Comparison of SequoiaView & Grokker: Study of Hierarchical Visualization Systems for Files, Assignments of Computer Science

Download Comparison of SequoiaView & Grokker: Study of Hierarchical Visualization Systems for Files and more Assignments Computer Science in PDF only on Docsity!

Homework 4:

Examining Systems for Exploring Hierarchies

Rajeev Dayal CS 7450 February 24, 2004

Introduction

This report is a critique and comparison of the SequoiaView and Grokker visualization systems. In the first two sections, the focus is on the hierarchical visualization and navigation elements of both programs. The third section concentrates on how useful both programs are at visualizing information for tasks relating to file system administration. A discussion of the interfaces of SequoiaView and Grokker is the focus of the fifth section. In the sixth section, new insights that I gained about my file system by using SequoiaView and Grokker are mentioned. The final section describes Grokker’s visualizations of the Web and Amazon.com

Visualization of Hierarchical Structure

When dealing with a hierarchical information structure, users want to be able to see the hierarchy of information. In SequoiaView, the hierarchy of information is not completely clear to the user until they try to select an item. It is difficult to tell which larger squares encompass the smaller squares. One of the options in SequoiaView causes the entire path to be shown on selection of an element. All of the parent nodes up to the root are highlighted with different colors. The color scale varies from red (close to the leaf) to yellow (close to the root). Although this mechanism is helpful, it is still difficult to determine which squares encompass other squares. The user must pay close attention to the colors of the squares. With Grokker, there is no confusion as to which child element belongs to which parent element. This is a result of using spaces between nodes. The clarity of Grokker’s visualization makes it easy for users to visually assess the number of children that a parent has.

Users need to be able to tell what each element in the structure corresponds to. In SequoiaView, only the children are labeled. One nice thing about this is that the leaves of the hierarchical tree can be immediately browsed – the user does not have to drill down in the hierarchy to get to the leaf nodes. Unfortunately, the parent nodes, which are the squares that encompass the children, are not labeled. To determine the directory that an encompassing square maps to, one has to look at the child’s path and map each subdirectory to an encompassing square. This task can be somewhat difficult to perform. Grokker is different in that the leaf nodes are not immediately viewable unless there are some short paths in the tree. As a result, the user may have to drill down in order to find information about the leaf nodes. Two advantages of this system are that parent nodes are clearly labeled, and the user is aware of the path that they took to get to the leaf node.

Navigation of Hierarchical Structure

SequoiaView’s navigation system is interesting because it provides the user with immediate access to the leaf nodes. This is useful in cases where the user can see that there are outliers at lower levels of the hierarchical structure, and he/she wants to figure out what these outliers are. To provide the user with information about the hierarchical path to the child, the path of the item is displayed, along with the series of encompassing boxes. As mentioned above, it is not always easy to discern what the hierarchical path is

For visualizing file types, SequoiaView uses color. This scheme is effective because the attribute that we are comparing is not mapped to a number; we only want to determine where the majority of files of a particular type are clustered. By enabling color schemes and performing a sort by file extension, the clusters of files are immediately clear to the user. Because size is encoded by the size of the cluster, one can easily tell how much space files of a particular type take up. Grokker provides a filter that allows the user to filter by the application that the file can be opened with. This could be used to determine where files used by a particular application are located. However, a compound “Opens With” filter cannot be used. As a result, Grokker is unable to visualize the locations of multiple file types, which is something that SequoiaView is able to do.

To determine how many files are in a particular directory using SequoiaView, the user has to look at the number of squares in the encompassing square that represents the directory. If the encompassing square is large, it is easy to discern that a directory has many files. However, if the encompassing square is small, then it is not easy to tell that there are multiple smaller squares within this encompassing square. In Grokker, number of files/subdirectories within a directory is easy to visualize because the size of the circle represents the number of items it contains (regardless of size).

User Interface

This section compares the interfaces of Grokker and SequoiaView with respect to file interaction, interface response time, and the manner in which the visualization could be modified.

Interaction with Files

SequoiaView is integrated very well with the Windows system. By right clicking on a file, the user can choose to execute it. The program used to run the file depends on the association that Windows has for that file. Users can also right-click on a file and choose to open up a Windows Explorer window with the path to the file already exposed. The properties of a file can also be viewed, and the property window is exactly the same as that of the one used by Windows Explorer. Grokker provides the same functionality by actually providing users with a browser similar to Internet Explorer. Grokker’s method is nicer in some ways because when you open certain types of files, the contents can be displayed in the browser window; a new application does not pop up and distract the user. Grokker even renders images in the browser view. The result is a nice thumbnail gallery of pictures.

Response Time

SequoiaView’s interface is very quick to respond to user interaction. Selection update is fast, and there are only slight delays in applying filters, sorts, or color schemes. Grokker is not as responsive as SequoiaView. Selection is definitely slower, but part of this is due to the fact that Grokker uses animations. As the number of items increases, Grokker

becomes noticeably slower. In one instance, Grokker became completely unusable because it was trying to render the images in my picture directory. The number of pictures was too much from Grokker to handle, and the interaction time was painfully slow. Another problem with Grokker is the amount of time it takes to “Grok” all of my files. SequoiaView can visualize all of my files in one-third to one-half of the time it takes for Grokker to visualize 2000 files.

Modification of Visualization

In Grokker, the user is able to delete and move items. Deleting items is useful for removing items from the display that do not matter, and moving items is useful for performing side-by-side comparisons between two items. Grokker definitely needs an undo feature for these types of manipulations. It would be nice if SequoiaView had the ability to move and delete items as well.

New Insights

This section discusses new insights that I gained about my file system by using SequoiaView and Grokker.

SequoiaView

By enabling the color scheme, I was able to easily see where all of my MP3 files were located. One of them had a different color than all the rest. By moving the cursor over it, I discovered that I had an MP2 file among my MP3 files. I filtered my files by *.java, *.cpp, and *.c, I discovered that I have done much more C/C++ programming than Java programming in the courses that I have taken

Grokker

By looking the visualization of the Documents and Settings directory, I was able to tell that I have a large number of cookies and temp files. This indicated to me that I needed to clean out these directories. I used the “opens with” filter in Grokker and discovered that the UltraEdit program can be used to open the most files on my computer.

Grokker’s Visualization of the Web and Amazon.com

I attempted to visualize a Web Search for the phrase “SSL”. The results were very interesting. Grokker grouped the results into 11 categories. One would assume that the names of each of these categories corresponds to a keyword has a high frequency in documents that contain the keyword SSL. One of the categories returned by Grokker was “certificate”. This is a valid category, because certificates are closely related to the topic of SSL. Another category that was returned was “Secure”. This made sense, since SSL stands for Secure Sockets Layer. However, neither “Sockets” nor “Layer” was returned

This document was created with Win2PDF available at http://www.daneprairie.com. The unregistered version of Win2PDF is for evaluation or non-commercial use only.