Estimation of Effort and Schedule for a project with almost no familiar elements

Question

To the best of my understanding the schedule for the project is derived using the effort estimates for the individual tasks (in person hours), and the number of relevant resources used for the particular task.

As discussed in this question, we are starting a project using a new platform, featuring a brand new API and feature-set. How do we estimate the schedule of the tasks involved in the project, knowing only the top level task breakdown? Also, if some efforts are allocated to the tasks, the estimations would have a wide probability distribution. What is the best way to communicate the same to management?

Answer 1

I usually give an R&D (Research and Development) estimate for unfamiliar territory. In this, I promise an actual estimate upon the elapsing of the R&D time limit. I spend the R&D time creating prototypes and proof of concepts, and then finally my overall architecture. By that time, you should be familiar enough to make a good estimate. Giving an estimate without communicating that you will have to spend time in R&D is asking for upset employers. They may not like this response, but I promise you they will hate the alternative-- i.e. you make a timeline that you have no idea that you can meet and you inevitably fail to meet it.