|
|
Ontological
Excavation:
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
IntroductionMicrosoft PowerPoint 2000 is primarily a tool for building and delivering presentations. While it does possess mechanisms that can be exploited to other ends (Bob Balzer gives a demo showing how he turned PowerPoint into a software development environment by using .COM listeners (Balzer, "Living with COTS", ICSE 2002), I hypothesized that PowerPoint will reveal an ontology that contains a primary cluster of core concepts related to presentations and slides and supporting concepts such as graphics and animations as smaller ontological clusters connected to it – characteristic of a Reef ontological structure. As such, PowerPoint should evidence high conceptual integrity both in the structural analysis of its ontology. This page contains the following:
Excavation ArtifactsThese are artifacts produced by Ontological Excavation. The XML versions were generated in Visio 2003 Professional and is probably optimized for Microsoft Internet Explorer. The morphology diagram for PowerPoint will probably not be readable even if you are on a Windows machine. However, if you are able to run the active scripts to get the search bar, you should be able to type in some key word - a label that can be found in the PowerPoint User Interface, for example - and the search will return all the objects in the graph that have that keyword. The abbreviations in the morphology refer to interface elements like buttons (B), Windows (W), Tabbed Panes (TP), List Items (LI), Radio Buttons (RB), Text Fields (TF), and so on.
Core Concepts of PowerPoint 2000The table below lists the core concepts of PowerPoint 2000 identified using ontological analysis techniques in order of structural importance::
Conceptual Integrity MeasuresI have been exploring metrics for measuring the conceptual integrity of computing applications. Thus far, I have identified two possible measures based on graph theory: conceptual coherence and conceptual complexity. I am also testing some combined calculations for the overall conceptual integrity. Conceptual Coherence - Conceptual coherence is a measure of an application's interrelatedness of its concepts, and uses average distance between nodes in a graph. The theory (explained in detail here) is that if a semantic network reflects potential data dependencies then a complete connected network contains concepts that are all interrelated and have an average distance of 1.0. The less related the concepts, the greater the average distance. Thus, the hypothesis is that removing those concepts essential to the application's domain model would make the resulting ontology less coherent, appearing as an increase in average distance. Conversely, removing peripheral concepts, not essential to the domain model, would make the resulting ontology more coherent, producing a decrease in average distance. Thus, conceptual coherence values reflect an ontology's "incoherence" where the higher the value, the more incoherent the ontology. Conceptual Complexity - An application's conceptual complexity reflects the average number of relationships per node (including attributes which are modeled as nodes in the ontology), and uses the average degree across all nodes in a graph (where a degree is simply the number of edges on a node). The theory (explained in detail here) is that a concept in a semantic network possessing many edges connecting it to its attributes or to other nodes has a high complexity versus a node with few edges. Thus, a complex concept is more likely to have interactions with many other concepts, raising the overall complexity of the ontology. For conceptual complexity, the hypothesis is that removing those nodes that help to simplify the ontology by organizing concepts will increase the average degree. Removing inherently complex concepts decreases the average degree of the ontology. Below are the results of a systematic study where the conceptual coherence and complexity was measured for the ontology ("Original"), then the core concepts were systematically removed, along with any isolated nodes and components that they produced when removed from the graph) and conceptual coherence and complexity were measured again.
Conceptual Integrity Metric - I am testing two calculations of conceptual coherence and complexity to provide an approximation of overall conceptual integrity. Currently they are labeled HZ1 and HZ2 (HZ stands for the Hsi-Zook measure). HZ1 is simply the product of coherence and complexity. HZ2 is the sum of the squares of coherence and complexity. These results are here mainly for completeness as these structural metrics will only be found to be meaningful with more data points and will probably have to be normalized against the size of the ontology.
