Wednesday, March 24, 2010

User Assistance Trends Panel

Rhonda Bracey, Nicky Bleiel, Teresa Goertz, Leah Gurn, Linda Urban

Joe: Definitely having conference next year, location still unknown. Most conferences I go to, most speakers are men. Our closing session is all women.

IT industry

Nicky: "Cube 2.0" greatest potential since Dilbert Doors, cube-land walls transformed into wrap-around computer, able to roll up and take home. Manipulate all applications with gestures. Specific gesture to open user assistance.

Teresa: Increasing use of mobile devices, for payments and savings, for medical assistance, for telemedicine, for micropayments.

Leah: Companies tying into more user-created information that are outside their control, higher noise to signal ratio, not necessarily a good thing, lots of illiterate rants and "me 2," IT will just throw up their hands--and this ship has already sailed.

Linda: When appliance not working, go up to it with e-book reader to access documentation and support, chips & Bluetooth in appliance or other device, direct link to support URL, location information and smart ads will supply addresses of local repair shops.

User assistance

Teresa: Increasing writer involvement with UI text and online help for mobile payment softare and mobile medical devices.

Leah: Embedded help will almost completely replace separate help, simplicity and ease-of-use the holy grail of user experience, much better understanding of user needs needs, scenarios, tasks part of development, help embedded in dialog box, window, website page.

Linda: User-generated content will increase UA jobs, new job titles and tasks (content curator, content weaver, content strategist), increased responsibility for understanding big picture around content, new skills to learn (user research, forum moderation, information architecture), must think beyond creating online help.

Nicky: Special IA bot will search web for all information (help, wikis, blogs, videos, news stories, etc.) on a product and organize it in easy-to-use display, search by version on request, can remove/rearrange items for customized display, metadata will make results relevant and updating easy.  "Do see a chance for "Clippy" to make a comeback, be our UA robot."

Tools and technologies

Leah: Increased use of XML and DITA will lead to on-the-fly translation options, companies will write tighter, cleaner online help that supports basic translation rules, online server-based computer translation will translate text-on-demand (cheap and fast) when users need topics. Useful especially for non-mainstream languages. Overall translation quality will go down, but make available in more places.

Linda: Authoring tool vendors will help write context-sensitive content at the point of need and mash ups to bring in needed information and links.

Nicky: Embedded live chat in the UA, specific to topic user is viewing, if users have to open help to get to chat, might get them to help,  if answer is in topic displayed, users forfeit right to chat again.

Teresa: Increasing awareness of SEO (search engine optimization) and incorporating principles into writing.

Turning Search Into Find

Matthew Ellison

Search is important way for users to find information in help.

Search is not the most effective tool for finding information, but it is the tool users prefer to use. Many help systems omit indexes.

We used to find things, now we search.

What are the obstacles to find? Don't know keywords, can't ask questions, too many results, no synonyms, and more. 

Google's predictive search as you type characters reminds of.... old-style Microsoft Help index. Search technology provides the best of using index in help. So used to predictive like in Google, when it doesn't happen elsewhere, think something;'s wrong.

Faceted search is kind of way of combining search and TOC.  Doesn't require hierarchy that TOC imposes. Classify information, using metadata, by specific characteristics (facets). Facets can take on values. Users explore information by choosing required facets. Combine facets to narrow search. Results presented in any order. Can be used to refine results of full-text search.

Factors that turn search into find

Stop words, exclude specific words from search matching, marginally reduces file size and increases search speed.

Facility to exclude topics from search, so topics appears only in specific contexts.

Search results synopses, show first few words or key extract from search result, enables user to assess relevance of search result.

Phrase matching, phrases in quotes, usually unnecessary.

Search filtering and faceted search, follows concept of information types, commonly used on e-commerce sites, achievable with concept keywords and search filters.

Ranking of search results, results determined by number of occurrence of keywords (indexing automatically promotes topic in results), metadata.

Metadata is the key to flexible and effective search, enables you to avoid zero search results, especially with synonyms.

Predictive search, reduces keystrokes, provides better "scent of information," discourages "long tail" searches.

