This article contains this release's feature updates, user interface changes and bug fixes.

Important changes

We wanted to remind our users that for the next release, two important changes will be deployed. These changes were already announced at an earlier stage. In addition, two new important changes have been added to this release. The overview of all important changes that may require action from administrators can be found here.

27th of August 2023 - Removed deprecated LTI 1.2 integrations
On August 27th, all LTI 1.2 integrations will be removed. From this date onwards, Ans will only support the LTI 1.3 integration.

• Fill-in question (if the ‘show answer’ option is enabled)
• Multiple choice (both one correct answer and multiple answers)
• Match
• Order
  
  

8th of October 2023 - The guess correction feature will no longer be applied by default to unanswered questions
Previously, if participants had not answered a question in an assignment, their responses were still corrected by the guess correction. With the upcoming change, a new setting will be introduced for the guess correction, which allows the correction to be disabled for unanswered questions. This setting will disable applying the guess correction for unanswered questions by default for newly created assignments.

29th of October 2023 - We remove the possibility for employees to access the answers of participants before a digital test has been submitted from the "Monitor participants during a digital test" beta feature
Based on our investigation, this option is not in line with the European GDPR laws. Other than this, the beta feature will stay unchanged. If you have any questions, please refer to your Customer Success Manager.

Changes to the platform

Added a search bar to the dropdown menu in the timeslot dialogue if it contains more than 7 options

We have added a search field to the new dropdown menus that contain more than 7 options. For now, this improvement only applies to the dropdown menu used to select participants within a timeslot.

Added the option to disable automatic enrolment in courses upon entry via LTI

We have added an option to block automatic enrolment in courses via LTI. This option is enabled by default to ensure that the existing LTI integrations keep working in the same manner. The option to disable enrolling users in a course has been incorporated in the LTI Integration settings, which can be found via the school settings.

Open beta changes

With the beta feature "Exclude exercises from the publication", exercises are now grouped by flow group

Previously, when using flow for a digital test and the "Exclude exercises from the publication" beta feature in the publication settings, we only provided an overview of all the exercises within the assignment. Now, when using the "Exclude exercises from the publication", the exercises are grouped per flow group and they can be recognised easier.

API changes

You can now retrieve who reviewed a specific submission via the API

The GET and PATCH responses from ​/api​/v2​/submissions​/{id} now include an array of reviewer ids. With this addition, you can now access and identify all the reviewers associated with a specific submission.

Bug fixes

• Fixed issue where assignments created with a template were archived on creation
  This occurred when the template assignment was archived. Now, new assignments are never archived upon creation.
  
  
• Fixed an issue where clicking on the "Read more" button did not work correctly when viewing an empty question bank assignment
  The "Read more" button did not work as intended at the question bank exercise overview. The button now links to an article within our help centre.

   

• Fixed an issue with syncing the placement of descriptions within an exercise
  While relocating a description added to an exercise from a question bank, the synchronisation process failed to consider the description's position, resulting in the inaccurate placement of the description.
  
  

• Fixed an issue where copying gaps that contained HTML in their name would cause the gap to be unusable
  Copying a gap that utilised HTML tags in its name led to the incorrect setting of the gap. This resulted in the gap's functionality being compromised when copying it into a question bank question or included within an assignment.

   

• Fixed an issue where the participants' list did not update when creating or updating a timeslot on the taking page
  The participants list now updates after saving a timeslot.
  
  

• Fixed an issue where a student, with multiple active sessions in ProctorExam, would become stuck on the loading screen
  Previously, students had the capability to start multiple Proctor Exam sessions, which prevented them from accessing their tests. This update prevents users from initiating multiple sessions and provides a warning regarding the presence of ongoing sessions.
  
  

• Fixed an error in email links to published results with Schoolyear integration
  Resolved an issue where clicking on a link in an email to view a published result for assignments with Schoolyear integration would lead to an error. The links now direct students to Schoolyear without any errors.
  
  

• Fixed an issue where the start reviewing button was visible in the preview of a peer review digital test
  Within a peer review digital test, instructors have the option to preview the test. After clicking submit in the previewed test, a new page would be shown giving the impression that they could also preview another peer review result. Start reviewing would result in a "Something went wrong error. To avoid any confusion, this action has now been removed.
  
  

• Resolved reliability calculation issue in peer-reviewed assignment
  Previously, the reliability of reviewers in peer-reviewed assignments was calculated based on their evaluation of both closed and open-ended questions. Because closed-ended questions are automatically reviewed, they should not have contributed to the reliability of reviewers. This has been updated so that only open-ended questions are taken into account.
  
  

• Fixed an issue where the assignment results page would briefly become unavailable after enabling a second review
  Previously, the assignment results page would briefly become unavailable after enabling a second review.

