INFO 202

Assignment 4 - Faceted Classification - Feedback

Author(s):
Bob Glushko
glushko@ischool.berkeley.edu

Course: INFO 202
Date: 15 October 2008
Title: Assignment 4 - Faceted Classification - Feedback
Summary feedback about the assignment. You'll get an email with a grade and specific comments about your work.

Summary

Overall you did a good job on an assignment that like the previous one was more challenging than it might have seemed at first.

Most of you handled quite well the key challenge of "making sure the system met your own needs while choosing facets that anyone could use without any help from you."

Some people chose facets that at first glance were too idiosyncratic or clever to meet the "use by someone else without explanation" test. For example, there were facets like "Right-handed or left-handed tool," "Number of letters in the common name for the tool," and "Can be taken on an airplane." But in each case an explanation was provided to justify the choice of facet and we ended up thinking that the design was reasonable.

Problem Areas

A couple of you misinterpreted the instructions and tried to force fit the new items into your original facet scheme, and that usually didn't work.

A few of you didn't follow the instructions to describe your specific experiences with the Facet Map software.

A few of you didn't explain the changes you made when you got the new items, or you said something perfunctory like "I had to add a new facet" without any justification or detail.

Overview of Grading

I prepared a very precise grading checklist for the TAs to use, and since they are very intelligent and disciplined folks they were able to apply it with almost perfect consistency.

The grading for this assignment is on a 10-point scale, and the average grade was 7.9. There two two perfect 10 grades, and I've asked to post one of them so that you can see that you can get a top grade by doing clean and elegant work and without having to write a PhD dissertation on facet design.

2. Designing the Classification -- Grading Instructions

We put strong emphasis on coverage, scaleability, and not being idiosyncratic, We gave up to 5 points here for the system of facets. We used the checklist of dimensions from the lecture here, so your feedback will typically say things like "not orthogonal" or "too idiosyncratic." Most people got 3 or 4 here.

4. Encoding in Facetmap

You got 1 point, for using Facetmap. Everyone got Facetmap to work, even though in some cases the XML you showed us didn't validate. Either Facetmap is lenient about this or you're more clever than we are.

You got 1 point for answering the question about how you dealt with the occurrence exclusivity restriction. Given that everyone got Facetmap to work, it was disappointing to see that some of you just blew off this question. FOLLOW THE INSTRUCTIONS. WE NUMBER ALL PARTS OF THE ASSIGNMENTS SO THAT YOU CAN LINE UP ANSWERS WITH THE QUESTIONS. IF YOU DON'T TO THIS ON THE MIDTERM AND FINAL EXAM WHERE THE PARTS OF QUESTIONS ARE BROKEN OUT, YOU WILL BE VERY SORRY. Please excuse the caps lock key but I don't know how else to get your attention.

The third part of this question asked you to evaluate how well your original system worked, and to explain any changes you needed to make when you got the new instances. This part of each assignment is designed to let you show us that you're learning the big ideas embodied in the assignment. So again, while a few of you got 3s here because you took the time to reflect deeply on the assignment (which doesn't imply that you wrote a lot of words), some of you clearly know a lot more than you are telling us. Please tell us more when we ask you what you are thinking, ok?