The simulation engine will handle a branching conversation. The
conversation proceeds linearly until the user is presented with a
choice. When the user makes a choice the conversation then follows that
branch linearly until the next user choice or the end of the
conversation. During the first pass, the user can see previous
statements, but cannot take a different course through the simulation.
At the end of the first pass, the user is directed to the analysis
portion. At the end of or during subsequent passes, the user has the
option to start again, use a back button, re-enter the analysis or
access "Library". Conversational elements have the form described on "Simulation" Content File Formats.
When the meeting ends, the learner is asked to analyze the conversation in terms of bias and procedural errors by selecting a statement from the transcript, and selecting or bookmarking its relevant bibliographic entries. If the learner has replayed the simulation and selected different choices at any of the decision points, new statements will show in the transcript, and new bibliographic entries will be displayed that relate to the new set of statements.
Build a front end in Director and Lingo which handles the display and animation of the current simulated speaker. The conversation is driven by an XML file ("Simulation" Content File Formats) and by the learners choices. The XML file is read by Director into some Lingo memory structure and traversed, as appropriate, through the conversation. The learner's choices are stored in memory and recorded as a transcript which can be accessed later by the learner. The simulation's statements are also recorded in this transcript.
Apparently, Director has some primitive XML reading methodologies, although it is not clear whether the available memory structures are sophisticated enough to handle transformation of the data.
Minuses:
- Director has a certain brittleness to it, this could max out Lingo's text handling capabilities
- There may be some issues with easily accessing the statement to reference mapping in Director.
Pluses:
- Creating application is fast and relatively easy.
- Easy to add media.
- The front end part of this task is well suited to Director.
Build a front end in Director & Lingo which handles the display and animation of the current simulated speaker. Also, build a front end for the analysis screen in Director & Lingo which handles analysis of the conversational transcript. The conversation is driven by an XML file ("Simulation" Content File Formats) and
by the learners choices. The XML file is read by Java (JAXB?) into a
memory structure. Director queries the java engine for the next
statement as necessary. The learners choices are stored in the director applications memory and
recorded as a transcript which can be accessed later by the learner.
(The sim's statements are also recorded in this transcript.)
Minuses:
- Using Java plus Director creates additional installation hurdles. Setting up communication shouldn't be too difficult, but we need to research it before we commit.
- The use of Java here may be overkill.
Pluses:
- Creating application is fast and relatively easy.
- Easy to add media.
- Java is great for manipulating xml.
Front end pages are generated based on learner actions and driven off of a ColdFusion backend. Creating a web page which can handle such a sophisticated interface may be tricky. This is something we may need to break into separate pages or into a flash animation (with data provided by Coldfusion). The conversation files are still XML, but they are parsed by ColdFusion. Graphics are handled by Flash or images on HTML. All links are live. The learners choices are stored in the ColdFusion database. (This means that we will need to deal with issues of "user comfort and anonymity" upfront.) We will also need to have some sort of user login mechanism.
Minuses:
- Web access required.
Pluses:
- Easy to update & improve the application.
- Application very visible and easy to access for demos.
- Easy to accumulate datum for feedback such as "what did other people say?".
- Web access guaranteed.
Front end is all web, each conversational step is created by moving to the next web page. A website is a natural analogy to this branching dialog. We could use Dreamweaver and Contribute to make it easier to create the pages. The learners choices are stored on a server and recorded as a transcript which can be accessed later by the learner. The simulation's statements are also recorded in this transcript. This means that we will need to build a backend processor to support this architecture. Creating a web page which can handle such a sophisticated interface may be tricky. This is something we may need to break into separate pages or into a flash animation.
Minuses:
- The conversational part is relatively easy, the analysis part is less easy.
- The conversational part is relatively easy, the analysis part is less easy.
- Difficult to add media.
- Web access required.
- Building the analysis page will not be easy.
Pluses:
- Creating application is fast and relatively easy.
- Fairly easy to change the conversational structure.
- Web access guaranteed.
Interaction Overview (Actors and Use Case)
Content Overview / File Formats
ADEPT Goals
ADEPT Tool Design
ADEPT Tool Development
ADEPT Project Schedule
ADEPT Minutes
ADEPT Prototypes
Georgia Tech Resources
Outside Resources