Calen Pennington (Deactivated)

Braden MacDonald

ClintonB (Deactivated)

Adam Palay (Deactivated)

Nimisha Asthagiri (Deactivated)

Julia Eskew (Deactivated)

Nimisha Asthagiri (Deactivated)

Marco Morales (Deactivated) dfriedmanR (Deactivated)

Steven Vergenz (Unlicensed)

Colin Fredericks

GiulioG (Unlicensed)

Chris Rodriguez (Deactivated)

Omar Al-Ithawi

C

Chris Rodriguez (Deactivated)

Related tickets (from Zachary Kramer (Deactivated)):

• PR-576
• PR-250
• BLD-135

Nate Aune

stv

TJ Keemon

Fred Smith (Deactivated)

ClintonB (Deactivated)

Xavier Antoviaque

Replace the Vagrant VirtualBox support with Docker.IO. Compared to VirtualBox, Docker is much more lightweight and it offers native performance for the app. Check https://docker.io for info about docker. As a stretch goal, I'd like to add setting up a clustered Docker.IO setup to allow dev. to have closer environment to production. This is especially doable with Docker because of it's relatively low memory consumption.

Note from Nate Aune: We've already built some Open edX Docker images (edx-lite and edx-full), but they're all-in-one, and we'd like to split the services into different images to have better separation between application and database, for example. Here are the instructions on how these images were created using Ansible. We started working on a multi-container Docker in this docker_multi_release branch of our forked configuration repo just MongoDB and MySQL for now.

Could possibly use Fig from Orchard (now owned by Docker) to handle the linking of different containers.

If you don't already have Docker installed on your computer, and you have a Mac or Windows machine, the easiest way to get it installed is to use boot2docker.

We haven't updated our fork of https://github.com/django-wiki/django-wiki (https://github.com/edx/django-wiki) since forever - Adam Palay (Deactivated) is that correct? Do we have a sense of how much work this would be?

Adam Palay (Deactivated): yeah since at least 2012. I know we'd made a few changes to our own fork, so we'd have to make sure they were compatible with new versions. But yeah, updating would be great.

btalbotR (Deactivated)

ClintonB (Deactivated)

Sass is indispensable to the design team, but the way we currently use it requires us to run a ruby process as part of edx-platform. Using libsass would allow us to remove that from our stack and give us dramatically faster compilation of stylesheets (some benchmarks show a 20x+ improvement there). Some wrinkles that would have to be resolved:

1. We use bourbon and neat (we can probably just move these to vendor includes instead of running the programs).
2. We're currently on Sass 3.3.5, Bourbon 4.0.2, Neat 1.6.0 – libsass support for 3.3 features is spotty, bourbon would have to be pulled down to 3.2 (which was meant as a libsass compatible bridge release to 4.0), and I'm not sure what the situation with neat is. We'd have to verify that we weren't using features that would break anything, or find workarounds if we did.
3. The way that the asset pipeline/watching happens would have to be adjusted. The edX UX team has some thoughts on this.

JesseZ (Deactivated), clytwynecR (Deactivated)

David Adams

We have two sets of selenium-based acceptance tests: the legacy tests built on lettuce and the more reliable tests built on bok-choy. We will be working on converting all the remaining lettuce tests to bok-choy.

If you want to learn (or improve your skills at) bok-choy or to really dig into a particular feature by helping to convert its tests, we welcome you to join our team! See this page for more details.

You could also make the xblock-sdk tests use bok-choy properly, as a warmup exercise.

Nimisha Asthagiri (Deactivated),

stv, dfriedmanR (Deactivated)

Julia Eskew (Deactivated), benweeksR (Unlicensed)

XBlocks can have children! Let's start moving more XModules to be XBlocks! For each particular XModule, this would mean:

• Removing XModule as one of the base classes it inherits from.
• Making sure they use the runtime properly.
• (optional) Extract it from edx-platform (e.g. templates, JS dependencies)
• (optional) Change it to obey XBlock conventions about not doing work in the __init__ (particularly capa and videos).

