The evolution of modeling research challenges

Abstract

In 2007 ICSE hosted a track called “Future of SoftwareDevelopment”(FOSD).Wewereinvitedtowriteandpresenta paper on the future of modeling for the track. The result-ing paper [1] described the state of modeling research, iden-tified some major challenges and proposed a research roadmap.Partsofthisroadmaparecurrentlybeingexplored,andprogresshasbeenmadeinaddressingsomeofthechallengeswe identified. However, there is still significant research “tobedone”withrespecttothechallengesoutlinedinthatpaper.Itisnotourintenttodiscusstheprogressthecommunityhasmadewithrespecttotheroadmapinthiseditorial(ourapolo-gies for deflating expectations in this regard; an editorial issimplynottheplaceforsuchdiscussions).Rather,wewouldliketousethiseditorialtostimulatediscussionsaroundsomeof the challenges that have arisen since we wrote that paper.Before we get into identifying specific challenges, letus first take a step back and ask “What is the grand chal-lenge addressed by researchers in the SoSyM community?”The answer has typically been “To significantly reduce thetime, cost, and effort required to develop complex software-intensive systems that meet stringent quality requirementsthrough use of models that are fit-for-purpose.” This is stillthe focus of many research programs, but we are also see-ing a broadening of the grand challenge to include not onlydevelopment problems, but problems that occur throughoutthe lifetime of a complex system, from conception to retire-ment. For example, in the FOSD paper we made referenceto emerging work on the use of models to manage software