So I recently ran a course that allowed parallel tracks of the Galaxy visual programming environment vs a Jupyter Notebook with predominantly Python.
We had participants fill out a SUS survey about their experience.
Galaxy scored a 7.2 while notebooks scored a 6.7 from 55 ratings, which I thought was shockingly close, given that the idea of Galaxy was to be significantly more usable.
A very, very common refrain from these new users (all undergraduates) was that they were overwhelmed by all the choices/parameters in tools, and that they wanted to learn the details on every single one in order to understand what they were doing. We reiterated numerous times that this was not important to do, that they should only note the parameters we note for them, but nevertheless this was a recurring theme.
Comparatively, notebooks/code literally only showed the parameters of import (anything not left as default theyâd have to actually check the docs for), so visually, it looks a lot simpler, and ultimately the users donât even realise what they donât know because it isnât staring them in the face.
Therefore, I propose the following solution - all non-essential parameters get hidden across all tools, UNLESS a user has selected a general âshow allâ button somewhere in their user preferences (or still better, a giant button on their main page that then essentially lets us distinguish advanced from beginner users, as I suspect this âsimplificationâ strategy may want to be employed else where. *Note that to enable this, the selection of ânonessentialâ parameters (or more likely, âessential parametersâ) needs to be easy to do (even for non-tool developers), and prioritised for tools used in training rather than across the board. It could/should be a n âopt-inâ situation.
Then we can ease beginners in, without overwhelming them with choice.
Also open to alternatives, but this data robust and something we should seriously action. We canât be that close to a notebook.