Practical ways to implement predictive search

Google Custom Search, use PHP and jQuery to add autocomplete, turnkey autosuggest technologies such as PredictAd.

Interaction Design Meets Expertise: Representation, Understanding, and Problem-solving

Axel Roestar, University of Washington

How do we support people in awhat they do best?

Design plays a role, Designers initiate change in the future, identify means to bring the change about. Lies at the root of innovation. Innovation challenges people, confronts them with things they haven't seen before.

Design is not art.

Connecting research and design

Cyclical design model. Thnk about design as an activity that never ends and never begins. Start with observing something in the world (something doesn't work). Come up with need for design. Develop concept, shift from observing to exploring. Look at the world and abstract observations. A concept is a hypothesis. Build prototypes, validate it with prospective users. Iteratively improve design based on testing feedback.

Even when product shipped, design isn't finished. Products are too complex to have them completely fleshed out before they go out in the world.

Pools of expertise in the cycle include innovator, practitioner, technologist.  Requires ability to move both forward and backward in the design process. If you were not part of the observation process, a designer starts from a very bad position because you're working with someone else's data. Ethnographic studies are difficult, but important that designers are following ethnographers.

3 U's of design: usefulness, understandability, usability.

Information is not a scarce resource, but attention is.

The Psychology of User Interface Responsiveness

aka The Psychology of Time Perception in Software
Steven Seow, Microsoft

What is objective is not necessarily what is perceived by users. Many layers to consider when talking about performance and responsiveness.

WYSMNBWYG: What you see may not be what you get.

Central course for attention. When you pay attention to non-temporal events, you lose information about time. When you pay attention to time, you lose information about details. Attenuation hypothesis.

Perception,  what the brain does with the information it receives from the senses. Couple of layers before it hits the brain. Something that's objectively described can mean different things to different people in different contexts.

Anything we can do to reduce perceived duration is a good thing.

No specific spot in the brain to perceive time. No need to question if perception is reality. A clock tells time, a metric tells a number. No judgment involved.

Tolerance, the maximum degree of something we are willing to experience to attain something.

"The download took 5 minutes."

Yesterday, it took 3 minutes: too slow.

File only 1MB: too slow.

Used to take more than an hour: download is great!

Maister's first law of service: satisfaction is a function of discomfirmation, or the difference between what was perceived and what was expected.

Responsiveness: the ability of a system to respond to user input and process internal operations without undue delay. 

Responsiveness relative to interaction in question, subjectively perceived & interpreted, non-exclusive (users aren't carrying stopwatches).

Relate a metric not to an interaction, but to a user expectancy. When you get a new interaction, relate it to the user expectancy, not the metric.

Classes of responsiveness:
  • Instantaneous (< 0.1 - 0.2 sec.)
  • Immediate (< 0.5 1 sec.)
  • Continuous (< 2 - 5 sec.)
  • Captive (< 7 - 10 sec.)
Continuous is about flow. Applicable to UA in areas of error messages. If too much text, too much to read, users will scan. Aim for 2-5 seconds, at 5-6 words per seconds, and aim for that size in error messages.

Not a mathematical model, but a psychological, empirical model, based on what users expect.

Time is a precious commodity. Users shell out not only money, but time, and have an expectation for their experience.

Uncertainty is the biggest poison. Information is a powerful antidote. The right amount of content at the right time.

Check out the book, "Designing and Engineering Time"

Double Scoop Case Studies, Theme: Single Source

I was going to go to the eBook conversion session, but when I got to the room, there were no tables, and I did not want to spend an hour with my notebook computer on my lap. So I went to my second choice.

Converting from Multiple Formats to DITA-compliant XML
John Kinsky, Intel

Managing a diverse group of writers spread out at many sites. Many documents, converted thousands of pages to DITA-compliant XML, and was able to do the task successfully. Supported multiple OSs, multiple products, and in many cases, multiple languages.

Know what DITA promises. DITA assumes all information can be reduced to bite-size chunks. (When all you have is a hammer, everything looks like a nail.) Minimalism is the order of the day. And DITA might not work for you.

DITA Open Toolkit is a publishing environment, not a magic solution. Requires knowledge of XML, XSL, DTDs.

