Week of 22 August 2016

The following information summarizes what is new in the edX platform this week.

The edX engineering wiki Release Pages provide detailed information about every change made to the edx-platform GitHub repository. If you are interested in additional information about every change in a release, create a user account for the wiki and review the dated release pages.


An error that prevented content library values from being imported along with ther course data has been resolved. When you import a course that includes randomized content blocks, the Display Name and Max Attempts values are now correctly imported along with other problem settings. (PLAT-1046)


Updates to Options on the Discussion Page

To improve learner and course team experiences with course discussions, this release includes several improvements to options on the Discussion page. The following images compare the placement of several options on this page.

Before this change:

  1. Add a Post
  2. Search All Posts
  3. List of posts
Add a Post button in top corner, search all posts field above the list of posts, and an All Discussions option above the list of posts to allow selection of a specific topic.

After this change:

  1. Add a Post
  2. Search All Posts
  3. List of topics, with options to show all posts instead, or to filter the list.
Search all posts field in top corner, Add a Post button to one side, and an option to show All Topics or all posts and a filter topics field above the list of topics.

For more information, see “Navigation Changes to Discussion Forum (partners)” on the edX partner portal, or “Navigation Changes to Discussion Forum (Open edX)” on the Open edX portal.

Performance Improvements for Grade Computations

Over the last two months, a series of incremental improvements have been made to grade computations. As a result, system performance for loading the learner Progress page has increased significantly over the past three weeks.

The graph that follows shows the effect of the first of these changes, which released at approximately 1:00 pm EDT on 29 Jun 2016. The graph shows a significant drop in response time, from eight seconds to approximately four seconds, immediately after the initial phase of the work was released.

Graph showing that the median load duration dropped from over two seconds to approximately one second. The load duration for the 95th and 99th percentiles were even more significant, dropping from over eight seconds to approximately four.

Subsequent performance improvements were then made that built upon this initial transformative change.

The graph that follows shows a comparison of how long it took to load the page, in seconds, for the two week period from 29 Jul 2016 to 11 Aug 2016 to the period two weeks immediately prior. The graph on the top compares median durations, and the graph on the bottom compares 95th percentile durations for these time periods.

Top graph shows matching trend lines of between 0.55 and 0.6 seconds for median load time over the first five days, then a drop of 0.15 seconds on 2 Aug and a general trend down to 0.35 seconds on 11 Aug. For 95th percentile values, trend lines matched at around 1.5 for the first five days, followed by a more gradual drop to under 1 on 11 Aug 15.

Other Changes

  • The Staff Debug Info option now appears only next to assessment types that allow course team members to reset the number of attempts or rescore the problem, such as the core CAPA problem types. (SOL-1809)
  • This release includes changes to enhance the accessibility of page navigation elements. Button elements now replace links and “href” elements where appropriate. (AC-530)
  • An error that prevented learners from being able to preview discussion posts has been resolved. Posts, including those with Markdown formatting, can now be previewed. (TNL-5198)


Courses that offer coupon codes can now specify valid email domains for accepting coupon codes. Coupon codes are accepted only for learners whose email addresses match the valid domains.