Let's face it - Photoshop is a BEAST. It is a powerful tool, but has an outdated WIMP interface that is highly complex and riddled with toolbars, panels, and dropdowns. This makes the software notoriously difficult to master. Much of Adobe Research efforts are in making these tools more accessible to novices and easier to learn.

I began by analyzing around 40 "actions", which are sophisticated image effects created my experts. These actions record specific Photoshop commands making it easy for even novices to produce very sophisticated effects on their images. Our goal was to see if we could modularize these actions.

A taxonomy was established and includes:

Effects, Textures, Cutouts, Brushes, Tweaks

In an effort to better understand the mental model of novice users I conducted a short paper prototype study with 10 participants. In this experiment, the participants were asked to put modules together in a way that felt most logical to them in order to turn one image in to another image. We gained the following insights:

Insight 1: Most participants were able to correctly choose the modules necessary for the image effect. This implies our taxonomy is logical and easily understood by novices.

Insight 2: Most participants tended to think of image operations flows as being horizontal, left to right.

Insight 3: Participants who had used Photoshop before were more inclined to stack modules vertically like layers, but still thought of operations as being horizontal.

Insight 4: “Cutout” was a more quickly understood term for novices than “selection mask”

The established modular system mirrors Photoshop in that layers are stacked vertically and can be rearranged, while operations/effects are a horizontal stack.