@guillemcordoba I put some more thought in your question about what would be useful in learning holochain.
As I said, having small Rust only toy programs are very interesting, because they are easy to read and only reside in the Rust ecosystem. It has less moving parts, and therefore a bit easier to get working.
I’m currently learning Rust and trying to understand basic concepts like Path, capabilities, …
And having to work only in Rust, relieves me of dealing with the extra complexity of Typescript, Npm, node and setting up related tooling.
But I can imagine that for people who are trying to learn how to build apps with UI’s and want to combine multiple dna’s and learn about the default conductor, this is not what they need.
For those purposes the same tactic might be usefull. Making examples or tutorials in a way that eliminates of minimize the use of Rust, could help people focus on the specific thing they are learning: e.g. writing testscenarios, playing with conductor functionality, …
I’m not sure it is possible, but if you could work with precompiled dna’s, that would relieves some of the complexity of setting up a Rust environment.
I know in the end, to make usefull hApps, it’s necessary to know the full stack. But having typescript only and rust only examples, might lower the barrier for developers to try out holochain.
I know from experience, that having a working toy example, builds confidence and triggers more curiosity. And on the contrary having quick start examples that don’t build or require a lot of set up, discourage beginners.
I think holochain RSM needs to provide solid, reproducible and stable examples if it wants to attract more people.
Case in point: the cmdchatter example worked perfectly, but now 2 weeks after it featured on the newsletter it gives a compile error. I know RSM is still evolving quickly and cmdchatter is from the dev community, but still I’m fearful these things with discourage newcomer and affect holochain’s reputation.