Outline of Class 24: Human Factors, Continued

Miscellaneous

• I've gone over your project proposals (except for one group that didn't tell me where theres was).
  • As mentioned earlier, detail was lacking in some. I've asked those groups to turn in new proposals.
  • One big problem was a limited sense of roles. I would have liked to know who is in charge of meetings, who is in charge of the joining of parts, who maintains the group code directory, and so on and so forth.
  • Another problem was a lack of weekly goals. I really would like to be able to turn to you each week and see how close to that week's goal you are.
  • At least one group impressed me by scheduling regular meetings and making it quite clear how they'd use the meeting.
• William got an eye infection yesterday (which is clearing up) and I seem to have caught something that is sapping my energy. What's the moral of this? I may not be around as much as I'd like over the next few days, and grading will continue to be delayed. Sorry about that.
• Read the selected chapters in Bentley for next week.
• There will be a Sunday night dinner this week. Let me know if you're interested. I'll also try to find another night for those of you who can't make Sunday nights.
• A number of you failed to complete assignment nine in time. I'd like explanations after class. Those of you who did complete the assignment generally did a very nice job, including finding an IS manager in Santa Rosa, a High-School Webmaster in Minnesota, and more.

Project Clio, Continued

• In the previous class, we began to discuss the design of a web analysis system, and came up with lists of suggested users and suggested features.
• Between then and now, you were asked to interview some potential client about such a system and to record their responses. There were 13 surveys recorded.
• Today, we'll go over responses and see where that leads us.

Initial Design and Survey Results

Suggested users

• System administrators, who care about load, use, and abuse
• Corporations, who care about whether or not their advertisements are being seen, and what types of pages are most frequently used (help pages, advertisements, corporate rosters, ...)
• The government, who cares about where users are and who may want to keep tabs on people and groups.
• Web designers, who want a better sense of how people actually use their web site.
• Individual users, who want to know whether the effort they've spent on their pages is worth it
• Institutions, who wonder what public perception their pages bring (which they hope to determine by seeing what types of pages are accessed)

• Were there other types of people interviewed?
  • IS Manager
  • Director of Minnesota Arts High School
  • Student imitating company

• In the interviews, we also considered computer proficiency and web proficiency
• Web proficiency
  • 1: 1
  • 2: 1
  • 3: 5
  • 4: 5
  • 5: 2
• Computer proficiency
  • 1: 0
  • 2: 1
  • 3: 5
  • 4: 4
  • 5: 4

Basic Features

• In class, we developed a number of types of features and design components, and then looked at some specifics of each. The general categories were counting basic information, methods of summarizing or limiting information, methods of presenting results, other issues.

Counting or identifying values

• What values might users want to count or identify?
  • Hits on each page or group of pages (12)
  • Hits from each machine or group of machines (10)
  • Amount of data sent (5)
  • Entries to system (9)
  • Exits from system (7)
  • Paths (7)
• Other suggestions included
  • Number of errors
  • Browser types
  • Source computer types

Reporting

• It was clear that some reports will want to limit or group pieces of data (e.g., put all the outlines for CS223 in one category).
  • Time-based (6)
  • Machine-based (1)
  • Page-based (7)
  • Otherwise group data (6)
• Other suggestions inclueded
  • Hit based. For example, you might want to limit reports to pages with more then 10 hits.

Presentation

• There are clearly a number of ways to present information. Here are some we came up with
  • Numerically (11)
  • Bar chart (7)
  • Web with different size/color nodes (2)
  • Animated (4)
• Other suggestions included
  • A continuous report of what's being hit (in a small window?)
  • Output in Excel/spreadsheet format for further analysis

Other issues

• We also considered a number of issues that are hard to classify.
  • Will the system permit saving and loading of data? (7)
  • Can we compare results across time? (5)
    • Different versions of the same site (0)
    • Same pages, different times (7)
    • Site-to-site comparison (1)
• Other suggestions included
  • Support different access levels.
  • Permit web-based viewing of data, so that outsiders can learn about hits to our site (with access levels)$.
  • Count loads that include images vs. loads that don't include images.
  • Too much detail in our system.
  • Time spent on each page.
  • Load over time specifically mentioned.
  • Can run on slower computers.
  • Easy to intall and use.

Process

• What, if anything, did you learn from interviewing people?
• How did you interview them?
• How much time did interviews take?

Project Clio, Revisited

Tasks for Project Clio, revisited

• Let's now revisit our list of tasks, creating new ones as we go, and thinking about how each type of interface might support these tasks.
  • "Tell me about the top ten pages for each of the past twelve months."
  • "Do most of the hits on those pages come from inside or outside of my system?"
  • "What percentage of my viewers load images?"
  • ...

Extending Project Clio

• Now that we have collected some potential features, designs, and uses for Project Clio, we'll need to begin to extend our design from features to interface.
• Our first question might be "How do we present the features in Project Clio"? We'll discuss this in class, but here are a few possibilities:
  • As a standard windows+menus system.
  • As a "little language"
  • ...
• A related question will be "how do we allow our users to specify which subset to group, select, or eliminate?"

Analysis

• Once we've come up with a general list, we'll have individual groups sketch different interfaces and describe how we'd support the various tasks listed above.

Implementation

• Another question: what additional support do we need, and how can we get it?
  • For "amount of time spent on a page"
  • For "type of machine" or "type of browser"
  • For "path through site"
  • For entry and exit?
  • ...
• Solutions?
  • Extend server to produce better logs
  • Find server that produces better logs
  • Update pages
  • ...