Citation Context

...One might expect the same in software engineering, but it seems the biggest area ofsvariability in practice today is the extent to which models are used. France and Rumpsprovide an excellent overview =-=[10]-=- of the current state of the art in modeling.sWe present the results of a survey of 113 software practitioners. The objectivessinclude understanding: the perceived difficulties of model-centric, or co...

Citation Context

...y on vast volumes of sourcescode, rarely looking at diagrams or models, let alone creating or editing them. This isstypified by the open-source community where file patches are the key unit exchangeds=-=[11]-=-. If modeling is so great, why are these people not doing it more?sThere has been a small amount of research into why modeling is not that prevalent.sAfonso et al [1] point out that “there is little p...

Citation Context

...results have not been encouraging. Arisholm et al. [5] concludedsthat the costs to maintain UML documentation in the software they studied onlysbalanced its benefits. This is echoed by Agerwal et al, =-=[3]-=-, [2] who conductedsusability experiments of modeling tools. Sjoeberg et al [14] provide a comprehensivessurvey of experiments, some of which relate to modeling.sReasons for the lack of adoption of mo...

Citation Context

...guidelines are the biggest concerns with UML.sSeveral researchers have conducted controlled experiments to measure the benefitssof modeling; but the results have not been encouraging. Arisholm et al. =-=[5]-=- concludedsthat the costs to maintain UML documentation in the software they studied onlysbalanced its benefits. This is echoed by Agerwal et al, [3], [2] who conductedsusability experiments of modeli...

Citation Context

...ts have not been encouraging. Arisholm et al. [5] concludedsthat the costs to maintain UML documentation in the software they studied onlysbalanced its benefits. This is echoed by Agerwal et al, [3], =-=[2]-=- who conductedsusability experiments of modeling tools. Sjoeberg et al [14] provide a comprehensivessurvey of experiments, some of which relate to modeling.sReasons for the lack of adoption of modelin...

Citation Context

...ective, and lesssadopted. These include a belief that it is just about “pretty pictures”, notsunderstanding the underlying paradigm well enough, and lack of attention tosconsistent style. Anda et al. =-=[4]-=- studied modeling in a safety-critical context, where itswould seem particularly appropriate. They reported that modeling yielded positivesresults, but the tool inadequacies and training costs were im...

Citation Context

...tches are the key unit exchangeds[11]. If modeling is so great, why are these people not doing it more?sThere has been a small amount of research into why modeling is not that prevalent.sAfonso et al =-=[1]-=- point out that “there is little practical evidence of the impact” ofsmodel-driven development in general software engineering. Berenbach et al [6]sdescribe common modeling problems that lead to model...

Citation Context

...nto why modeling is not that prevalent.sAfonso et al [1] point out that “there is little practical evidence of the impact” ofsmodel-driven development in general software engineering. Berenbach et al =-=[6]-=-sdescribe common modeling problems that lead to models being less effective, and lesssadopted. These include a belief that it is just about “pretty pictures”, notsunderstanding the underlying paradigm...

Citation Context

...s questions. Participants were able tosignore these questions if they did not have substantial experience with modeling tools.sA technical report of the questions and complete results is available at =-=[8]-=-.s3. DemographicssParticipants averaged 14 years experience (80% > 5 years, 20% > 20 years). Aboutstwo thirds were from Canada or the USA. The rest of the world was fairly wellsrepresented with partic...

Citation Context

... to maintain UML documentation in the software they studied onlysbalanced its benefits. This is echoed by Agerwal et al, [3], [2] who conductedsusability experiments of modeling tools. Sjoeberg et al =-=[14]-=- provide a comprehensivessurvey of experiments, some of which relate to modeling.sReasons for the lack of adoption of modeling seem to boil down to: a) weaknessessin tools, modeling languages and proc...