A design of self-questioning mechanism for information requirement specification.

Huang, I-Lin

ABSTRACT

Incorrect requirement specifications are widely recognized as the major cause of information system failures. In order to improve the correctness of requirement specifications, various requirement specification techniques such as Data Flow Diagram, and Object Model, have been invented to help information analysts capture, understand, and represent information requirements. However, information analysts' cognitive abilities are still the most important determinant for the correctness of requirement specifications. Empirical studies have showed that novice information analysts cannot use requirement specification techniques effectively; and hence their performance of specifying information requirements is significantly lower than that of expert information analysts.

Self-questioning has long been recognized in the field of learning research as a strategy that can improve students' cognitive abilities on reading comprehension and problem solving. In order to improve the cognitive abilities of novice information analysts, this research argued that novice information analysts should be trained to incorporate self-questioning mechanism into their requirement specification process. On the basis of the theories on human cognition, this research also proposed several design strategies for self-asking questions that can guide novice information analysts to make more effective model-based reasoning and hence to achieve a higher correctness of requirement specifications.

INTRODUCTION

Specifying information requirements for the target information system under development is an important step in information systems development. The major output of specifying information requirements is a set of information requirement specifications (or simply requirement specifications) that state the desired functional and performance characteristics of the target information system (Roman, 1985). Basically, there are three purposes of requirement specifications: (1) facilitating an understanding of the target system, (2) guiding the process of information system design, and (3) serving as a basis for all communications concerning the information system being developed (Hsia, Davis & Kung, 1993; Schemer, 1987).

Incorrect requirement specifications are widely recognized as the major cause of information system failures (Dardenne, van Lamsweerde & Fickas, 1993; Davis, 1988; Dorfman, 1990; Greenspan & Mylopoulos, 1982; Scharer, 1981; Vessey & Conger, 1994). It has been reported that about two-thirds of information system failures can be attributed to the mistakes made in requirement specifications (Fraser, Kumar & Vaishnavi, 1991; Shemer, 1987). According to an estimation, incorrect requirement specifications may cost fifty to one hundred times more than what would have been required if the errors are not discovered until system implementation (Roman, 1985; Shemer, 1987).

In order to improve the correctness of requirement specifications, various requirement specification techniques such as Data Flow Diagram, and Object Model, have been invented to help information analysts capture, understand, and represent information requirements (Couger, Colter & Knapp, 1982; Davis, 1988; Wieringa, 1998). However, the cognitive abilities of information analysts are still the most important determinant for the correctness of requirement specifications (Schenk, Vitalari & Davis, 1998). It has been reported that expert information analysts can use requirement specification techniques effectively by retrieving and applying more relevant modeling principles, making more critical testing of hypotheses and finally achieving requirement specifications with higher correctness (Allwood, 1986; Koubek, et al, 1989; Vitalari & Dickson, 1983). On the other hand, novice information analysts cannot use requirement specification techniques effectively because they have difficulties in correctly identifying important concepts in problem statements. The inadequacy of cognitive abilities makes novice information analysts unable to perform model-based reasoning effectively, and therefore leads to more errors in requirement specifications (Batra & Davis, 1992; Batra & Sein, 1994; Sutcliffe & Maiden, 1992).

It is believed that there are two cognitive characteristics that account for the difference of cognitive abilities between novice and expert information analysts: (1) reasoning processes, and (2) the knowledge organizations that support the reasoning processes. In addition, the learning process is slow for novice information analysts to reach the expert level of reasoning processes and knowledge organizations (Huang & Burns, 2000; Schenk, Vitalari & Davis, 1998). Therefore, the research question for this research is how to help novice information analysts achieve more correct requirement specifications even with relatively inadequate reasoning processes and knowledge organizations.

In the field of learning research, self-questioning is regarded as a cognitive strategy that can help students focus attention, organize new material, and finally integrate the new information with existing knowledge (Doerr & Tripp, 1999; Glaubman & Ofir, 1997; King, 1989, 1992; Wong, 1985). Empirical studies also showed that self-questioning can improve students' abilities of reading comprehension and problem solving (Doerr & Tripp, 1999; King 1992; Wong, 1985). Therefore, this research argued that novice information analysts should be trained to incorporate self-questioning mechanism into their cognitive process of information requirement specification. When novice information analysts can ask right questions to themselves during requirement specification, they will be able to learn how to think like an expert, organize their knowledge like an expert, and finally specify information requirements like an expert.

In order for the self-questioning mechanism to effectively guide novice information analysts to specify correct information requirements, a set of self-asking questions should be available to novice information analysts and meet at least two criteria: (1) can lead to correct requirement specifications, and (2) can fit the cognitive behavior of novice information analysts naturally. Currently, the common practice of requirement specification is focused on deriving information requirements from the answers of the generic questions for the constructs of requirement specification techniques. However, the generic questions do not fit the cognitive behavior of novice information analysts naturally. Consequently, the generic questions may induce from novice information analysts the answers that lead to incorrect requirement specifications. Therefore, on the basis of the theories in human cognition (Gernsbacher, 1990; Graesser, 1995; Kintsch, 1988; Ortony, 1978), this research proposed that basic objects, coherence, and systematicity are the three cognitive principles that should be the guides for designing effective self-asking questions. For demonstration, a set of self-asking questions deriving from the three cognitive principles were designed to help specify information requirements by Data Flow Diagram.

The rest of this article will be organized into four sections. First, the self-questioning as a strategy to support human reading comprehension and problem solving will be discussed. Second, basic objects, coherence, and systematicity as guides for designing self-asking questions will be elaborated. In addition, Data Flow Diagram will be used to demonstrate the design of self-asking questions. Third, a brief example will be used to illustrate on how the self-asking questions support the process of specifying correct requirements. Finally, the conclusion will be made in the final section.