• Fixed an issue where a second review would be created for a preview result
  When previewing an assignment and submitting the demo result, an extra result would be created when the Second Review feature was enabled. This is now no longer the case.

• Enhanced lockdown browser protection for assignment result domains overview
  Addressed a vulnerability that allowed students to circumvent lockdown browser restrictions and access the results domain overview in cases where the assignment was intended to be protected. This fix reinforces the integrity of assessment conditions by ensuring that the lockdown browser functions as intended, maintaining a secure environment for protected assignments.

• Fixed an issue where the p-value of a question was still shown while the question no longer had graded answers
  When a question in a question bank was unlinked from any course, the p-value was not reset to zero and instead displayed the p-value from before the question was unlinked.

• Removed the "Available" label at the overview of the published assignments for students
  The "Available" label for published assignments on the homepage has been removed as it caused confusion for students due to its conflicting styling with the "Live" label. Removing this label simplifies the interface of the homepage and helps students distinguish different types of information more easily.

• Removed non-functional "approve" and "disapprove" options that were visible to students in the settings of a published result
  Previously, students with result export privileges could access the results settings page, where they encountered inactive "approve" and "disapprove" buttons. These buttons were misleading, as they lacked functionality for the student. The buttons are no longer visible to students.

• Fixed an issue where the "Assignee" dropdown list was cut off when assigning a comment
  This occurred when assigning a comment to a contributor in a question bank exercise. Now users can scroll through all contributors.

• Fixed an issue where adding a protection rule label to results would still allow changes to the result files
  An assignment label with the protection rule Disable results changes did not prevent editing or deleting result files even though the banner indicated that this should not be possible. With this fix, the protection rule "Disable results changes" now correctly prevents any changes to the result files.

• Fixed an issue where restoring a grading scheme to a previous version from the log would cause an error
  This error would only occur if students' answers had already been reviewed. The error no longer occurs.

• Synchronised submission attributes in API V2 response and logs
  Fixed a problem where the attributes in API V2 submission responses were different from the attributes in the corresponding submission logs. This has been sorted out, and both endpoints now show the same attributes consistently. Some new attributes have been added to the submission response, like adjustment, raw_score, auto_graded, none_above, archived, trashed, and trashed_at. This makes sure that when users use API V2 for getting and putting data, everything works together smoothly and completely.

• Fixed an issue where error messages would not be displayed when linking assignments via Deeplinking
  An error message will now be shown instead of the default broken page error by the browser when encountering an error while linking an assignment via LTI Deeplinking. The error message contains information explaining why the Deeplinking request failed.
  
  
• Fixed an issue where the general feedback in the QTI import would lose HTML tags
  During the import of multiple-choice questions using QTI, the general feedback would undergo HTML tag stripping, leading to the removal of its formatting.

• Fixed a vulnerability issue where remote code execution was possible in the code editor of exercises
  When creating exercises, the user had the possibility to define certain variables with the help of a code editor. This allowed potential attackers to execute commands on the server. The code editor functionality runs on a separate server, so the vulnerability does not impact Ans' integrity, nor could the attacker access Ans data. We have performance an impact analysis and have found the vulnerability has not been exploited.

• Fixed an issue where the timeslot displayed on the cover page of a written assignment did not always display the correct time
  When adding a timeslot to a written assignment or bubble sheet assignment, the default timezone shown on the cover page was always set to Central European Time (CET). We now make sure to take the time zone into consideration when displaying the timeslot on the cover page of a written assignment or bubble sheet assignment.

• Fixed an issue where an administrator would not be able to assume a user's account if the user had Two-Factor Authentication enabled
  Previously, if an administrator needed to assume an employee account and Two-Factor Authentication was enabled, the administrator would be prompted to input the Two-Factor Authentication code, even though they didn't possess it. With this fix, administrators can now successfully assume accounts without encountering the authentication code issue.

• Fixed an issue where upside-down scans, in certain instances, were unable to correctly recognise the page number after being uploaded
  In some rare instances when a scan is uploaded upside-down, the code on the top right of the page would be incorrectly identified. This would cause the scan to display scan issues.

• Fixed an issue where selecting the archived filter after already searching for an item would not yield any results
  When attempting to search for an item and then selecting the "Archived" filter, no results would be returned. Selecting the "Archived" filter first and then performing the search would return results as expected. With this fix, filtering and searching can be performed in any order.

• Fixed an issue where changing the ownership of a result to an unknown student would not correctly assign the ownership
  In a written assignment, if the ownership of a result were marked as "Unknown student", the result would be bound to a new student account with an incremented student number. With this fix, the result will instead be bound to the student number "0000000".