View from 50,000 ft:
  1. Analyze current document
  2. Determine topic types
  3. Restructure existing, legacy content (better to do in your existing tools; it's faster)
  4. Convert from legacy format, seciton by section and topic by topic
  5. Publish document from new sources (concurrent with doing the same with your old sources)
Study your document. Determine your legacy document type, topic based or book/chapter based. Topic-based easier if you've been writing online help.

Choose topic types. Concept, task, reference, generic. DITA guys will tell you you can't use the generic topic. But generic topic type solves problems, all other topic types derived from the generic topic type. Use a matrix tracking device (spreadsheet) and assign topic to topic type, plus time value it will take to convert.

Topic type analysis on only part of the overall process. Found a lot of theory, but little practical info for how to deal with conversion.  Output dictates the ground rules. Decide before you begin to convert. HTML requires different top-level DITA map structure than PDF. Multiple formats from single source requires additional work and testing.

Choose a path: easy or hard. Easy path, convert everything to concept or generic topic types. Hard path, follow recommendation from nearly everyone else, convert everything to appropriate topic type (which we should). Easy path gets you there quickly, but requires you go back at some time and restructure the content.

Middle path follows the easy path for the bulk of the conversion, follow the hard path for select group of topics.

Use specialization? Sooner or later have to answer the question. Need skillset (XML, XSL, DTDs) for "yes" but "no' restricts your flexibility (but allows you to use out-of-the-box topics.

Have you thought about all possible single-source conditions? Reduced special formatting variations? Decided on image types and sizes? restructured legacy content before conversion? Built a prototype or proof of concept?

Tame the beast. Have one person dedicated to the task. No matter what your conversion estimates are, it's going to take longer. Get help from others who have gone through it. Restructure content in original format before converting. Convert a few to[pics, build, view output. Repeat restructure and convert, over and over.

Voltiare" "The perfect is the enemy of the good." You have to know when good enough is good enough. Keep in mind why you chose to use DITA in the first place.

Developing Product Documentation in a Confluence Wiki
Bruce Michelson, Corda Technologies

Previously outsourced documentation, developed using FrameMaker to PDF and RoboHelp to FlashHelp. Directed to look into wiki.

Goals: collaborative authoring, better and stronger searching, avoid FlashHelp installation, improved documentation. hassle free and low cost production, support for translation.

Considerations: features, search engine and accessibility, export and import tools, backup and version control, security, maintenance and authoring.

Define access; groups; global, space, and page permission.

Developed wiki personas: viewer, contributor, translator, reviewer, publisher, administrator.

Customers, if they have an account, can create or edit pages.

Don't have global search and replace, like when product names change. 

Sense of loss of control by writers because anyone can contribute. Lots of fights with developers. Burned bridges by moving source to wiki. Bottleneck with reviewing.

Tuesday, March 23, 2010

Where User Experience and Software Engineering Meet

Andrew Ko

In method, I'm an HCI (human-computer interaction) researcher, bit in practice, I'm an SE (software engineering) researcher. In HCI, how do we get the right design, in SE, how do we get the design right. Question is how do we merge these.

Why is software evolution difficult?

From study at Microsoft, observed 25 hours of coding and bug fixing, took 357 pages of (handwritten) notes, logged 4231 events in a spreadsheet. each some piece of information a developer or manager looking for at the one time.

Looked also at reasons why people switched tasks, such as face-to-face meeting, email, all sorts of reasons. Found that work is fragmented. People are interrupted every 5-10 minutes. Only that amount of time to focus. Also, blocked about every 10 minutes. They need info that they can't get themselves.

22 information needs. 5 least often satisfied:
  • what code caused behavior (36%)
  • what code caused program state (615)
  • what is program suppose to do (15%)
  • in what situations does failure occur (41%)
  • why was code implemented this way (44%)
One reason why it's hard to fix bugs is because answering some of these questions requires getting people together for meetings to make decisions, meetings where stakeholders must be involved.

Software development is tacit. Plans and specifications are unwritten, developers have to communicate (a lot) to make progress. It takes time to coordinate communications. Can't make people coordinate better just with tools. Software quality depends a lot on quality of communication. Communication and cognition are inherently faulty and unreliable .