A curious result from this analysis is evidence suggesting that removing Presentation improves both the coherence and complexity of the ontology. Intuitively, I believe that if the concept of Presentations were removed from the PowerPoint problem domain, a user could still create slides, handouts, and display slide shows but it wouldn't be called a presentation. It would be something else. Plus, since PowerPoint 2000 has the capability of saving presentations in other formats, the concept of a presentation might just serve as a general organizing concept. Not curious are the metrics that show that removing the Genigraphics Wizard and Online Broadcast Tools reduce the incoherence of PowerPoint. Use Case SilhouetteA use cases are from the Unified Software Process. They are essentially scenarios modeled as a series of actions that systems perform to achieve a goal for a user. A use case silhouette highlights those concepts activated by elements in the morphology. For PowerPoint 2000, we used Microsoft PowerPoint 2000 for Dummies for Windows by Doug Lowe. PowerPoint 2000 Use Case Silhouette Statistics
We have hypothesized that a set of use cases reflecting average usage of an application should invoke a concepts in the ontology with a frequency that should parallel the structural importance of this concept within the ontology. In other words, core concepts identified from the ontology should also be the most important concepts in a typical set of use cases. We determined frequency by counting the number of times a concept appeared in a use case. The table below lists the most frequently accessed concepts in the use cases along with their centrality values and whether they have membership in the set of core concepts. Partial List of Concepts Ordered by Times Referenced in Use Cases
Many of the concepts on the list do not appear as core concepts in the ontology. Some of this can be attributed to how the use cases were written. For example, Current Slide appears frequently because a use case will start “Move to the slide” – which refers to the current slide as opposed to a more general reference to slides. In the case of the Views, many use cases would have an instruction “Switch to the Slide View”. Clipboard functions – copying and pasted – were also referenced often throughout the cases as methods for placing certain objects. The use case silhouette does show that the Slide Master has much more importance in PowerPoint usage than the structural metrics suggest. The author referenced using slide masters for many things, including creating backgrounds, templates, changing footers, and so on. We first checked to see if we had made a modeling error in our ontology. The figure below shows the Slide Master ontology.
While we stand by our use of an association to show the relationship of the Slide Master to Slide – that the Slide Master formats all Slides in a Presentation – this example shows a potential weakness of our representation. The Slide Master formatting affects all the slides in a presentation and thus, it would follow that a user would take extra care to ensure that the Master was designed correctly. On the other hand, this example demonstrates how use case silhouetting can reveal potential areas of concern for developers that would not have emerged in an abstract requirements analysis or testing methodology. At least three of the core concepts can be found on both lists: Presentation, Slide, and AutoShape [Draw] Object. Text, Selection, and Color are also core concepts and Slide Show does appear as an important concept in the Use Cases, as do Line and File. While this partially proves our hypothesis, there are a number of concepts that do not appear as frequently in the use cases as their centrality values would suggest. We have listed these below Core Concepts Absent from Use Case Silhouettes
Slide Object, PowerPoint File, Notes Page Object, Animation, and [Configuration] are partially inferred concepts that we structured to model certain important generalizations. Because they were constructed somewhat indirectly from the PowerPoint application for modeling convenience, we would expect most use cases not to refer to them explicitly. However, AutoShape objects, Text Boxes, and ClipArt are considered Slide and Notes Page Objects. Slide Objects have animations and action settings. Animation ties together the concepts of Slide Transition, Entry Animation, and Preset Animation. Presentations and Slide Shows are also PowerPoint Files. The [Configuration] models the possible global options that can be set in the PowerPoint Tools Menu. If we accounted for the complete generalization and aggregation in the counts of those particular concepts, they would appear with more frequency across all the use cases. The other missing concepts seem to confirm what we had proposed earlier – that the features whose removal from PowerPoint 2000 improved its conceptual coherence are peripheral to the application, despite their centrality values. Online Broadcast, the Genigraphics Wizard, and WordArt all have sufficiently large ontologies that they appear as core concepts in the ontology. Now, a use case silhouette consisting of a set of ‘likely’ uses of the application show that these are not very important to users. One use case was devoted to each topic, on average. It is likely that if we were to perform the same analysis using actual user data that these concepts may disappear entirely from the analysis.To supplement the analysis and make the excavation artifacts more tractable for viewing , I created visualizations of the large graphs representing the morphology and ontology of the application. Visualizations were generated by UCINET (Borgatti, S.P., Everett, M.G. and Freeman, L.C. 2002. UCINET for Windows: Software for Social Network Analysis. Harvard: Analytic Technologies). The morphology and ontology graphs were saved as DL-format files (essentially adjacency list representations for the graphs they represent) which were then imported into NetDraw (Borgatti, S.P. 2002. Netdraw. Harvard: Analytic Technologies). After applying a spring embedding algorithm, which sets the length of an edge between two nodes as a function of the other nodes in the neighborhood, the diagram is saved as a KineMAGE. MAGE is a program for visualizing organic molecules and protein configurations developed by David Richardson through the Biochemistry Department at Duke University, NC. I added additional colorings of the kinemages to highlight salient features in the diagrams.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
