...
We consider 6 high-level tasks related to our problem, 3 essential and 3 non-essential. We list the three non-essential tasks as a reminder that they still need to be implemented, even if they are not significant for GR1.
Essential Tasks
Reading
Besides viewing the content, users also need to be able to organize the content (such as with tags or a directory structure), and control the lifetime of the content (marking items as read, marking as “read later,” etc)This task focuses on how the user traverses their content, and how they read it.
- Why is the task being done? The user wants to know what’s going on in their world.
- What does the user need to know or have before doing the task? That what they’re about to read is likely to be interesting/useful
- Where is the task performed? in a smart phone’s web browser, potentially on the go
- What is the environment like? Noisy, dirty, dangerous? potentially
- How often is the task performed? Within a session, repeatedly and rapidly, and they may have several sessions per day
- What are its time resource constraints? As fast as possible (as quickly as a few seconds), but perhaps lasting several minutes for longer content
- How is the task learned? Primarily by recognition. Most users are familiar with reading material over the Web, and this experience should carry over.
- What can go wrong? accidentally lose the item they just read and having to find it again
- Who else is involved in the task? no one
...
Users will often be interested in only a subset of information (written by a specific person or persons, or covering a certain topic), and thus want to run a filter over their content to identify this subset.
- Why is the task being done? The user wants to restrict the given information to that subset.
- What does the user need to know or have before doing the task? They need to be logged in, and know enough about the subset to describe it or pick a filter.
- Where is the task performed? Either on a desktop/laptop, or on a phone browser.
- What is the environment like? Noisy, dirty, dangerous? It could be, if the user is on the go with their smartphone or tablet.
- How often is the task performed? This depends on the user’s profile; some may do it all the time, others only once or twice per session.
- What are its time resource constraints? The feedback loop should ideally be very fast, comparable with Google Instant Search, so that refining the filter is easy.
- How is the task learned? The user can learn it by recognition; we plan to design the interface so its controls are very similar to Google Instant Search or other similar mechanisms.
- What can go wrong? The filter could be incorrect (too broad or narrow), or it could have a typo. A fast feedback loop will improve safety by ameliorating these issues.
- Who else is involved in the task? No one else in this case.’
- What subtasks are there? Navigating to the textbox (or other form element used for filtering, entering the author name or topic to keep (or reusing an old filter), scanning the subset displayed, and revising the input if needed. Optionally, users could save the filter for future usage.
Saving News Items
Users will need to be able to save organize the content (such as with tags or a directory structure), and control the lifetime of the content (marking items as read, marking as “read later,” etc). This also includes saving the information they find important as they read through updates in their news feeds. It should be noted that the term "save" may end up being misleading. The user's intent is to make it possible to revisit the item in question in the future. Saving it somewhere is one way of doing it, and how files are managed on the desktop. The exact implementation of this feature may be a less conventional model, such as tagging like what users can do with emails in Gmail to make them easily accessible later. Prototyping and testing will shed more light on which approach will be easiest for users to learn.
...