How can users help software evolution

Looked at bug reports (more than 500K) in the Mozilla community. Did quantitative analysis the characterize report trends, and qualitative analysis to explain resolution trends. About 45% or reports submitted by "reporters" the type of user that submits the majority of community bug reports, are duplicates. Most "reporter" reports are not fixed. Many also are "as designed" or not reproducible.

The percentage of bugs entered by "reporters" has gone down over time. Most "reporter" reports were tech support for power users' tinkering. They also rarely gave enough useful information to reproduce problems. Also entered problems that had already been solved.

Open bug reporting is useful, but lots of overhead to process bad reports, only a skewed subset of users report bugs, users who report bugs are bad at providing useful information, and text isn't precise enough to express useful information.

The challenge is to enable users to submit bugs that are precise, structured, aggregatable, with not training.

Microsoft Help Preview

Paul O'Rear

After agonizing about which session to go to--I was liking the session about gesture interfaces and thinking very seriously about the DITA case study (but Tony Self drags on when he speaks and the slides didn't look like they'd apply to my needs so much)--I decided to settle here.

With new release of Visual Studio, VS10 releasing new help system.

Looking at MS Help timeline.
  • 1990 with Windows 3.0, RTF source, .HLP files. 
  • 1997 brought HTMLHelp, HTML-based, with .CHM files. (The poster child of the famous Bill Gates memo to make everything "Internet.") 
  • 2002 was Help 2.x, HTML-based, .HXS files, but for Visual Studio, intent to replace HTMLHelp.
  • 2006, AP (assistance platform) Help, XML-based, .H1S files, "Longhorn" help, dependent on .NET platform, used in Windows Vista, MAML (Microsoft Assistance Markup Language)
  • 2010, Help Viewer, XHTML-based, .MSHC files, shipping with Visual Studio 2010.
Started in 2008 to try and create as simple as architecture as possible.

Help 2.x feedback was not good.

Goals for Help Viewer:
  • Help is quick and easy
  • View online and offline in familiar, browser-based experience
  • Ensure offline Help content relevantr and up-to-date (requires srtrong updating scenario and important componenet of architecture)
  • Get fast and relevant search results with both F1 and full text searches
  • Benefit from simple, standards-based architecture
Viewer is your default browser.

Help Library Manager generates runtime TOC, keyword index, search index for content cached locally.

Help Library Agent services ms-xhelp:/// protocol from applications & retrieves content from online or local based on user setting.

Help Viewer features

Contains along the left areas for search, a table of contents, related topics content, and an index. On the right is content. In Visual Studio, right content area has sections, some which have tabs. VS help shipping without index, and causing controversy. Had to cut because of time, resources, and user data suggested people using Google, but beta user feedback suggested needed index.

Classic view, lightweight view, and script-free view for online MSDN content.

Help Library Manager
When first installing content, user choose where to install, becomes default. Link to install content from online.

Don't have to have help content on distribution CD, but can install from online.

Content hierarchy start with products, then books, packages, and topics.

Two primary modes to install content: interactive and silent. with interactive, customer selection gets installed, and with silent, the installer decides. Latter useful for IT. Silent "remove content' supported.

Interactive security supports MSHC or digitally signed cabinet files. Silent install requires signed cabinet files and is not intended as an end-user scenario.

File types
  • MSHC, Microsoft Help Container, XHTML fies compiled into a Zip file, renamed for association with help viewer
  • MSHI, Microsoft Help Index, index fragment to expedite content merging
  • MSHA, Microsoftr Help Assets, manifest of MSHC files, used as  a setup definition file
  • CAB, Cabinet files, standard file for combining files and digitally signing authentication
Content requirements
  • XHTML 1.0, allows for greater separation of content from rendering
  • Must leverage a variety of MSHC-specific attributes to enable TOC, multiple language support, F1 support, etc.
  • Topics must contain attributes for ID, catalog disambiguities, word breaker to use when indexing, title, branding support
  • Optional attributes for TOC, F1, keywords, category, description
One of the differences from past, no separate files for TOC, index, etc., but all info is coded in the XHTML topic file.