Why Psychopy Coder and Not builder?

posted on 1 Apr 2024

Here, I post my points, ideas, and also some interesting news, including papers and research. You may like them!

I want to say "I like Psychopy" because it is Python-based and has many useful features. Psychopy this is just the sweet side of the story, and there is another part,

Problems with Psychopy in general?
First of all, running Psychopy on Python 3.8 and specifically using wxpython, which doesn't run on Python> 3.8, is a problem. 

Why not builder?

I wouldn't say I like using a builder for the whole experiment. I use Builder to design faster, continuing creating some stimuli or specific points, but the whole code is better in the code environment for so many reasons:

• Data obtained by code is much easier to handle,
• and the number of libraries psycho-py builder imports is much more than what you may need to run
• the builder system creates a lot of data that is mostly not used in any analysis

To point out an example. I did a whole experiment with all I needed in 765 lines of code, including comments, alternative options as a command, and many more, and if you just save a builder with a space io, you have a minimum of 500 lines of code! a load of packages and system load, especially if you use eye tracker and EEG at the same time, causes multiple crashes for no reason, like freezing; 

Once I went through the log file, I could see that irrelevant library conflicts with eye-tracker, especially for someone like me whose RT and also sync of EEG trigger is important (remember we are talking about milliseconds, even 10 ms). This causes problems.
And finally, control over sections is making it harder.

And there are a lot more to discuss but I hope these are enough point toward using Coder not builder,