Feasibility Rationale Document
International (French) Cross-Cultural Teaching Model

Team 16
Last modified: December 3, 1997
CS577A


Table of Contents

  1. System Overview
    1. System Objectives
    2. System Scope and Context

  2. List of Documents

  3. Product Rationale
    1. Business Case Analysis
      1. Development Cost Analysis
      2. Operational Cost Estimate
      3. Estimate of Value Added and Relation to Cost
    2. Stakeholder Concurrence
    3. Requirements Satisfaction
      1. Capability Requirements
      2. Interface Requirements
      3. Quality Requirements
      4. Evolution Requirements
    4. Operational Concept Satisfaction

  4. Process Rationale
    1. System Priorities
    2. Process Match to System Priorities
    3. Consistency of Priorities, Process and Resources

  5. Risk Analysis

  6. Appendix

  7. Glossary

1.0 System Overview

The International (French) Cross-Cultural Teaching Model (ICTM) can be viewed as a teaching vehicle by which instructors can more effectively provide course instruction via the internet. For a more detailed overview of the ICTM, please refer to the System Overview of the OCD (Section 1.0).

1.1 System Objectives

Briefly speaking, the objective of the ICTM is to simplify and consolidate the process by which instructors may provide course-related multimedia materials to their students, thus indirectly providing the students with an intellectually rich, diverse atmosphere through which they can more effectively study and learn about, in this case, the French culture. For further details on the objectives of the ICTM, please refer to the System Objectives of the OCD (Section 1.1) .

1.2 System Scope and Context

The ICTM could potentially develop into a large scale software product comprised of components that provide users with extensive functionality; however, under the proposed time constraints, the scope and context of the ICTM system is limited to basic functionality that still achieves the fundamental goals of the system. The scope and context of the ICTM are further detailed in the System Scope and Context of the OCD (Section 1.2) .

2.0 List of Documents

Operational Concept Definition
System and Software Requirements Description
System and Software Architecture Definition
Software Development Plan (Life Cycle Plan)
Prototype
Feasibility Rationale

3.0 Product Rationale

3.1 Business Case Analysis

The operational costs associated with the International (French) Cross-Cultural Teaching Model are estimated and analyzed in the following sections. The objective of the business case analysis is to compare the value and cost of the current system that is employed by the French and Italian department against the value and cost of the proposed ICTM system. For further detailed analysis on the motives for changing from the current system to the proposed ICTM system, please refer to the Rationale for New Capabilities of the OCD (Section 4.3) .

3.1.1 Development Cost Analysis

The cost of development for this project is essentially zero because this project will be developed for free by a team of graduate students enrolled in CS577b. In addition to the free labor costs, the equipment expenses are also negligible since the system will be developed on either USC provided hardware or on student-owned computers. However, if a server cannot be found to house the ICTM system, then the development cost will increase considerably to cover the cost and purchase of an appropriate server. Inherent to the development costs of this project are also minor expenses in the form of storage media, paper, etc.

3.1.2 Operational Cost Estimate

The primary operational cost of the current system involves hiring student assistants to develop and maintain web pages for multimedia course deliverables as well as issuing course materials via more traditional methods such as handouts, in-class slides, etc. As stated by the customer, the costs for web development are currently minimal: theoretically $7.25/hr but, because the customer only hires work study students, the customer only ends up paying about 50% of the total hourly rate. The design of the page took one week (20 hours) and the student assistant is currently contributing about one hour per week on the project, since the "template" is already in place.

One of the objectives of the proposed system is to simplify the web page creation/modification process enough so that instructors can input material themselves. This can lower labor costs by reducing the amount of time student assistants will have to spend on data entry and site development. Offsetting this reduction in operating costs, however, is the additional cost necessary for account maintenance where both student and professor user accounts will have to be established to prohibit unauthorized access to the system.

But the primary objective of this system is to allow multiple local or remote instructors to input their material into this collaborative, instructional site. This can reduce operating costs in two ways: 1) it decreases the supplies and paperwork needed to accommodate the additional information and 2) it reduces the amount of time the student site assistants will have to spend on data retrieval and entry.

