Yesterday night, there was this terrible urge to redo one of my old prototype software projects which I had done some times back. Then I thought, let's do it with a twist this time. Let me try to capture the journey and share it here on this site. Let me try to pen the ecstasy. Coming to tangible talk - the business driver originated one afternoon when I was engaged in one of my heated discussions with some my close friends (Project Managers, professionally) about the concept of capturing time. My argument was that 'the actual time' spent was not being captured and without proper data how can we quantitatively manage and optimize on time management? As with all degenerative, coffee break discussions the talk diverged quickly into 'How do you define accurate time?', 'The effort to capture time to such granularity can't be expended'. The discussion did not end conclusively (None of my discussions with managers ever have, btw). I spent three uncomfortable nights brooding on the topic of quantitative time management. I spent the next two days hiding in a corner and writing code. This is what I came up with on the fourth day.

This fragmented chart looks like a DNA analysis, but infact is the Gantt chart for my time spent on activities on 05-Jun-2006. See how you can see that I came into the office, fired up my eMail application and scooted for a smoke :). Since I had the data captured in an analyzable format, I could project many different views of my time usage. Some of the important things one can clearly deduce is that I kept hopping between tasks, hardly able to spend more than 45 minutes on any task - a highly fragmented existence and oh yes, it was one of those 14 hour days. Productivity factor = 56%. Number of times hopping was done between jobs = 39 !!. The revelations kept flowing in.

Let's stop analyzing my work pattern. The only feeling of accomplishment was that now I had a tool which could show me my time statistics and the best part was it cost me around 15 seconds or less to capture my daily statistics.

Although I had the tool, I was never satisfied with the code or the design. It still haunts me. It was done in too much of a hurry and I would love to redo it again and this time relishing the feeling of walking towards the goal - I have tasted the goal and it's the journey I want to relive now.. and take you with me on the lovely ride. I hope you will enjoy. Lookout for the 'TimeTracker' sub section on the softwares section. I will be adding the tab soon.

Time for the punch line - I don't hate processes, I don't hate CMMI, I think they are beautiful and can help make the world a better place. I hate the way we try to interprete and implement them. Quantitative prediction is not about capturing data but rather being able to quantitatively predict the outcome of the process by employing statistical means on the captured data. Ever tried extracting quantitative data from a bunch of MS Excel timesheets (and employing statistical means)??