Tasks

  • Die Hauptkritikpunkte der Reviewer waren die Generalisierbarkeit des Ansatzes und gewisse Designentscheidungen, die nur halbherzig motiviert wurden. Besonders das erste ist ein interessanter Punkt und eine Sache die wir leicht reparieren können.

  • *Generalization*:

    • Zur Generalisierung würde ich nochmal einen Absatz schreiben (vermutlich als weiteren Unterpunkt zu Kapitel 6), der versucht, Listening Histories in den generellen Kontext von Lifelogging einzusortieren und v.a. auch sagt, für welche Arten von Daten der LastHistory-Ansatz noch funktionieren könnte bzw. was man ändern müsste. Alternative wäre diese Besprechung mehr in den generellen Text zu integrieren. Meinungen? Ideen?

    • [DB: added section 6 (LastHistory Beyond Music) addressing that]

  • Quer durch die Reviews werden verschiedene Designentscheidungen bemängelt:
  • *Y-Axis*: "I was thinking that it might be nice to allow the user to put different variables on the y-axis for some added flexibility" (R1)

    • [FS: I think the 2D time-based visualization is one of the strongest points of the project and fundamentally enables the visualization of individual listening history entries and thereby of time-based patterns. Maybe this should be emphasized more? Also, I am not sure how "different variables on the y-axis" could work - what would the visualization look like?]

    • [DB: mentioned that in passing in section 6. TODO: write more about that in section 4.2.3 ("focus on time")]

    • [DB: fixed with the new 4.2.3?]

    • [DB: definitely fixed it with the discussion in chapter 7]

  • *Rectangular zoom*: "Another design feature I thought about was some kind of limited zooming capability to a rectangular region chosen by the user [DB: similar to Hochheiser/Shneiderman Timebox widgets 2004]. That might help in examining densely packed regions." (R1)

    • [FS: In my experience with 3-4 minute songs, a 15" or larger display using the full screen height nicely fits a full day without overlapping nodes, which was the reason why a vertical zoom/scrolling was not included (see also Diplomarbeit section 3.5.1). It might still add some value in terms of visually focusing on a specific area, so it could be mentioned in future work maybe?]

    • [DB: 4.3, fixed]

  • *Zoom+Pan*: "Zooming and panning are described, but not in detail - 1d or 2d?" (R2)

    • [FS: Maybe click+drag should be mentioned for panning?]

  • *Arcs*: "The curved lines to other appearances of a song look cool, but I'm not quite sure how helpful that really is. Perhaps another visual presentation of that information might be more effective." (R1)

    • [FS: The arcs enable the user to distinguish the individual lines connecting two sequences even in the case where the sequences are close to each other horizontally, because the lines never go into nodes from above or below the nodes (which would make multiple lines overlap), but always from the side. Maybe this should be mentioned as a design decision.]

    • [DB: added that. fixed?]

    • [DB: re-read the paragraph and we give a lot of reasons for using arcs. fixed.]

  • *Personal Mode*: "The distinction between personal and analytic mode seems a bit artificial. Why not have those features as individually selectable operations?" (R2)

    • [DB: We should make clear that personal and analytic mode are only shortcuts for several settings. Maybe argue with the casual users thing?]

    • [FS: Sounds good]

    • [DB: added section 4.2.1 on that, so, fixed]

  • *Scalability*: "Effectiveness of the approach when large amounts of data is involved. ... I can imagine the timeline-based representation would get quite busy and become incomprehensible with a large data set. ... " (R3) [DB: R3 writes in even more detail about it, basically wants to turn the thing into a visual analytics paper]

    • [DB: added a discussion about that to section 7 (future work). fixed.]

  • Pattern finding: "It is unclear how similar events are displayed? For example, how could a user go about to find out what genre of music s/he likes during the holiday season? Would the user need to first find all corresponding photos/calendar information? ... If there are multiple time spans, is there an easy way for users to see the comparison of listening histories for the multiple selected time periods?" (R3)
    • [DB: Discuss restrictions of this approach and the lack of automatic input - future work? I'd also argue with the whole: the main information is locked in the user's brain approach]
    • [FS: Not sure I understand the question. The application enables searching for time periods (e.g. "every June") and highlights them for the user to compare. The colors visually indicate the predominate genre for each time period. Does that not solve the reviewer’s question?]

