Last week I had the pleasure of attending the 30th annual COCOMO Forum in Arlington, VA. As always it was an informative and enjoyable exchange of software and system estimating topics and challenges with some of the brightest practitioners in the industry. Topics ranged from technical debt and software maintenance, early phase effort and schedule estimation, COSMIC function points and various data analytics topics. In addition to the presentations there were several lively workshop panels discussing and debating future directions for the COCOMO and CoSysmo models (contemplating the 3.0 versions of each of these models.
While all the talks were quite good there were two that I found particularly enlightening. Bob Stoddard presented his research on “Causal Modeling of Observational Costs Data: A Ground Breaking use of Directed Acyclic Graphs.” Based on the work of Judea Pearl and others, Bob informed the audience about new and emerging technology with the promise to significantly improve analyst’s ability to understand cost data in ways that traditional correlation does not facilitate. For more information on this technology – check out this Carnegie Mellon website ( www.phil.cmu.edu/tetrad/). Another talk that was quite intriguing was presented by Stevens Institute of technology graduate student Ye Yang. Ye discussed her research into Crowdsourcing. Crowdsourcing is an emerging area of software engineering that relies on calls for participation to accept tasks for any task in the software development process (https://en.wikipedia.org/wiki/Crowdsourcing_software_development) Check out this link for more information on crowdsourcing. With the emergence of cloud technologies this model for software development has been gaining popularity in industry and with government agencies (https://en.wikipedia.org/wiki/Government_crowdsourcing#Department_of_State). She presented her findings on several research questions around pricing of crowdsourced tasks and the potential motivating factors that would entice one into competing for these tasks.
This is not the first time I have attended this conference, hopefully it will not be my last. Although attendance is not huge, the mix of software and system cost gurus with the bright and enthusiastic graduate students working with them, is always impressive. I continue to learn valuable lessons about the cost estimating challenges in the field and on the horizon. Though the proceedings are not yet available they should be soon @ http://csse.usc.edu/new/events/event/30th-cocomo.