XBlock is a component architecture by edX.org for building courseware.
This is a pre-alpha release of the XBlock API, to gather input from potential users of the API. We like what is here, but are open to suggestions for changes. We will be implementing this shortly in the edX LMS.
This repo contains the core code for implementing XBlocks as well as a simple workbench application for running XBlocks in a small simple environment.
EdX courseware is built out of components that are combined hierarchically. These include components like the video player, LON-CAPA problems, as well as compound components like learning sequences. We are developing a second-generation API for these components called XBlocks. Although they're in a prototype stage, we like the API, and want to collaborate with others to develop them into an industry standard. This is our proposed API and specification for XBlocks.
How does this differ from existing industry standards like Learning Tools Interoperability (LTI) and SCORM? On a high level, XBlocks is a Python language-level API, and it provides sensible defaults for things like storing data. XBlocks could be wrapped up in LTI, and one could make an LTI XBlock. The core reason to write an XBlock is that it is deployable. You can give us the code to an XBlock, and we can embed it in our courseware. LTI would require you to give us a virtual machine image which ran it.
This code runs on Python 2.7.
-
Get a local copy of this repo.
-
(Optional) Create and activate a virtualenv to work in.
-
Install the requirements and register the XBlock entry points with (you may need to sudo this if you don't use virtualenv):
$ pip install -r requirements.txt -
Run the Django development server:
$ python manage.py runserver -
Open a web browser to: http://127.0.0.1:8000
When you open the workbench, you'll see a list of sample XBlock configurations (scenarios). Each will display a page showing the XBlocks composited together, along with internal information like the "database" contents.
The workbench doesn't use a real database, it simply stores all data in an in-memory dictionary. The data is all lost and reset when you restart the server.
If you want to experiment with different students, you can use a URL parameter to set the numeric student ID, which defaults to 1:
http://127.0.0.1:8000/?student_id=17
Different students will see different student state, for example, while seeing the same content.
The docs for the XBlock API is on Read The Docs: https://xblock.readthedocs.org .
There are distinct layers to this code. In the docstrings, we've tried to make clear which layer the code lives in, though sometimes the lines are blurred:
-
XBlock The sample XBlock code. This is the most important, it is the code that most third parties will be writing, and demonstrates the XBlock interfaces.
-
Runtime The runtime code that we think will be common to all runtimes. This is the behind-the-scenes code edX will write to make XBlocks work. This layer may not be real code here, but we'll need to write real code to perform these functions.
-
Workbench The workbench-specific code we wrote to make everything work. This code is the least real, in that it might be just stubs, and will be done very differently in real code.
-
Thumbs This is a sample XBlock written as a separate installable kit as an example of how third-party XBlocks can be structured.
Making an XBlock can be as simple as creating a Python class with a few
specific methods. The thumbs module demonstrates an XBlock with state,
views, and input handling.
You can provide scenarios for the workbench to display, see the thumbs.py sample for an example, or the xblock/problem.py file. The scenarios are written in a simple XML language. Note this is not an XML format we are proposing as a standard.
Once you install your XBlock into your virtualenv, the workbench will automatically display its scenarios for you to experiment with.
The XBlock mailing list is edx-xblock on Google Groups. You can also write to edX directly at info@edx.org.