search

openedx-unsupported / ease

EASE (Enhanced AI Scoring Engine) is a library that allows for machine learning based classification of textual content. This is useful for tasks such as scoring student essays.
GNU Affero General Public License v3.0
216 stars 96 forks source link

Remove "obsolete", as ease is a requirement of ora2. #72

Closed cahrens closed 7 years ago

cahrens commented 7 years ago

I don't think this should have been marked obsolete. edx-ora2 declares ease as a dependency (though I don't think it is used by any code we are actually running in production).

@edx/teaching-and-learning: please review

FYI @staubina

staubina commented 7 years ago

👍

robrap commented 7 years ago

@cahrens You make it seem like we could drop the requirement from ora2 and mark this as archived. Why wouldn't we want to move in that direction if it is unused?

cahrens commented 7 years ago

@robrap Sorry, my comment was overly simplistic. This is my understanding of the usage of ease, though honestly I haven't digged into it much.

  1. ease is used for AI grading
  2. There is a lot of code and some tests in edx-ora2 for AI testing
  3. We don't expose the AI option in Studio
  4. We never completely finished the AI work (at least to the point where we are comfortable advertising it)
  5. But if one sets up their problem in XML, they can enable AI

The summary is that it is not simple to rip it out, and it is not entirely clear to me that we wish to rip it out.

pmitros commented 7 years ago

@cahrens While removing ORA is clearly out-of-scope for this PR, perhaps some directional guidance would be a good question for @marcotuts? My feeling is we either want working features, or we want to remove things we're not going to use from the platform. Getting AI grading over the line would be nice, and something I'd be supportive of. But if we're not going to prioritize that, and have a feature which we're not even sure whether it works at this point, it makes sense to remove it. That's just technical debt and unnecessary complexity. If we only prioritize this work in a few years, it's still probably simpler to move EASE out of archive, and redo the integration from scratch. By that point, any integration code we have today will have a half-decade of bit rot.