psychopy.org | Reference | Downloads | Github

Advice on identifying a Development versus 'Stable' release of PsychoPy

#1

Hi Developers of PsychoPy,

Can you advise how we identify a stable version of PsychoPy that can be made available to staff or students in a university for teaching or use in final year projects.

Other open source products use the words development and stable e.g. Datavyu and GPower.

Our institution moved to version 1.85.2 over the summer for use in this academic year and we are affected by the pyglet bug (some demos don’t run and a couple of experiments that had been created for teaching). We are unsure whether to move to 1.85.4 or to revert to the version we used last academic year 1.84.1.

Our heaviest usage is in teaching is in S2 so we have an opportunity to make a change at the end of S1 but value your advice on identifying a stable release of PsychoPy now and in future.

Thanks in advance for your help and advice (and all the effort that has been put in to developing PsychoPy).

SRob

#2

Sorry to hear that. We’ve been had some problems with Pyglet over a few versions. (In the version in 1.84.1 they had a different bug where text objects gradually consumed memory and could lead to a crash halfway through a text-heavy study). In this particular instance I would recommend 1.85.4 over 1.84.x

The general advice is that we don’t add new features to minor increments so upgrading within one of those is less likely to introduce new problems. New versions where the major version number changes (1.NN.x) typically involve new features and, sometimes, refactoring and are quite prone to introduce new problems. So the general advice is:

  • Avoid 1.NN.0 releases on production machines (.0 releases are the first in a bug-fix series)
  • Try to upgrade between studies (your summer upgrade plan is what we aim for in my dept too)
  • Be wary of updates where the major number changes

I’m afraid that advice wouldn’t have saved you from the pyglet bug where large text didn’t render properly on windows. That was something that we didn’t fix in the first releases of 1.85 series (took me a while to track it down).

best wishes,
Jon

1 Like