Task Landscapes
Traditional software design decomposes work into task hierarchies — goals into sub-goals, sub-goals into steps, steps into actions. Neat, logical, and perfectly suited to linear, well-structured tasks. For complex professional work, it's a fantasy.
Barbara Mirel offers a better unit of design: the task landscape1 — a configurable terrain of activities, decision points, and paths that users traverse differently depending on situational conditions, emerging insights, and professional judgement. There's no single correct path through a task landscape. Instead, there are regions, landmarks, choice points, and well-worn trails that professionals navigate with a mix of pattern recognition and improvisation.
Mainline Tasks vs. Enabling Tasks
The most important distinction in task landscape design is between mainline tasks and enabling tasks.1
Mainline tasks are the core analytical or creative operations: filtering data, applying effects, modelling geometry, compositing layers. These directly produce the work output, and they're what design teams typically model and prioritise.
Enabling tasks are the supporting operations that make mainline tasks possible: annotating views, bookmarking states, comparing versions, writing notes, seeding the workspace with reminders, verifying data integrity, managing file organisation. In practice, these consume enormous effort and are essential to expert performance — yet they're systematically overlooked in design specifications.
If your design supports "filter, sort, and select" but not "remember why I filtered this way," it will fail complex users.
For every mainline task in your software, ask: What does the user need to remember, compare, annotate, or verify to make this operation meaningful? Those are your missing enabling tasks.
Choice Points
A choice point is a moment where the user faces multiple legitimate next steps, and the right choice depends on domain knowledge, situational factors, and emerging findings. A 3D artist might refine topology before texturing, or skip ahead to rough textures to test the look first. A data analyst might chase a suspicious outlier or park it to focus on the main trend.
Software can't make these decisions — but it can support them:
- Make available options visible (feed-forward)
- Provide enough information to choose wisely (context)
- Allow the user to change course later without penalty (provisionality)
- Preserve the state of the unchosen path for later return (bookmarking)
Cross-Pattern Traversal
Users of complex software don't stay within a single workflow pattern.1 They dynamically cross between patterns of inquiry, returning to an earlier analysis with new insights gained from a different investigation.
A video editor reviewing colour grades (Pattern A) notices an audio sync issue and switches to audio editing (Pattern B). After fixing the sync, they return to colour grading — but now with awareness that a particular shot needs re-timing, something they discovered during the audio work. The return to Pattern A is informed by Pattern B.
This is the norm, not the exception. Software must support smooth travel within and across task landscapes:
- State preservation when switching contexts
- Easy re-entry to previous working states
- Visible markers showing what's been done in each area
- No penalty for interrupting one pattern to pursue another
Mapping Your Software's Task Landscapes
When designing or evaluating IA for complex software:
- Observe real users on real problems. Watch where they backtrack, switch context, reach for external tools, or pause to orient themselves. These moments reveal the shape of the task landscape.
- Identify mainline tasks, then identify the enabling tasks users perform around them.
- Map the choice points — where do users face decisions with multiple legitimate paths? What information do they need at those moments?
- Trace cross-pattern traversal — when users switch between different types of work, what state needs preserving? What context do they need to resume?
- Audit for gaps — where do users leave the software to accomplish something (notes apps, spreadsheets, physical paper)? Those gaps are missing enabling tasks in your landscape.