Therefore, it is estimated that there will be a net reduction in operational costs.

3.1.3 Estimate of Value Added and Relation to Cost

Although the International (French) Cross-Cultural Teaching Model does not generate revenue for the Department of French and Italian, its value can be measured by the enhanced instructional environment that it provides.

The comparison of the value added versus the relation to cost between the current system and the proposed system can be noted in two ways: 1) cost in terms of time spent in "clerical" and "housework" tasks (tagging, deleting outdated materials, checking outdated links) will be significantly reduced if not eliminated in the proposed system and 2) which will in turn allow for a system which allows for greater concentration in content. The advantages are thus qualitatively enhanced, while quantitative measures become less important.

By simplifying the web page creation/modification process and by placing the teaching model on the internet, authorized instructors from around the word, in this case Paris, could input their material to this site, comfortably and efficiently. With multiple instructors contributing their material to the teaching site, the breadth and quality of material is augmented, thus providing students with a more comprehensive educational experience.

In addition, the International (French) Cross-Cultural Teaching Model benefits students in that it consolidates course-related information into one site, making it more convenient for students to access material and thus encouraging students to explore and make use of the site. Furthermore, by incorporating a search element, a translation element, and a communications element (bulletin board) into our proposed system, students are provided a more effective, a more efficient vehicle by which they can learn about contemporary French culture.

3.2 Stakeholder Concurrence

The three primary stakeholders in this project (user, developer, customer) used the WinWin negotiation process to identify and define the requirements of the proposed system. Project stakeholders specified their win conditions at the beginning of the negotiations. Subsequently, issues related to win conditions were identified and options were proposed to resolve the issues. The conclusion of the WinWin negotiation produced agreements to the win conditions and options that establish the functionality for our proposed system.

3.3 Requirements Satisfaction

3.3.1 Capability Requirements

The capability requirements for the ICTM proposed system can be divided into four categories: general capabilities, student site capabilities, instructor site capabilities, and optional capabilities. The general capabilities are those capabilities which are common to both the student and instructor sites. The student site capabilities are those capabilities contained at the student site which are intended primarily for student use. But because instructors also have access to the student pages, they are able to make use of the student capabilities as well. Students, on the other hand, do not enjoy the same privileges as the instructors and therefore cannot gain access to the instructor site, nor can they make use of the capabilities contained at that site. Thus, he instructor site capabilities are restricted for instructor use only. The optional capabilities are categorized by those features which, though they enhance the overall functionality of the ICTM system, should be considered only if resources, technology, and/or time becomes available.

Note that the capabilities of this system are available only to authorized users - students enrolled in the course and participating instructors.

General Capabilities:

Student Site Capabilities:

Instructor Site Capabilities:

Optional Capabilities:

3.3.2 Interface Requirements

The International (French) Cross-Cultural Teaching Model is designed to be a semi-autonomous system. The ICTM does not require an interface with SIRSI, but does require extensive interfacing with a web server. The various interface requirements for the ICTM are as follows:

3.3.3 Quality Requirements

3.3.4 Evolution Requirements

The primary evolution requirement for the ICTM is the capacity for interdepartmental usage. This would involve porting the ICTM to support other single byte and multi-byte character sets. Correspondingly, the search, translation, and on-line help should also support the department's language that the ICTM is being ported to.

Another evolution requirement is the capacity to perform queries on text contained in images, which would then eliminate the need for keyword descriptions associated with each image. JSTOR presents one intriguing possibility that could be implemented in the near future.

In all likelihood, the ICTM will not support link updating meaning that obsolete or non-existing links will not be automatically updated by the ICTM system. Instead, it will be the responsibility of the system administrator to update these obsolete links. Once technology has evolved to allow such a capability, the link updating shall remain an evolution requirement at this stage.

In addition, any evolution in hardware or web browser technology should not impact the ICTM to a significant degree since the ICTM only requires that the server hardware have a network interface card and that the browser retain its basic functionality.

