Laurence.Tait@clc.org.au

Bug: Vector graphics don't work with repeats

Discussion created by Laurence.Tait@clc.org.au on Oct 6, 2020
Latest reply on Oct 11, 2020 by Laurence.Tait@clc.org.au

Vector graphics don't work in repeats after the first loop.

 

For the initial run through, they work fine, as can be seen in the image "first_repeat" below. 

 

On any subsequent loops through the repeat, the Vector graphics question appears to be selecting as intended. However it is not actually registering as a selected answer, as can be seen in the image "second_repeat" below.

 

This bug has resulted in the loss of a large amount of data. Not only has the data for this question not been saved by the app, but any subsequent questions that had "relevant" or "hidden" conditions that were dependent on the vector graphics questions were never displayed, and thus never answered.

 

A user can force the vector graphics answer to be selected by changing the answer from that given in the previous loop through, and then selecting the correct answer again. 

 

For example, if the answer in the first repeat is "Yes", then to enter "Yes" for the second repeat, the user needs to first enter "No" as an answer, and then enter "Yes".

 

This behaviour makes the problem worse, since it results in intermittent and unpredictable "correct" behaviour that makes it very hard to isolate the source of a problem, or even realise that a problem is occurring.

 

This is not a work around, since it completely screws up any questions that use the "Once" condition in the calculations. It is also incompatible with questions that lose their relevance or visibility in response to the question being answered.

 

If a field with a Vector graphics question is "required" the app will display the question as having been selected, but will refuse to let the user move on, as can be seen in the image "required_repeat" below. This behaviour leads to some users asking why the survey allows the choice of multiple answers, when they are required to choose only one. The implication is that users are interpreting this behaviour to mean that they are required to answer ""No" even when the correct answer is "Yes".

 

 

Outcomes