In the last class of CGT512 (a HCI and UX class I’m taking this semester), when we discuss about the history of GUI, we briefly mentioned the Memex microfilm viewer described by Vannevar Bush in 1945. Memex reminds us that human brain works with associations and links rather than alphabetical orders or clearly defined categories. I’ve been having some interests in tagging, categorization, and information retrieval. I’ve also encountered some categorization problems in my life and research recently:
1. I moved to a new place recently–an exciting house! Okay, during the process of packing stuff, I tried to be very very organized. I put summer clothes, winter clothes, spring/fall clothes, shoes, kitchen stuffs, bath stuffs, etc. separately in different boxes, and marked them. But in the end, I lost my patience to categorize every tiny bit of things into the exact categories. I just couldn’t find a proper category for them or they belong to several different categories. Or I found something belong to a certain category but that box was already sealed. I also need to consider whether they belong to things that I will use quite often recently or can leave in the basement for long, but this sometimes conflicts with the previous categories I have set. I had such a headache trying to be organized and make my life easier looking for things when settle down at the new place, but there are still things I cannot remember where exactly I packed until now that I’ve been in the new house for almost three weeks. What I kept thinking during the process was that there gonna be a better way to organize stuff! According to the Memex philosophy, maybe I shouldn’t categorize the stuffs based on their same properties, I should organize them based on associations, but how? It seems the modern grocery stores are also thinking about this problem of how to organize their shelves to increase profit, such as the famous example of putting pampers and beer together.
2. I’ve been working on a project analyzing some Twitter posts by engineering students. I retrieved all the tweets under a Twitter hashtag #engineeringProblems. I want to find some common rules in these students’ vocabulary so I can retrieve more relevant tweets in the future, but I couldn’t put these vocabulary under clearly separated categories using the topic modeling algorithms. The same words or phrases jump around several different categories. Should I manually force them into clearly separate categories? Or should I just study keywords co-occurrence (association)? Or should I do both? How language is developed and organized?
I went back to re-read one of Dr.V’s blog posts written almost a year ago on tagging and information categorization. My understanding is that human brain works based on associations, and human brain also desires for control and clear overall pictures. I become very much intrigued about what a superb task management and information retrieval system interface will look like enabling both associations in human brain and also give the sense of control and clear categorization ?
Cindy Chen's Blog 
I organize my house based on activity – especially the kitchen. The placement of objects should support, as much as possible, the flow of activity. So you can reach for what you need easily when you need it. Retrieval is only one goal of categorization, but not the only one.