Estimating the size of software and the effort to be spent to develop it is very important for the accurate planning and calculation of software projects. At the beginning of the software development process, size, effort and cost estimations are the most important problems that developers and administrators face.
One common approach in agile software development is the so-called "planning poker" (see for instance http://www.mountaingoatsoftware.com/agile/planning-poker). Collect a couple of experienced specialists in a topic (something like 5 persons) an make them silently (!) make their guess/estimation per feature in pre-defined, "quantized" time units. Common are units of 1,2,5,10,20,40,100 days. everything longer than 100 days is "don't know how long", meaning normally the boundary conditions and requirements are not known well enough. One common practice is to use poker cards having these numbers imprinted so the "players" have to pick one silently and are not allowed to change their pick when revealed. If the numbers of the participants are all the same or in a narrow range, everything is fine, then take it. If the spread is large, let the lowest and the highest estimator discuss why they picked what they picked to get an idea of their reasoning.
When doing expert-estimation never forget to make sure that time is planned / estimated for integration and testing. This can easily double the time planned for pure software development. As well do not forget things like requirement analysis + UI design (upfront), customer interaction and prototype presentation and internationalization if needed. All these things are likely to be forgotten if you ask software development experts / programmers only.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software