Motivation:
In Wiki-style learning environments new content has to be reviewed by expert users to be termed as reliable. This process is essential for maintaining the quality of the content published. However, this process of “Review” itself can be difficult and slow when we concern ourselves to articles seeing rapid contributions as the articles need to be continuously evaluated by the experts to retain reliability. Therefore, we need to ask ourselves that – Is there a way we can speed up this quality check mechanism hence encouraging participation while ensuring content reliability?
This project is aimed to tackle the above problem by developing a Rating Engine for WikiToLearn (a wiki style learning platform ) based on various parameters like User’s Opinion, Author’s Credibility,Page Interconnections and Revision History to evaluate the reliability and quality of articles on the platform.
Timeline:
The project was developed over a period of 3 months with regular feedbacks from the WikiToLearn mentors with an average programming effort of 40 hours a week.
Learning Outcomes:
I have been exposed to a numerous contemporary tools and frameworks and aligned my work closely to the industry standard. I acquired considerable amount of skills in the following domains:
- RESTful Web Services framework - Jersey
- NoSQL database (OrientDB)
- MediaWiki Web API ecosystem
- MediaWiki Extension Development
- Container Technologies (Docker)
- PHP and jQuery
- Shell Scripting
- GIT
- Agile Software Development practices
- Open Source etiquettes
Challenges:
As the rating engine was developed completely from scratch, it was a fairly complex task. I faced friction on a lot of fronts some of them were:
- Deciding suitable frameworks to develop the project and read undocumented code.
- Developing a Rating algorithm that will integrate old content with the new one.
- Understanding the Media Wiki Architecture well enough to write an extension.
Impact:
Now the users of the wiki platform will be able to assess the quality of any content just by looking at the score generated by the Rating Engine.
The entire Wiki Structure is now modeled into a logical graph.
MediaWiki Extension in Action