The Magical Number 7+-2

July 25, 2007

From time to time I witness the wide gap between theory and practice in the field of interaction design and information architecture. I often witness this when simpletons (those who take a non-theoretical and non-scholarly approach to everything in life because they always think everything is black and white) attempt to practice what should be left to those who are better educated. This often comes in the form of misapplying good research findings to areas of practice for which it was never intended.

Ask any information architect or interaction designer and they will undoubtedly tell you that they have heard of “the magical number 7 +-2” principle. However, despite its wide use and popularity, I’ll bet that few would be able to tell you the origins of it. The idea comes directly from a 1956 research study by the cognitive scientist George Miller. In his study, he suggests that the capacity of short-term memory is within the range of 5-8 items. Miller was primarily writing about short-term memory span for digits, but an item could include something more complex such as a long word. He also introduced the idea of “chunking” which means we are able to group similar items together to form one item (such as a string of letters to form one word).

Other studies have been conducted since then that suggest that the capacity for short-term memory is much less. Also other studies suggest that the capacity may vary depending on the type of information (visual, spatial, phonological). What also complicates matters are issues of interference and attention.

The reason why applying such research to the practice of design and information architecture could be dangerous is because there are so many confounding variables in a real-world environment. This is, of course, assuming that the principles were applied correctly. However, the findings of Miller’s research study is often misapplied, as is apparent when some designers and information architects say things like . . . “because people can only hold 5-9 items in short-term memory, I decided that there should only be 7 menu items on this website, or 7 icons, or 7 bulleted items in a category, or 7 subdivisions on the homepage . . . ” and on and on and on. What does Miller’s study have to do with the number of tabs on the top of a web page? Such practices are not the implications of his study! Please spread the word.

Source:  Yvonne Rogers, “New theoretical approaches for human-computer interaction.”

Menus in Minutes: A recipe info. system for home cook enthusiasts

May 30, 2007

“Menus in Minutes” was the name of my group’s final project in “Systems Analysis and Design” in Winter 2007. Tonight I had intended to write a postscript that includes my personal reflections, comments received, and new insights I have gained as to how our project could be improved if we had another 10 weeks to work on it.

However, I realized that I never posted a single blog entry, so in order to correct this, I will post a few entries that will cover the following topics:

  • What the project was and how it all came together into the form you can see (you can download the written report).
  • Comments we received from Professors Agre and Borgman, as well as comments received when interviewing for internships in April.
  • New insights I have gained from new classes and a considerable amount of reflection.

All of this is in retrospect of course, so my new found 20/20 vision will be mixed in with my group’s thinking at the time.

Let me know what you think!

“When is the pen mightier than the keyboard?”

April 4, 2007

I attended a lecture recently with this title given by Andreis Van Dam from Brown University. It was a demonstration lecture on computer software applications his research team is developing which uses a pen rather than a mouse or a keyboard to interact with the software. The demonstrations were amazing to say the least and I was awed and inspired by what I saw.

Dr. Van Dam said that pen-centric computing takes advantage of the pen which is more than a high resolution mouse. The pen interprets “digital ink” in an appropriate context for characters, symbols, gestures, etc. He said that the best example of this so far is how the Palm Pilot feature called “graffiti” that interprets users’ gestures for alphanumeric characters. He also said that pen-centric computing is best if it is used with speech and other modes.

One application he showed us was one in which you can write music. You use a pen on something like a tablet PC and you make simple gestures to indicate the notes you want. So how do users know which gestures to make? Well, the software is designed so that the gestures correspond to what they are already familiar with in that given context. One thing that was suprising was that users are only willing to learn 12 new gestures: they are not willing to spend alot of time learning all the gestures, especially if the gestures are unintuitive.

I will write more entries on this lecture at a later time.

Searching iTunes

February 24, 2007

I have had iTunes for more than a year, and it is the main software I use to buy and organize my music.  My information science courses have begun to make me think about how “information packages” are indexed.  In other words, what parts make up the information package, what are those parts named, and what level of granularity?

From the interface, I can ascertain that the songs are indexed by name, time, artist, album, price, popularity, and genre.  These elements/fields and the content in them can be though of as metadata, and together, these parts desribe the entire song.  However, users are not able to search all of these fields.  The search box does not allow users to limit their searches to song name, or album, or anything else.  Users are forced to type something into their search box, and whala, all of these things pop up.  The results you get could be because your search term is part of a song title, an album, or whatever.  Either way, it is all mixed up.  Try searching for “beetles” and you’ll see what I mean.

Besides allowing users to search the fields that iTunes already indexes, it should also index the lyrics of songs (just the keywords), so that users can search by lyrics.  How many times have you known the lyrics of the song you wanted, but couldn’t remember the title?  The access points to song records should not exclude keywords in the lyrics.

Code vs. User Interface, part 2

October 4, 2006

I feel that I was too harsh on user interface in my last entry, so here I will trumpet the benefits of user interface over code. One benefit in particular: the “intellisense.” Intellisense (I think that’s the correct term), is shen the program guesses what the user is trying to do. You will often see intellisense when typing in the html code and a menu of choices appears. For example, when I inserted the “style” keyword into the anchor tag, a menu of all the available styles for the anchor tags appeared. this made it easier for me to remember all the possible styles available for the anchor tag. Without Dreamweaver or a user interface, I would have to have all the possiblities memorized (or at least, a good reference book).

Code vs. User Interface

September 29, 2006

I have finally formed an opinion on whether it’s easier to use a graphical user interface (GUI) or to just hard code something. The short answer: if you have the know-how, it is usally easier to just hard-code. I ran into this issue when creating this website. Technically, I created it in Dreamweaver, but I didn’t use the GUI that much. I mostly used (X)HTML to create these pages. In fact, when I first started Dreamweaver to create my website, I didn’t even know how to begin. Do I just type on the screen? What do these buttons do? What menues and sub-menus do I go into to create style sheets? It was very confusing. The Dreamweaver GUI is so overwhelming because there are buttons, menus, and icons to click on. There are also many textboxes in which you input information. So immediately I gave up and started coding everything.