3.4 Operational Concept Satisfaction

Several mainstream scenarios are detailed below addressed from three points of view: the system administrator, the student user, and the instructor user. In addition, an exception-handling scenario is included. Each scenario assumes proper validation on login.

System administrator scenarios:

Student user scenarios:

Instructor user scenarios:

Exception-handling scenarios:

4.0 Process Rationale

4.1 System Priorities

Assuming that this project will be assigned a five person student team working at part-time, the system capabilities are divided into three categories based on the customers' requirements and potential time constraints: essential, desirable, and optional. Each of the capabilities are further grouped into four component modules based on functional similarity: collaboration, communications, tools, and system.

Essential capabilities:

The essential capabilities are the core capabilities of the ICTM system and must be implemented prior to May 1, 1998. The essential capabilities are listed below:


Desirable capabilities:

Following the implementation of the essential capabilities and if time permits, as many desirable capabilities are to be incorporated into the system as possible. The desirable capabilities are listed below:


Optional capabilities:

The optional capabilities should be viewed as potential future enhancements to the system. The optional capabilities are listed below:

4.2 Process Match to System Priorities

Using the Process Model Decision Table, the Spiral Model in conjunction with a risk-driven waterfall approach will be the primary software development processes that will be used to develop the International (French) Cross-Cultural Teaching Model. The decision to pursue the Spiral Model is shown below in table 4.2.


Growth envelope Understanding of requirements Robustness Available technology Architecture understanding
Limited to Large
High
Medium
Partial COTS
Medium

Table 4.2


The most important factors that contributed to the decision to pursue the spiral process model were the growth envelope of the system, the robustness of the project, and the technology available to the system. The growth envelope of the ICTM is fairly large because the system should be able to handle large amounts of multimedia information that will be contributed to the site by a multitude of instructors. In addition, the ICTM may be expanded interdepartmentally between foreign language departments and perhaps between various universities. The ICTM can also be expanded to support a large user-base. But as the ICTM stands currently, it is focused on the curriculum of two courses within the French and Italian department at USC with a projected low user-base. The ICTM is also designed to be fairly robust. Minor errors on both the student and user ends will be gracefully handled and should not halt the system, preventing other users from continuing to explore the student or instructor sites; however, major problems such as unexpected server crashes leading to downtime are not assumed, which is why our proposed system falls short of being categorized as being highly or fully robust. Lastly, the ICTM is dependent on commercial-off-the-shelf software applications in the form of browsers, translation programs, and Java (for developmental purposes).

A risk-driven waterfall strategy will be embedded within the spiral model to accommodate the two incremental development phases of the ICTM. The first incremental release will contain the primary core (essential) capabilities of the ICTM system which is comprised of the primary functionality available to both users as well as those specific to the instructor users: web page creation, web page authoring, web page selection, a bulletin board, the form emailer, logging in, translating, and fetching multimedia files. The second incremental software release will contain the secondary core (desirable) capabilities of the ICTM which is primarily comprised of functionality that will be available at the student site: time stamping of pages, an on-line help manual, and the form emailer.

The Overall Development Strategy of the Software Development Plan (Section 2.1) goes into further detail about how the Spiral and risk-driven waterfall approaches will be used in developing, testing, and implementing the system capabilities into the ICTM.

Note that the number of components is heavily weighted in the first increment and the Detailed Schedule of Deliverables of the Software Development Plan (Section 2.2) reflects this. The first incremental development phase is scheduled to begin on roughly on February 1, 1998 and release on April 1, 1998. The second incremental development phase is scheduled to being roughly on March 16, 1998 and release on May 1, 1998.

4.3 Consistency of Priorities, Process and Resources

COCOMO II was used to analyze and determine the consistency of the priorities that were specified for the system capabilities throughout each of the LCA documents. After analyzing the COCOMO II results, it was determined that the priorities assigned to each of the ICTM system modules -- collaboration, communications, tools, and system -- was consistent.

