search

theRAPTLab / meme

DEPRECATED as of 2024-05-21. Use `theRAPTLab/meme-2023` instead. MEME development framework, using Electron, Webpack, and Express to create an "appliance-style" app server to end users on a LAN. Migrated from GitLab June 2023.
MIT License
0 stars 2 forks source link

Feature Request (?) "Model Sharing" between different MEME Servers #31

Open benloh opened 4 years ago

benloh commented 4 years ago

In GitLab by @daveseah on May 15, 2020, 11:44

This is a possible extension from #23: the ability to provide a cool "model" to another party (e.g. Joshua share a model they're using with Rutgers). In this case a "model" includes all the resources, etc that are linked to the graphical representation and underlying database. It does NOT include "admin data" (classes, students, teachers, criteria, resources)

It's unclear what the researcher/teacher needs to have copied as-is, and what needs to be reset/cleared. Beed to work out what this is before we can write a technical spec.

benloh commented 4 years ago

In GitLab by @jdanish on May 15, 2020, 11:53

@daveseah I've been thinking about this and it strikes me there are a few possible approaches as well. In case this helps.

  1. The model just includes the model, but not the resources/links to resources. This is the most likely use-case short-term, and I presume is easier to handle. And if we want we can re-link the resources. It is a bit antithetical to the whole idea of the design (that evidence really matters), but ... it's not like kids will be sharing these across sites.
  2. The resources are included, but then we likely run into a redundancy or mapping issue on import (yes, the thing called E1 in this is E1 in this, or no the E1 in this is E2 in this).
  3. We somehow handle these separately (I can move the evidence over, or the models, or one than the other, which might simplify 2, but you still have that issue).
  4. We assume we have the same evidence and anything that doesn't match up basically links to something blank and it's up to the user to fix it.
  5. Some other. For example, if it is a user-readable file format, you can change the evidence links by hand. And if they are empty they get cut.

If we need to, I think we'd be quite happy with 1. If we can get the other issues resolved, even better, but it wouldn't be worth say, blowing 1/3 of the budget on it.