Re:You can Estimate a Software Engineering Project (Score 1)

Maybe projects should be defined in the number of parameters necessary to define it. If you have a bridge or buiding, the CAD diagram should give you a good estimate of the number involved. Now consider a small SW project. the number of parameter is a measured by the number of variables. The complexity could be measured by the number of branches. The coupling factor is the extent to which functionality can be broken down to reusable modules. Basically, my point is that designing a software program should be weighted based on the difficult involved. Most bridge designers would refuse to give you a quote unless he were to spend a good deal of up front time to figure it out, unless he has access data of the same exact conditions. Yet, programmers are expect to come up with estimates in a couple of days, in a world where technology changes faster than Italian fashion. (Consider that in less than a decade, Unix has just given way to Windows, GUI's and event-driven programming popped up, Object-oriented programming has taken off, the internet has become a platform to support, C gave way to C++, gave way to Java, and might be replaced by C#.) That's like assuming that designing a bridge from wood 100 years ago is identical to designing one from steel and concrete now.