Arguments for OpenEdX vs. Canvas?

[Jane Manning]

I've been asked a few times what the arguments are for using OpenEdX over Instructure's Canvas product, and I've got a rough-and-ready answer, but not one that's super-detailed, since I haven't spent a lot of time looking at Canvas.

The differences that come to mind for me are:

Structural: 

Canvas is the product of a for-profit rather than a non-profit, and keep some features (e.g. analytics) out of the open source version.  

I don't have a sense for how realistic it is for people outside of Instructure to expect to have their pull requests accepted - looking at their top 10 contributors at https://github.com/instructure/canvas-lms/graphs/contributors it looks like there are two who don't list themselves as instructure employees, so perhaps it's possible (or maybe these people are contractors working for instructure - I just don't know).  

Canvas doesn't seem to actively use github issues or some other public location as the place where instructure employees do bug-tracking. (OpenEdX doesn't either, but it's early enough days that there's hope that that may change.)

There's a single provider of commercial hosting and support for the platform for people who don't want to run it themselves: Instructure. 

Feature-wise:

Canvas seems much weaker on features that are important for MOOCs - e.g. it doesn't have anything like the OpenEdX video player, its range of auto-gradeable assessment types is much weaker, and its discussion forums don't seem to allow any sub-forum structure, or be able to be included on a particular course page like OpenEdX discussions can.

On the flip side, Canvas has strong on-campus LMS features - gradebooks that are downloadable / uploadable, strong support for manual grading by instructors, ability for students to turn in work as a group, etc.

So my guess is that it's the suitability for the MOOC use cases, and perhaps the desire to use an open source product that isn't controlled by a for-profit, that have driven most of the OpenEdX adopters, and most either don't have on-campus classes, are using something different on-campus, or are using OpenEdX in conjunction with other tools on-campus.

Can someone who's looked into this in more detail than I have either verify that I've got it roughly right on the differences, or let me know what I'm missing / misunderstanding?

thanks a lot -

Jane

[Yarko Tymciurak]

Jane -

I don't know that you've missed anything, but my exposure is light.

A friend I've talked about MOOCs with a bit had his publisher offer a new book from him on Canvas - so my experience is as a student.

I was less than thrilled, both as a student, and as a tech-observer.   I think you've got the forum comments, and video (he had nothing more than a 10 minute intro into the next reading, which all did "on their own").

- Yarko

[Claude Coulombe]

Greetings,

I've recently got a discussion about Canvas after the publication of a recent comparative study from Indiana University.

Indeed, Canvas is a very good LMS platform! In terms of features, it is quite possible that Canvas outperforms most of its LMS competitors at least in their current versions. Canvas platform also benefits from having been created around 2008 so after Ajax and Web 2.0 but before the cloud main emergence.

Personally, my main complaints against Canvas are:

As a LMS provider : 

1) Weakness of the development community (654 on the list, most of whom are users not developers) 

2) Effective control of the Instructure company on the development of Canvas (open source or not) 

3) In terms of technology, the choice of Ruby on Rails. Honestly I prefer Python, but it is a questionable / debatable choice... ad infinitum...

As a MOOC platform:

1) Not originally conceived to be scalable (Canvas MOOC early adopters have got some real issues)

2) Not enough modularity, I really enjoy the XBlocks architecture of Open edX

3) Weakness of the learning analytics which are crucial for MOOCs

4) Weakness of the assessments API and architecture