5.3.17 issue confirmed

Do you know if the bug can be fixed? If so, how long would it take? To know if I can wait for the bug to be solved to share my questionnaire, because of deadlines I have

Thank you in advance

The order is determined by a seed which is saved along the response, so the order can be reproduced later.
If we change the way the random order is seeded it would invalidate the data for people who already collected responses using the current system.
That is why we cannot just change it on the fly.

@c_schmitz the issue here : it's the same order for each question in same survey.

the seed must take account different question id : no ?

The questionnaire has not been launched yet, I was waiting for the bug to be solved to share it! So we can still modify it without any problem

Problem is that question IDs can change, for example when restoring a survey archive. You would need to store a different seed value for each question for each response, to have a reproducable randomization
I don't see any short-notice solution here.

Oh right, question code then (by default) ?
(i know question code can be updated)

Before : random is always different , by question but when reload too.

By question is good,
By reload is bad (really bad)

Now :
by question is bad (except if option)
by reload is good …

Maybe adding a option here ?

Currently : even set different subquetsion code didn't make random different …

I really this we need more option here

1. Randomized each time loaded (same than 3.X and before)
2. Randomized with param : can be
   1. Extra option (allowing different question to have same random)
   2. Subquestion code (concatenated ?) : same list have same order , maybe default ?
   3. Question code (then same order between reload but different than other)

And we need a SEED set when testing survey (this is an issue)

@Arthur__ : currently you need to use 3.X … randomized each time loaded.

I don't really understand what I have to do on my survey.. This is my first time using it, sorry �

@Arthur__ There is no current solution or workaround. We are discussing a future approach, but that will not be available on short notice, sorry.

@DenisChenu
I would suggest to have a METADATA field (instead of the current SEED field) in the response table where we store all this additional information in JSON format
THat would also be future proof for more metadata, if needed. It will just not be searchable, easily. So it would not be good for data that needs to be searched/index/queried.

Personally i like the extra options … and i think it's needed.

1. We need a 'randomized all time diffrent' like 2.6 until 3.X
2. We need a 'randomized by id' : same response get same order
3. We need a 'randomized like other question' (it's an old feature request)

I look at seed system … i don't see where we can add some var to existing SEED :(

@Arthur__ : about potential workaround :

If you host yourself : setup a 3.X version : LTS Release 3.28.13 on https://community.limesurvey.org/downloads/

If you hosted by limesurvey : i don't know, but i think you can ask to have a 3.X ?

Okay thanks for your help but the version can't be installed in the short term, so I'll share the questionnaire like that!