ion. While scientific visualization deals with things that are real, or that could be seen if they were, InfoVis usually does not. What does a bank account look like? How do you depict the results of a survey so that you can understand them and draw conclusions? The question of coming up with a visual representation from nothing in InfoVis demands a dif ferent approach than the quasi-photographic view of SciVis . Interdisciplinary work. Visualization is inherently interdisciplinary. People who mostly come from a computer science background try to produce images that are perceived easily, and that communicate data effectively to the user. We draw on perceptual psychology, cognition, illustration, visua l communication, etc. Yet this is hardly acknowledged especiall y in SciVis, while InfoVis researchers started conducting us er studies and basing their work on literature in other fields lo ng ago. The number of artists and other non-computer scientists is also much larger than in scientific visualization, w hich tends to scare them off with the level of technical knowledge that is required for entry. Women. Look around you, how many women do you see? If there are quite a few, you stumbled into InfoVis or an InfoVis session at Vis. If there are mostly men, you’re at Vis. Does this bother you? It certainly should, we need a lot more variety and color in our field, and InfoVis is doing a lot better here. Some of the points in the above list are connected, and most of them present strengths as well as weaknesses. We can certainly ac t as if InfoVis and SciVis were the same, and hope that by doing so the differences will simply disappear – but that is not going to h elp. In fact, we can profit much more from realizing and understand ing the differences, and trying to leverage them. Who says an Inf oVis paper can’t be extremely technical, or a SciVis technique ca n’t be documented from start to end and thoroughly evaluated? Like it or not, the historical division between InfoVis and S ciVis is here to stay. But this should be taken as an advantage, as so mething that can drive innovation, rather than an obstacle. 2.5 Communities, Divides and Bridges (Helwig Hauser) It is not in the nature of panel discussions to significantly c hange the course of history. If at all, panels can stimulate a broad er discussion amongst scientists about relevant issues. But this , at least, is very important. And the community divide between SciVis a nd InfoVis definitely is a relevant issue, and even though alrea dy previously, e.g., at Vis 2003 [5], panelists discussed this iss ue, it is still necessary to explicitly discuss what possibly divides or in tegrates SciVis and InfoVis. Thinking about it and discussing the iss ue, we can try to delineate what we just think from what actually is t rue. This, hopefully, then leads to a more informed (and a less emo tional) point of view. When being open to new points of view, when being ready to expand the horizon of our usual considerations, we can ident ify a lot of useful input from outside worlds [1]. Many InfoVis sol utions proof very useful when it comes to the visualization of multivariate, scientific data (see, for example, Fig. 1 for a InfoV is-SciVismixed analysis of scientific data). Increasingly, also resu lts from the SciVis world, such as the general-purpose use of GPUs, fo r example, or the use of semitransparency for visualization, in spire the InfoVis world. Also, we have to face the fact that research to pics change and evolve over time, both also in the fields of SciV is and InfoVis, and we are departing from known grounds and enter new ones [1]. SciVis data, for example, becomes more vers atile, i.e., multi-dimensional (e.g., time-dependent and 3 D), multivariate, multi-modal, multi-typed (e.g., also including d iscrete, categorical, nominal data, etc.), aso. InfoVis, on the other ha nd, faces new challenges when dealing with really large data, when goi g 3D, or when being confronted with dense data distributions w hich tightly relate to the scenario of continuous data as in SciVi s. Another point of view is the one of practitioners. For many of those who actually need visualization, it really makes no po int f whether something is SciVis or InfoVis – visualization solu tions are needed – and if they stem from different fields, that’s just fin e, but not really relevant (in the first place). In many application cases, we anyhow see that the visualization part only makes up a limi ted part of the complete solution – other aspects like data acqui sition and management, documentation and reporting, ergonomics o f the GUI, automation, etc., make up important parts, as well. Fro m this point of view, for example, it not really makes a whole lot of s en e to finely distinguish between SciVis and InfoVis. A similar p oint of view, for example, also is the one of funding organization s – it might pay off to sell our technology as Vis (not as SciVis or In foVis). The same also holds for application people. Do the li fe sciences need SciVis or InfoVis? They probably need both. Finally, there is yet another issue, which shows up when we ta ke a step backwards and consider the panel topic from a slightly larger distance: in addition to SciVis and InfoVis, there is a lot mo re which also is visualization, or at least tightly relates to it. Kno wledge visualization, for example, which much more focusses on present atio than on exploration and analysis, and cartographic/geogra phic visualization, which – in principle – has important aspects in co mmon with InfoVis (but with a strong and central relation to geogr aphical, i.e., spatial references), make up just two related fields (a mongst others more), which also should be considered. Accordingly , it is well possible that in future, we not only talk about SciVis an d I foVis, but aboutxVis, with x being a lot more, or, dropping the x, just about Vis.
[1]
Kwan-Liu Ma.
Visualizing Visualizations: User Interfaces for Managing and Exploring Scientific Visualization Data
,
2000,
IEEE Computer Graphics and Applications.
[2]
Jarke J. van Wijk,et al.
Smooth and efficient zooming and panning
,
2003,
IEEE Symposium on Information Visualization 2003 (IEEE Cat. No.03TH8714).
[3]
Matthew O. Ward,et al.
Information and scientific visualization: separate but equal or happy together at last
,
2003,
IEEE Visualization, 2003. VIS 2003..
[4]
Melanie Tory,et al.
Rethinking Visualization: A High-Level Taxonomy
,
2004,
IEEE Symposium on Information Visualization.
[5]
Melanie Tory,et al.
Rethinking Visualization: A High-Level Taxonomy
,
2004
.
[6]
Kwan-Liu Ma,et al.
Scientific Discovery through Advanced Visualization
,
2005
.
[7]
Helwig Hauser,et al.
Toward new grounds in visualization
,
2005,
COMG.
[8]
Gordon Erlebacher,et al.
Overview of Flow Visualization
,
2005,
The Visualization Handbook.
[9]
Jarke J. van Wijk,et al.
The value of visualization
,
2005,
VIS 05. IEEE Visualization, 2005..
[10]
T. J. Jankun-Kelly,et al.
Is There Science in Visualization ? Panel
,
2006
.