| Author | Message | ||
     IT-Architect Dipl.-Ing. (M.E.) Muhammet Oeztuerk |
As we know customer requirements for a software product are to be categorized in the following group: * Functional * Technical * Quality We also know we are able to measure the functional dimension of a software product & software project with FPA. The result oft function points count are comparable worldwide because of the standard and strictness of the function points rules. But there is no standard method to measure the another two dimensions (technically, quality). This two dimensions are still part of the customer requirements, and therefore contribute to the requirement size of the software and to the development cost oft the software . We try to size this two dimension with different methods. Some of us uses GSC or another uses the some cost factors of the COCOMO estimation method to catch up this two dimensions. Both non of the results of GSC or COCOMO are comparable! Indeed what we need are two methods with strict rules like FPA to measure this dimensions in order to size a software from all aspect. My first suggestion :-) about the names of methods : - TPA (Technical Point Analysis ) - QPA (Quality Point Analysis) IT-Architect Dipl.-Ing. (M.E.) Muhammet Oeztuerk Stuttgart/Germany | ||
| Luigi Buglione |
Hi Muhammet, after the F/Q/T requirement types classification proposed by IFPUG (http://www.ifpug.org/members/Framework.pdf), on section 2.4 there is also the suggestion to use the ISO/IEC 9126 Parts 2-3-4 metrics for the "Q" side, while there is an implicit suggestion to derive with a GQM approach (www.gqm.nl) the measures and metrics for the "T" side. So, probably there is no need to have a series of "xPA" methods, but only to establish the entities of interest and to choose the better size unit for your measurement goal. Another approach could be to size the project as a whole entity from a Project Management perspective and in this case a proposal could be PSU (Project Size Unit -- www.geocities.com/lbu_measure/psu/psu.htm), for early project sizing and estimation and that can be used also jointly with FPA. Estimation should be done comparing size and effort referred to the same entity: right now, usually regression analysis are done comparing the project functional size with the overall project effort, not only that one referred to the functional side, producing a certain amount of error. But it's a long discussion on this point... Best regards, Luigi Buglione, Ph.D. SEMQ: http://www.geocities.com/lbu_measure Luigi Buglione http://www.geocities.com/lbu_measure |