  • *User Study*:

    • Ein interessanter Punkt ist dass wir in Section 2.3 mögliche Aufgaben und Muster für die Daten definieren, die in der User Study aber nur noch teilweise auftauchen. Also sollte man mal schauen, ob sich ein Mapping zwischen den beiden herstellen lässt.

    • "I am a bit disappointed about the case study especially the type of tasks performed by users in the study ... What about the support of tasks mentioned in section 2.3?" (R3)

    • "Can the authors provide some cases which are no-intuitive and not easy to be found by the method?" (R4)

  • *Online Study*: Außerdem müssen die aktuellen Ergebnisse der Onlinestudy noch eingearbeitet werden.

  • *Formatting*: "The authors have no indentation on the paragraphs and I believe that there should be based on the TVCG formatting guidelines." (R1)

    • [DB: fixed]

  • Outlook: "What if the visualization results are presented to others who are not the owner of the listen [sic!] histories? Is there any difference in the performance?" (R4)
    • [DB: TODO: at the beginning of the user study section, repeat again that we develop for end users and that we haven't tried it with external analysts (though it might be possible)]
  • Wording: Typos und komische Formulierungen sind natürlich immer drin, also sollte man auch nochmal drübergehen und danach suchen.

  • *Intro*: First paragraph is weird - Sebastian: put it in the end of the introduction, sell it as a contribution (we provide an answer to these questions!) and pick it up again in the conclusion

    • [DB: fixed]

  • *Figure 2:* Make text bigger (hard to read right now)

  • User gender [Sebastian]: Use 's/he' throughout the whole paper.

    • I did not change it, when the text is about a certain user from the study.

  • Text widows [Sebastian]: Delete single words in new lines. (on Monday)

  • User Study [Sebastian]: Read over it again and check structure and wordings.

  • Further references: Add Lee Byron and Martin Wattenbergs Stacked Graphs

-- DominikusBaur - 07 Jun 2010

Changelog

  • 2010/06/12:
    • Fixed formatting with latest vgtc template

  • 2010/06/16:
    • Added reference to Sellen + Whittaker CACM 2010, put analysis and personal mode in contrast to their five Rs. Also mentioned them as reference for visual images work good for cuing (actually: Conway, Autobiographical Memory 1990, but that's one more reference frown )
    • Replaced some 'memory triggering' with 'cuing' (more appropriate psychological term) in section 2.1

  • 2010/06/21:
    • Wrote section 6 ("LastHistory beyond music"), as an extension and generalization of the approach
    • Added explanation to "Online Reception" -> "Setup", as to why we didn't log user activity but used a questionnaire instead
    • Changed title of section 2.1. from "General overview" to "General aspects"
    • Added explanation why genres have their downsides to section 2.1
    • Rewrote "Task and Patterns" -> "Analysis"
    • Added 4.2.1 "Analysis and personal modes" as an introduction to the concept of two sets of options and to explain the overlap between them.
    • Added references to section 6 (sanderson1999, ayers1995)

  • 2010/06/24:
    • Added a paragraph to section 4.2.3 (Focus on time) discussing other values for the y-axis

  • 2010/06/25:
    • Rewrote first part of 4.3 about zoom+pan. More detailed info on zooming, click+drag to pan. Also discussed adding rectangular marquees for zooming in on two dimension.
    • Minor changes to 4.3 -> "Sequences" regarding arcs, should be fine now
    • Added part about different mappings for the y-axis also to chapter 7 (paragraph about sequences)
    • Added a new paragraph to 7 (future work) about the integration of automatic analysis (addressing especially reviewer 3).
    • Enlarged figure 5 (Personal Mode) as the song circles mentioned in the caption were not visible.
    • Added overlay capabilites for different filter results as possible future work (section 7).
    • Put first paragraph of intro to intro's end and rewrote it a bit (too much?)
    • Added Sebastian's comments (mostly typos, phrasing)

  • 2010/06/26:
    • Sebastian: fixed figure 2
    • Updated numbers and results in the online study section
    • Removed the weird sentence about brain extensions and cognition amplifiers from the introduction
    • Rewrote and restructured section 5.1 -> "Results" according to table 1

  • 2010/06/28:
    • Wrote summary of changes
    • Changed width of figures 3,4,5 to 0.9 (instead of 1.0)
Topic revision: r23 - 30 Jun 2010, DominikusBaur
 
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Medieninformatik-Wiki? Send feedback