on a finished, working binding generator (crystal_lib, bindgen, etc.) could be a wise choice. Making a OpenCV (C++) binding for Crystal could be another one.
NOTE: We already have a couple of “killer apps”: Kemal and Amber. Unfortunately, they both attract mainly Ruby, Rails and Sinatra programmers/users who mostly use the existing Crystal code to replace Ruby. This does not help Crystal to evolve.