I think we should be careful when advising young or inexperienced people hacking for the joy of it, lest they make mistakes that are hard to recover. Time is precious -- better to choose wisely than learn from experience.
> Reverse engineering is fine...from a legal [perspective] in a lot of jurisdictions
Sorry, what jurisdiction permits one to violate a contract against reverse-engineering?
This is a pretty big problem that can't be hand-waved away. Companies everywhere have mere copyright on some materials that they make available if you agree e.g., not to reverse-engineer the interface or distribute the materials, essentially hoisting copyright to contract. This is how Java has worked for decades.
> Reverse engineering is fine. It certainly is from a moral point of view
In this case, the server is how Obsidian funds their work, offering free and paid parts. They're clear about it up front; they don't entice you and switch terms after you're committed. So how is it morally justifiable in this case?
> most have unenforceable clauses
The "enforcement" is not that they win their case, but the threat that they can take you to court and force you to pay a lot of money to escape the contract, especially in cases where you have significant investments or income. Unless you're dedicated to being so irrelevantly small that no one cares, why waste time building IP you can't benefit from, or telling the world that you don't care?
For people starting out, I would recommend picking viable projects, i.e., relevant to others, not illegal, not on dying technologies. They can exemplify not only your skills but also your judgment, and introduce you to the whole lifecycle of OS development if they have any legs.
And it may be best to find an unoccupied niche; in this case, there are alternatives already (better to invent than re-invent). (Unless you aspire to being a giant-killer :)
That's in part why the standard young-entrepreneur pattern is to pick new applications of emerging technologies, particularly those that irreversibly disrupt old industries or build new markets.
In this case there are many note-taking applications with syncing server. The unique project today e.g., might be to update an existing sync alternative to maintain a semantic model and then publish semantic cross-references, or publish a results of AI queries over the evolving set of highly-trafficked semantic topics.