1. Seems many questions in the paper are really about the modeling issue in SV technology, which is a necessary condition for a successful viz system. I think it is reasonable to try investigating and simplifying human mental models of the real world.
2. It would be nice that a SV system can provide functionalities for model/representation manipulation, such as transformation, simplification, etc.
1.Are there any examples of visualizations tools which attempt to use "cognitive structures" for the basis of design, and if so, have they been evaluated?
2. Are there some visual elements that we can assume everyone understands or can be taught, e.g. linked list diagrams?
1. In regards to convention, should we really have some kind of standards like ANSI to C? Is it necessary good? Like it has been indicated in this paper, people visualize differently. Importance is to have the right mapping between the context and the representation. Perhaps we can have a set of conventions in individual domains?
2. For novices, it seems that they learn more or better if they program the visualization themselves. Should we concentrate in designing tools to facilitate that, instead of designing the animations to try to communicate the ideas to them?
3. I am not sure about the arguments for taking "a version to bed". Nowadays there are computers everywhere (a laptop can be conveniently placed in your reach while you are in bed). We often think of great ideas while we are taking a shower or even in our dreams (which we often forget when we wake up). Perhaps we can start looking into extracting images from our brains instead of concentrating on designing visualization? :) (Maybe attaching a digital storage unit in our brains to record such images and download to computers whenever we want to. :)
1. Providing software visualization tools originated from the experts' imagery can be very helpful for people who are not experts themselves and even for communications between experts. But how surely do those abstractions or represent ations from research reflect the real pattern? Or maybe the imagery obtained is somehow reconstructured version?
2. From cognitive aspect, to determine what kind of and how many different kinds of representations to be used in a visualization tool is very important and always a paintaking step. Will it also be very essential to work on presenting those things in a meaningful way?
1. Have software visualizations failed to focus on usability and interactivity, which could be a reason there hasn't been overwhelming succeess of them thus far? 2. Has the appeal for software visualizations thus far been mostly because graphics are more interesting to look at than text?
1. Does the use of multiple views in visualization answer the problem of conflicting conventions in visualization systems? 2. While it may not be possible to capture the non-visual parts of the mental model constructed in the programming task, could visualization provide enough proper visual cues to inspire the completion of the model?
1. I am surprised there is not more mention of Object-Oriented techniques in this chapter. There is a fairly standard notation, Unified Modeling Language (UML), for OO analysis and design. Such syntax enables general visualizations of a wide variety of problem domains. Indeed, the advantage of OO is that it is a more natural of programming. (Recall the endorsement our guest lecturer gave for the the "Gang of Four" patterns book) 2. In the discussion on "Is SV a way into 'the expert mind'..." (page 457), item #1 discusses "set aside" information and #3 "external memory." These seem to be the same thing. What is the distinction here?