myGrid User Interview Methodology

To gather user requirements for myGrid we have used a very simple technique. We use semi-structured informal interviews. An interviewer has a series of questions and uses these to form an interview with a user. It is not like a questionairre - the questions merely act as a guide to the interviewer. He or she is free to let the interview follow digressions. this means what is gathered may well not be consistent between interviewees, but this is not a statistical survey!

Different interviews were devised for different user classes within the myGrid UserTaxonomy. This meant we could target the classes in which we had most interest.

• bioinformaticians
• Narrow bioinformatics users
• Occasional bioinformatics users
• Tool builders
• System Administrators
• Bioinformatics tool builders

The interview sessions were planned to take 45 minutes to one hour. The number of questions were restricted to keep within this time limit. Each interview was started with an explanation of the interview (particularly emphasising that it was not a test). Then myGrid was explained to a varying degree, depending on the user class. An example of myGrid explanatory blurb is available. Most interview questions started with explanations of the user's job and tasks; the main difficulties in the technical aspects of their job and the tools and resources they used. Many of the questions were orientated to myGrid servivces:

• Notification
• Provenence
• Personalisation
• Service discovery

The different classes of interview are available below:

• Occasional bioinformaticians
• Specialist bioinformaticians
• Tool builders
• System administrators

User interviews were recorded (with permission) and the interviews later transcribed. From these transcripts, features of interest were extracted by members of the User Group. These features were collected under emerging themes -- initially these were based on the myGrid services and the tasks of the particular user class.

-- AnilWipat - 05 Jan 2003