Reflection 3
In this paper, the author developed a tool, IMAXES, that allows users to create visualizations of data in old and new ways by manipulating axes using a hand controller. The formation of the visualizations utilizes a set grammar that the authors developed which allows for deterministic and declarative outputs. They test their novel technology by doing a user study with a wine expert. He tested his general knowledge by creating and analyzing familiar visualizations. Then, he made new visualizations and learned of new possible correlations involving red and white win. Overall, the results were promising, and the functionality of the tool for a practical purpose was a success.
Because it is a fixed grammar, there is a limit on the possible types of visualizations, but it is impressive that they were able to create new emergent visualizations. It is also interesting how they have some preset low level dataset that they load to help ease users into the process. The paper describes the collection as being clustered together in a bookshelf-like manner. Furthermore, they can also combine and use these visualizations with each other (i.e. tree linking). One major challenge for users is learning how to interpret the results of these new types of visualizations and dealing with a higher load complexity. It would be a point of interest to have the manipulation of the axes involved actually be in the immersed environment. Regardless, there was no existing tool prior that allowed for such easy manipulation of axes, attributes, mixing, and outputting of various types of graphs.
For their evaluations, it would have been interesting if they could compare the rate at which their tool creates visualizations to other popular tools, or if they could compare how the emergent visualizations fair against more standard ones. There is not a lot of metrics used for an evaluation. The nature of the tool is exploratory which was consistent with the demonstration of their use case with the wine data. It appears quite risky to not have a usual evaluation of their tool and to mostly just show results based on their design decisions. From a paper perspective, it is important to understand the conference you are submitting to for example UIST versus CHI and knowing the qualities that the accepted papers generally have. Overall, this was an interesting read and innovative research in the relatively young field of virtual/augmented reality.