As shown below in figure 4.3, the COCOMO II results estimated that it would most likely take 5.5 full time software personnel (FSWP) 11.6 person months to complete the entire International (French) Cross-Cultural Teaching Model system. Note that because COCOMO II does not allow a SLOC count below 2000, the system and communications modules were combined into one component module consisting of a combined 3233 SLOC. Please refer to figure 6.0a in the Appendix for the original overall ICTM SLOC estimation.

Figure 4.3a ICTM COCOMO Model

This estimation does not agree with the original estimation that this project would only require a team of five graduate students a period of four months to complete. However, we feel that this estimation is inflated for the following reasons:

  1. The largest ICTM module, in terms of SLOC, is the collaboration module which has been estimated to consist of 5837.5 source lines of code. Approximately 87% of the overall SLOC count for the collaboration module involves the web page authoring portion which in itself is estimated to have a SLOC count of 5075; however, we feel that the SLOC estimation that we originally assigned to this specific component is inflated since this estimation is based on the CS577b team developing this collaboration java applet from scratch. We are under the impression that the source code for some java applet can be found on the web that will significantly reduce the SLOC (up to 50%) for the web authoring portion of the collaboration module.
  2. Again, the original SLOC estimates are based on the CS577b developers coding the ICTM from scratch. But the ICTM final prototype has done a good portion of the HTML coding of the various module and assuming that the CS577b development team chooses to build their functionality on top of our prototype, this should significantly reduce the amount of coding that they will have to do.
  3. There was a considerable amount of discrepancy in our original combined SLOC estimates, particularly in the web authoring and search components. Each team member then presented his reasononing for each of the individual SLOC estimates and a new set of estimations were obtained as shown in figure 6.0b in the Appendix . The discrepancies were eliminated and the overall SLOC count for the ICTM system were subsequently reduced.
  4. It is a distinct possibility that the graduate students working on the ICTM project next semester can be more productive than FSWP.

Figure 4.3b below shows the adjusted COCOMO model which uses the new SLOC estimates in figure 6.0b in the Appendix and has removed the desirable and optional capabilities in order to focus on the schedule for the essential capabilities of the ICTM system.

Figure 4.3b Adjusted ICTM COCOMO Model

The new COCOMO results estimated that it would most likely take 5.0 full time software personnel (FSWP) 8.2 person months to complete the essential capabilites of the ICTM system. This estimation does not take into consideration software reuse. Based on this new estimation, we still firmly believe that it will be feasible to implement the ICTM system in the alloted four month period given a team of five graduate students.

5.0 Risk Analysis

The sections below summarize and identify the top risk items associated with the proposed ICTM system. For further detailed analysis, please refer to the Software Development Plan (Section 4.1)


  1. Personnel shortfalls:
    At this point, three of the four team members from the CS577a team have elected not to enroll in CS577b next semester. This means that virtually a new team of five graduate students in CS577b will be assigned to develop the ICTM posing a potential risk. Each team member's ability will vary depending on their educational backgrounds and work experiences, thus making it nearly impossible to accurately predict how the personnel will factor into the system development process. In addition, compatibility conflicts could arise between team members decreasing the overall efficiency of the team. Loss of personnel is another concern that could greatly impact the status of the project, particularly if the loss occurs during critical phases of the life cycle.

    Resolution:
    To reduce the risk of personnel shortfalls, particularly in the area of technical expertise, it is strongly suggested that each CS577b developer be proficient in the one or more of following areas:

    • Java programming
    • CGI scripting
    • HTML
    • Networking and concurrency

  2. Schedule constraints:
    Schedule constraints are a major risk item since the development team is fixed to a twelve week schedule next semester and COCOMO estimates have projected that the ICTM system will take nearly 10 months to complete.

    Resolution:
    To successfully resolve this risk item next semester, the CS577b project team should adhere to the incremental development plan stated in the Software Development Plan and scrub requirements (after consulting with the customer) if necessary. More importantly, the CS577b should attempt to reuse as much code as possible especially in regards to the java applet that will serve as the collaboration module. The CS577b should also consider building on top of the prototype that the CS577a team has developed. Much of the HTML coding has already been done and the customer has expressed her satisfaction with the prototype interface.

