Randomizer issue

Thanks for your report - I'm afraid we can't reproduce this behavior. Values generated by Next variant jump correctly in a random order. Here is the video of what we observe, this is the intended behvior:

https://www.dropbox.com/s/ql2qus5hifss...r.mp4?dl=0

And here is the filter used in the video:
https://www.dropbox.com/s/xqfcvcfhb6pn...ffxml?dl=0

If your experience is different, please send us your video and the steps to reproduce it.

Actually no, the issue is still present in FF9. Here is a video showing the problem:

https://www.dropbox.com/s/76nxatk8q787...1.mp4?dl=0

Notice how the IntSliders always move down a step after setting randomization from high to low and then getting stuck there. There's something weird going on, I'll do some more testing.

Thanks emme, it would be impossible to comprehend without the video. Forwarded to the developers.

Thanks GMM.

I've noticed inconsistent behavior with IntSliders, where they sometimes keep oscillating between two adjacent values and sometimes move one step down and then get stuck there. Perhaps this is by design and just some weird logic, but still smells a bit buggy. If I figure out how to replicate this consistently, I'll let you know.

Sounds like on low randomization an IntSlider with 4 steps should move within +/- 0.4 range and always round back to the original value, right? That's the behavior I would expect.

Good to hear you'll be working on this.

Thanks for the upd ate. The randomizer logic works more consistently now. That being said, I still don't think this design is optimal for user experience.

When the randomization is se t to low, I would expect a relatively low change in the control values. Currently a 2-step IntSlider will change between its values even with low randomization. That's a high relative change.

I suggest the relative difference should always be within the randomization percentage range. So, on low setting (+/- 10% change) IntSliders with less than 5 steps should never change (given they would round back to the old value). The basic logic (for relative values in the range of 0.0-1.0) would ideally be something like:

new_value = current_value + random(-1,1) * randomization_percentage

Absolute values would then be derived from this.

I don't know if this makes sense - I'll have to give it more thought later. Anyway, I appreciate the update. Happy new year!