A common task when trying to understand pieces of architecture inside a site is to spot, document and depict changes over time. This task pulls together different actors, with different agendas, and covers a rather heterogeneous set of challenges. Ever since XIXth pioneering works like Minard's thematic cartography, or Marey's graphic method, we are entitled to believe that depicting dynamics of change requires specific means - in other words that the analysis of changes implies rethinking traditional visual tools like cartography, sections, 3D graphics (etc.). In the "information age" too, with the emergent "visual analytics" research field, innovative graphic solutions are put to the fore that renew our capacity to analyse and make decision on spatio-temporal data sets.However when talking specifically about architecture, focus is still mainly put on representing a state ("my building in 1605"), rather than changes - may the model be virtual or tangible. Naturally tangible models may to some appear as "communication old timer". And it is true that the development of computer applications (GIS, CAD, virtual reality, the internet) and related technologies (LIDAR survey, 3D laser scanning ), have driven most actors to adapt their work methodologies. Seemingly, a time came when actors stopped thinking "what can you do with a virtual model that you can't already do with a tangible model?"Yet, our claim is that, with mature computer technologies and related devices (touch screens, smartphones, etc.), with records of successes and failures, it might be time today to re-think this over calmly. In this contribution we intend to turn the question around: "what can we do with a tangible model that we can't already do with a virtual model"? In other words, we wish to demonstrate that maybe there are tasks - both communication and reasoning tasks - that are best tackled once freed from the screen as unique interface.We will introduce an experimental device, called "tangible chronology" developed in order to represent changes that occurred on Krakow's market square (24 artefacts, 64 successive changes) over a period of 750 years. The device combines a master board, tangible models of the artefacts, with a coding of their position on the board and in ordinal time, and a tangible timeline for each artefact with a coding of different types of changes, of durations, as well as of uncertainty (in the dating). Initially designed for the blind, our first proof-of-concept prototype was then extended to match the requirements of edutainment tasks in the context of museum animation.Naturally our aim is not to question the usefulness of computers: tangible models we present come out of a fully computer-operated design and production process. Furthermore, tangible interfaces have in fact become a hot research topic within the computer science discipline itself where limits of the now traditional HCI devices ("mouse/keyboard/screen") have long been identified.Backed up by experiment, our contribution will discuss in what tangible models could serve content holders or academics specifically in historic sciences, and in what their making there calls specific attention and methods.