  3. Translation provider risks:
    At this point, it is unknown whether a commercial on-line translation provider exists that will reliably and accurately translate HTML documents at little to no cost for the customer. Accurately translating documents from one language to another requires knowledge of not only the language itself, but also the subtle nuances that oftentimes accompany it. Currently one translation provider has been found whose Power Translator for UNIX seems to be a viable option for a on-line translation service. However, it has been discovered that the Power Translator does a very literal translation which thus reduces the accuracy of the translation. A tentative agreement has been reached with Vivigy, Inc. that would provide free on-line translation service for the ICTM system in exchange for advertising space on the ICTM student and/or instructor sites. The Vivigy, Inc. Power Translator for UNIX demo can be tested at http://www.vivigy.com/Demomain.htm and the person to contact at Vivigy is Rod Young (ryoung@vivigy.com) . If the CS577b ICTM team cannot reach a final agreement with Vivigy, Inc. then the risk of finding a translation provider becomes very high because an alternate translation service has not been found. Again, it is very important that the CS577b ICTM team find a translation provider that is on-line since a platform specific translation COTS software application cannot feasibly be distributed to all registered users of the system.

    Resolution:
    The client has agreed to help search for a reliable on-line translator that the CS577b team can use next semester. At this point, the client has expressed her desire not to use Power Translator due to its literal translation; however, as a last resort, this may seem a viable solution which would require a warning posted somewhere on the ICTM site stating the potential loss of meaning with the translations.

    Note that the order that this risk should be resolved should be in the following order: accessibility (on-line), interface, accuracy/reliability, and cost. In other words, an on-line translator that provides service at a fixed cost should be considered over a platform specific, software package that is provided for free.

  4. Server hardware risk:
    At this point, a server has not been found that will house the ICTM system. Both the Center for Excellence in Teaching and the French and Italian department do not own their own servers.

    Resolution:
    If a server has not been found by the time the ICTM system has undergone development next semester, the developers can, for the time being, develop the system on UCS servers. Once a server has been found, all of the files and documents associated with the ICTM project can then be moved over to the new server.

  5. Incorrect user interface:
    The user interface is one of the single most important elements associated with the ICTM. Thus, it is unknown whether the team members assigned to this project next semester will develop an interface that accurately reflects what the customer desires.

    Resolution:
    Prototyping and incremental development

  6. Gold plating:
    Gold plating is the process by which the specifications and capabilities for the proposed system are overambitious, leading to excessive functionality that may be cost ineffective and/or infeasible within the given time constraints.

    Resolution:
    Requirements scrubbing, prototyping, cost-benefit analysis, and designing to cost

  7. Continuing stream of new or changing requirements:
    During the course of system development, there exists a distinct possibility that the proposed requirements may be altered or may be modified to accomodate new ones. Unfortunately, the advantages of early and frequent prototyping are offset by the risk that the customer may want to modify or add requirements to the preexisting set.

    Resolution:
    High change threshold, information hiding, and incremental development by defering changes to later increments

  8. Real-time performance shortfalls:
    It is difficult to analyze the performance of the various features of the ICTM since so many variables are involved when determining the performance of a function. For instance, it is impossible to accurately determine the performance of the translation feature since it is unknown at this point what translation provider will be used.

    Resolution:
    Simulation, benchmarking, modeling, prototyping, instrumentation, and tuning

6.0 Appendix


Figure 6.0a Team 16 SLOC Estimation

As shown in figure 6.0a above, there are a number of discrepancies in the SLOC estimates particularly in the web authoring and the search components. Thus, each of our team members submitted their reasoning for their respective SLOC counts and a new set of SLOC estimates were obtained as shown below in figure 6.0b below. For example, all team members, except the architect, believed that the search component of the ICTM system would involve a considerable amount of coding; however, we all agreed that since the search has been refined to a simple string-search (UNIX grep), the SLOC count should be reduced to match what the architect had originally estimated (50).


Figure 6.0b Adjusted Team 16 SLOC Estimation

7.0 Glossary

Ref. OCD (Section 10.0)


ICTM home page ICTM prototype Back to top