This preview shows page 1. Sign up to view the full content.
Unformatted text preview: work package needs to be made. This is done by using the quantification of each task in each work package to estimate how many labor-hours are needed to complete that task. The sum of the labor-hours for each task in a work package is the number of labor-hours needed to complete the work package. Knowing the number of labor-hours needed to complete the work package (and the material and equipment needs), one can estimate the cost of the work package. This is called detailed cost estimating. The cost estimate for a work package takes into consideration the fact that different types or grades of labor on the work package may have different labor rates. The cost estimate also takes into consideration the cost of material and equipment needed to complete the work package. How this is done will be explained shortly. Many schemes have been proposed over the years to come up with estimates for a project, or even for all of the work packages, that avoids this level of detail in estimating. Of course, if a project is extremely similar to a previous project the same project team has already done, and if good historical records exist about the labor-hours and the costs of the previous project, then one could perhaps use this information to obtain a fairly accurate estimate of the labor-hours and the cost of the new project. In practice, this is rarely the case. Many of the schemes for avoiding detailed cost estimating are based on the use of cost models for projects. Companies that perform certain types of projects regularly can be expected to gain some expertise in estimating and managing these types Page 45 of projects. This often leads to the belief that the way to avoid the effort of detailed cost estimating and possibly even to get more accurate estimates is to collect statistics on these historical projects and to use them in some way to predict what the average costs will be on future projects of the same type. This works sometimes when there is little difference from one project to the next. But some advocates of this approach go so far as to propose using this historical statistical information to build a general estimating model that can be used to estimate the labor-hours and the cost of all future projects. For instance, B. Boehm's book Software Engineering Economics advocates using the COCOMO cost model for estimating software development project costs. These estimating models rarely give consistently good results. The models' authors often claim there is evidence that they do. But experience with some of these models and with trying to verify the claims of their advocates has convinced the author that most of the claims are greatly exaggerated. These models usually do not take proper account of the highly diverse levels of experience and capabilities of the individual project team members, and it is they who execute the work packages. These team members, on the other hand, often know their own capabilities better than those who employ the models. And if these team members take part in the estimation of the work packages for which they are going to be responsible, they often give estimates that realistically fit their capabilities and experiences. On...
View Full Document
- Spring '10