(Floe) Notes from chats with SMEs on UI options design
w/ Clayton (Sept. 30, 2010)
What are user preferences?
- Includes the kind of stuff you already see around (e.g., BBC's accessibility)
- Gives users the ability to specify high contrast presentations, big fonts, etc.
- These are relatively simple presentation attributes
- In the Access4All scheme, user preferences are settings that might be used to inform screen reader behavior
- Also provided an outline mode (i.e., table of contents)
- More interesting user preferences would include radical transformations of content
- E.g., deliberate suppression of content
- E.g., on the full web version of Amazon.com, you have hundreds of controls, especially below the fold, but on the iPhone version, you get a much simpler presentation
- You'd never end up using most of the controls on the full web version
- Would be good for users to have choice of presentation profile; e.g., use mobile presentation even if one isn't on a mobile device, simply for the ease of use
Users and reading
- A person's ability to read has more than one component to it, but a critical component is the ability to skim
- E.g., with Google search results, users tend to do a lot of rapid non-reading, and then cognitively zoom into something of interest to read
- A lot of people lack the ability to scan, and using the web is a completely different experience if you lack it; using the web becomes hardly possible.
- Most people aren't even aware that the content is covered with other stuff--you just reject things that you're not interested in
- People who can't skim: it's like being in a tar pit; you have to work through everything
Suppression of content
- We need to mark up the content to point out what's critical content
- Need to differentiate between critical and non-critical content
- With that, we could:
- Suppress non-critical content entirely, or
- Split it into multiple pages (e.g., first page for critical, second page for non-critical), or
- Introduce controls on non-critical content
- This means two things:
- Content providers/authors need to be able to assign content priority
- The presentation must be able to adapt to different content priority requirements
- This would also be really useful to automagically create mobile sites
- Heading structure isn't enough for prioritization of content
- For typical users, you want to put as much as you can on the same page, prevent splitting it and having to look for it
- For users who have difficulty reading/scanning, it's the opposite
Preferences of user preferences
- Users need to have more than one set of preferences because of:
- Situational differences
- Privacy concerns
- Anonymous profiles that aren't linked to traceable identity; but also preferences that need to be linked to identity if they want to bring it with them
Preference control
- In the field of low-vision, fine-grained control is necessary
- Different levels and options of colour control are especially important (some combinations are more readable to some than others)
- But supporting a portfolio of "starter profiles" is a great idea
- Suspect that some people won't do much tuning
w/ Gary (Sept. 30, 2010)
The idea behind UI options
- Represent user's preferences across different applications
- Make customizations on one, apply across different applications
- Context from which this arose: web portals; one overarching application, but pulls in different applications from outside; apply preferences to those applications too
- Allow more than just the ability to do "simple" things
- Influenced by BBC's display options, Transformable, and TILE
History of UI options
- Two major design iterations
- Driven by conversations at Fluid all-hands meeting and conversations with Jutta and Clayton
- Design was informed by Jutta, Clayton, all-hands meeting, and other best practices
- No users were interviews
- No user testing was done
- Why were tweaks chosen over presets?
- Started with presets, taking a cue from BBC's display options interface
- But too difficult to determine a good preset, or how many you would need; so reverted to just tweaks
- Users have to decipher the package of a preset--there's a cognitive burden of having to pick a package first
Design considerations
- Took care about how to present the options to users to make
- Problem is with how to categorize preferences and options, with plain language
- Also, some things overlap, so it's not easy to stick it in box or another
- Many controls are possible; need to balance between having enough controls to change the interface vs. having so many controls
- In the current design, only showed one type of control at a time, instead of an intimidating mass of controls
- But sometimes changes to one preference in one area impacted another
- Live preview was introduced to see the impact
- If you change preferences of the active interface, it might adversely affect usability of the current interface
- Why not preview on the page actual? Because controls take up too much space, and you can't see the original content
- But it's difficult to represent every kind of interface element in one preview window
- Text size is a popular adjustment
- Wanted to give users ability to freeform change, but not see the changes until saved/applied
- Should users be able to go backward in history? Be able to undo?
- Should a reset bring you to the initial point you came into, or to the previous setting?
- What is a default setting? Where/how is it set? Is it based on a theme, or user's own stylesheet setting, or an accepted standard?
- If a user can't read text below 32 pt, how can you use UI options to make changes?
- Struggled with level of detail of options (e.g., serif vs sans-serif; pt size; line height; etc.)
- Can we reduce the number of controls?
- In choosing a preset, and tweaking it, and choosing another, does the user lose everything they just tweaked?
- Would social presets be useful?
- No guarantee that the things the users pick is the best thing
- We're giving users the ability to make poor design decisions
- What if the user makes the interface unusable?
Where the designs left off
- More chats needed with SMEs
- User testing