- JK7409-001[1].jpg (139.41 KiB) Viewed 2148 times
Unless you're some kind of masochist, dealing with software "help desks" online or on the phone is not one of life's pleasures. Unless you have such a simple problem you could probably solve it yourself, you're quite likely to have your query written off as unanswerably idioscyncratic - leaving you in limbo.
To the rescue, perhaps, comes Andrew Ko of the Information School at the University of Washington, who has worked out a way of what he calls "mining the whining" of help desk users to reveal the frequency and prevalence of problems with software - the idea being that techs can then fix the seemingly idiosyncratic bugs they didn't know about.
Ko's plan is to extract, aggregate and organise key words from online help requests using software called Frictionary, which ranks re-occurring phrases by frequency to see what problems are the most prevalent. Ko presented his idea at the Conference on Human Factors in Computing Systems held in Austin, Texas last month.
To test Frictionary, he trained it on 90,000 support requests for the Firefox open source browser - and then asked Firefox's principal designer what he thought about its findings. Frictionary's ability to show bug changes over time was useful, he said, particularly regarding Firefox's ability to install or update as versions changed. He thinks Frictionary will also help open-source communities prioritise work, rather than merely focus on what interests them.
Development of Frictionary continues. In the meantime, why not check out these true help desk stories over at Silicon Glen.