Future aritecture of GroMoPo Discussion

[rreinecke]

Hi @Castronova, @aaraney, @samzipper, @kbefus,
sorry for the delay. I came down with COVID. So as promised here is an issue to discuss the further development of the GroMoPo App.
In my view the most sensible thing would be to go with the least "intrusive" option of utilizing the Hydroshare user authentication in the app itself to have some sort of access control.

All ideas that we discussed were a bit more involved.

Robert

[samzipper]

Thanks for starting the issue @rreinecke ! And hope you are feeling better! I am back from vacation and full of GroMoPo good vibes.

The least intrusive option sounds good to me. I worry slightly about creating a barrier (people would need to have or create a HydroShare account to contribute to the GroMoPo database) - that might be enough friction to push away some people who are on the fence about whether it is worth the effort to share or not.

What if anyone could submit data, but someone with a HydroShare account has to check off on it? That way, people who are already invested in the project, like the students leading the data entry, would not need anyone to verify there entries. Anyone new who comes across the website could still submit with no additional steps, but then it would trigger a review from one of us. I think the number in that second category would be small enough that I would be happy to take on that responsibility.

I don't know whether that is a practice middle ground from an implementation perspective, though! Would love the experts chiming in there.

[rreinecke]

Thanks @samzipper !
I agree that it would introduce an additional barrier but it would also make a lot of things much easier:

• We don't need to store any authentication information in the app itself -> which currently is not possible because the app and all its contents are pulled from github. So we would definitely need to move away from the free streamlit service and develop our own deployment system.
• Also if people can submit data that cannot directly be pushed to hydroshare. So we would need some kind of in-between storage solution.
• And we need to implement the whole review system

So in the end we would end up with something like that, but with the added possibility (and complexity!) of using a hydroshare account:

I agree that we would find people to take on the review side of it. But I fear that we currently do not have the man power to implement this solution. Its feasible but someone needs to implement, test, and deploy it.

[samzipper]

Gotcha! That makes sense to me. In that case, I agree with your initial diagram as the most practical short-term solution. If we run into barriers down the road, we can figure out what type of adaptations are needed.

@Castronova @aaraney, how does that sound to you?