I never got to meet Peter Kollock in person, but my decision to pursue sociology was influenced by his sharp work on social dilemmas and virtual communities, the latter being something few sociologists were thinking of at the time. He was a role model for me in the Mertonian sense; I’m just at the beginning stages of a sociology career, and Peter has been an example of what to aspire to.
In particular, his 1998 Annual Review of Sociology paper on social dilemmas was huge for me. It’s a clear discussion of many kinds of collective action problems, and demonstrated an approach to the study of cooperation and competition that I preferred immensely to the more formal and abstract treatments I’d read by economists. I still go back to this paper all the time.
Often with his former student Marc Smith, Kollock showed why virtual communities are amenable to sociological study (Marc posted a touching memorial on his blog, connectedaction). Kollock and Smith’s 1996 chapter  on Usenet presciently tackled bandwidth as a common good, as well as socialization, monitoring and sanctioning in virtual groups. Kollock also did some of the earliest work on eBay; , looking at how reputation works in a world with near-infinite exchange partners and few channels for sharing social information.
Kollock was early in working on some of the most important problems people studying the internet are facing now. His work helped us start to understand reputation, trust and exchange in a world where exchange partners are many, instead of few, weak and sparse instead of strong and dense, and distant instead of close; and his contributions will be missed.
 in Susan Herring’s edited volume, Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural Perspectives.
 in Ed Lawler, et al.’s 1999 edited volume, Advances in Group Processes, vol. 16.