Some example XModules to target:

• (easy): purely structural blocks that don't really use the runtime much: Sequence, Chapter, etc.
• (intermediate): capa, which already has had a lot of the underlying work done so that Google's runtime can make use of it.
• (hard): video module, which has a lot of dependencies on things in platform.

Calen Pennington (Deactivated): Please add anything you think pertinent.

Note: Studio's lack of support for editing non-xmodule xblocks will be a major impediment to merging any of these ports.

Build a CI pipeline for devstack. Goals: to test on a regular basis (at least once/day, at most as many commits as possible of edx/edx-platform and edx/configuration):

• Creating a devstack instance from scratch
• Re-provisioning on a longer-running devstack instance

Should also run smoke tests against the devstack target once provisioning completes, with some retrofitting on the test_build_server role in the configuration repo.

Note: due to virtualization limitations, this cannot be done on EC2 completely end-to-end; however, portions of it (such as the ansible provisioning) can.

Implement Cassandra as a significantly more scalable backend for the XBlock user_state scope in the LMS runtime. The expected integration style is to implement a new CassandraKeyValueStore, replacing the DjangoKeyValueStore presently being used.

In terms of data model, this sketches out a proposed starting point:

• Row Key: Student + Course Run
• Column Keys: Module Type + Usage Key (Module State Key) + Timestamp
• Column Value: UTF8 String (JSON serialized state)

KaTeX is a new math typesetting library built by Khan Academy engineers, which purports to outperform MathJax by up to 20-30x (ex. http://jsperf.com/katex-vs-mathjax).

KaTeX as yet only supports a subset of MathJax functions and is therefore not a drop-in replacement for existing courses. As a mitigation we could provide a course-level switch where the author/admin can choose between the two libraries.

Richard Machielse (Unlicensed)

S (Deactivated)

steveR (Deactivated)

ChristinaR (Deactivated)

Xavier Antoviaque

Rework Studio's editing framework so that it is not dependent upon xmodule descriptors but can work for any arbitrary xblock. Some thoughts:

• Allow xblocks to declare that they wish to support this new style of editing (don't break existing editors)
• Have Studio render the Save and Cancel buttons and have Save call a method on the xblock
• Allow an xblock to declare which tabs it wishes to show and Studio will ask the xblock to render each one when needed
• Provide a default "Settings" tab view (in Studio only, initially) that is metadata-driven
  
  • Allow an xblock to opt in to the "Settings" tab and indicate which fields it would like to have shown.

Out of scope thoughts for the longer term:

• Have a separate repo for an xblock-editing framework (or maybe more generically a model-driven editing framework)
  • Studio would use this framework, but then so could other runtimes
• Have Stevedore entry points so that any package can register new xblock field type editors

Add the ability to either:

a) Replicate the state of a course to other instances of edX, and merge changes from other instances. This would allow each member of a team of developers to have his/her own devstack, and can thus make edits to a course without fear of overwriting another team member's drafts/changes. Perhaps some sort of git integration with the course import/export tool?

b) Indicate what users are currently editing a course page, preferably along with the ability to see their changes in real-time in the draft view. This could allow out-of-band communication between team members to make sure no toes are stepped on.

If path a) is taken, this could tie in with the Studio history viewer proposed by dmitchellR (Deactivated).

Marco Morales (Deactivated)

Diana Huang

This is a two part project:

1. Create a merge-able PR to move the LMS user account settings to a dedicated page, accessed from the User drop-down menu
2. Getting creative with the new user account settings page

Renzo Lucioni (Deactivated)

Alasdair Swan (Deactivated)

Develop Xblock and LTI plugins to integrate Cognii's natural language processing assessment API into Open EdX platform. The main tasks include:

1. Get the open edX system working on Amazon AWS (using the AMI)
2. Implement the Xblock or LTI plugin for external assessment service

1. Combine the existing pages into a single 2 Text field form.
2. Build off Learner Account page that francesR (Deactivated) is leading, to create a progressive registration form that will contain all the extra fields currently on our registration page.

1. make deployment easy using ansible + git
2. box as wireless access point, configuration in ansible
   

rePatcha XBlock

http://repatcha.org/

https://github.com/TechNikh/xblock_repatcha

1. needs third party auth - github
2. add short text field to Annotator.js
3. vote up/down a patch