SELF-QUESTIONING AS A LEARNING STRATEGY

In learning research, self-questioning has been promoted as a strategy to improve students' abilities of reading comprehension and problem solving (Chin & Chia, 2004; Doerr & Tripp, 1999; Wong, 1985). When applying the strategy in reading comprehension or problem solving, the students are instructed to ask themselves questions before, during or after reading a text (for reading comprehension) or a problem statement (for problem solving). In this section, the discussion will be focused on three aspects of self-questioning in learning research: the benefits, the timing, and the questions of self-questioning.

The Benefits of Self-questioning

The benefits of self-questioning have been explored in learning research from three perspectives: active processing theory, metacognitive theory, and schema theory (Wong, 1985). Firstly, active processing theory suggests that it is the quantity, rather than the quality of self-asking questions that improves students' performance of reading comprehension or problem solving. The theory argues that actively asking questions during reading comprehension or problem solving can help students focus their thinking on the reading material, and hence improve their performance of reading comprehension and problem solving. The argument of active processing theory is supported by the evidence that the questions asked by students themselves facilitate understanding better than those asked by the instructors. Even further, more questions asked during reading or problem solving result in better comprehension and retention.

Secondly, metacognition theory suggests that the function of self-questioning is to help students monitor their understanding of the reading material (King, 1989; Nolan, 1991; Ozgungor & Guthrie, 2004). Therefore, the right self-asking questions as a metacognitive strategy should be able to help students focus on the important aspects of the material they read. Through a questioning--answering--questioning cycle, students can effectively analyze the content, relate it to their prior knowledge, and finally evaluate it and reassign their cognitive resources accordingly. This comprehension-monitoring process can not only let students know what they have learned, but also make them become aware of what they have not yet understood. When students fail to answer the questions they themselves post, they can take remedial action by asking themselves related questions, or asking questions of other information sources.

Finally, schema theory believes that students' reading comprehension or problem solving depends on their ability of activating prior knowledge (or called schema) in their minds (Chin & Chia, 2004; Wong, 1985). Therefore, the right questions asked by students themselves should be able to trigger the related concepts and experiences in students' minds, help students integrate their knowledge with the ideas in the reading material, and as a result, achieve better reading comprehension or problem solving.

The Timing of Self-questioning

It is suggested that self-questioning can enhance students' reading comprehension and problem solving no matter it is performed before, during, or after reading a text. Normally, students are advised to ask questions during reading a text. The function of self-questioning under this situation is mainly seeking information to (1) remove an obstacle in a plan or problem, (2) resolve a contradiction between ideas, (3) explain an unusual or anomalous event, (4) fill an obvious gap in students' knowledge base, and finally (5) make a decision among a set of alternatives that are equally likely (Graesser & Person, 1994).

Asking questions of predicting the ideas of a text before reading it can also facilitate reading comprehension (Nolan, 1991; Osman & Hannafin, 1994). It is believed that prediction can activate a cognitive plan to guide the students to understand the text. In addition, prediction can motivate students to read the text to confirm their prediction (Kletzien & Bednar, 1988).

Finally, self-questioning after reading a text can be helpful for students' reading comprehension because it provides students with a way to test their understanding; that is, it helps them to evaluate how well they understand what they are studying (King, 1992).

The Questions of Self-questioning

The main function of self-questioning is supposed to facilitate students' comprehension by guiding them in focusing attention (active processing theory), identifying important information (metacognition theory), and integrating the important information with existing knowledge (schema theory). In order to achieve the function effectively, two types of questions have been proposed: domain-specific and domain-independent questions. Firstly, domain-specific questions are particularly useful in activating the prior content knowledge of the target text in students' minds. Therefore, domain-specific questions are very efficient in facilitating the integration between the new ideas and the students' prior knowledge. However, this approach will become very ineffective in situations that students have little or no background knowledge (Osman & Hannafin, 1994).

On the other hand, self-questioning with domain-independent concepts can improve students' comprehension in different domains. For example, when reading narratives, students can be taught to incorporate general story grammar elements such as leading character, action, obstacle, and outcome, into self-asking questions (Singer & Dolan, 1982). In this approach, students answer a general story grammar question for each element included in the story. Examples of general questions are as follows (Swanson, 1998): (a) Who is the leading character? (b) What is the leading character trying to accomplish? (c) What obstacles does the leading character encounter? Using the general questions as models, students create and answer their own specific questions based on the particular story they are reading, and hence reach a better understanding of the story.

In the next section, this research will discuss the three cognitive principles that can guide the design of effective self-asking questions for improving novice information analysts' ability to specify correct requirement specifications.

DESIGN STRATEGIES FOR SELF-QUESTIONING MECHANISM

During information requirement specification, users often provide information analysts with a verbal or written problem statement concerning the users' information requirements for a particular problem domain. The problem statement reflects the users' cognitive model of the problem domain. The task of the information analysts is to capture, understand, and represent the users' information requirements from the problem statement with the help of requirement specification techniques. From the perspective of human cognition, the cognitive behavior of the information analysts can be viewed as a reading comprehension process (Huang & Burns, 1998). Empirical studies on the modeling behavior of information analysts show a strong association among the activities of gathering information, identifying relevant facts, and conceptual modeling (Batra & Davis, 1992; Sutcliffe & Maiden, 1992). This strong association reflects that information requirement specification is basically an understanding process.

Because self-questioning is proven to be an effective strategy for improving reading comprehension and problem solving, this research argues that novice information analysts should be trained to incorporate self-questioning mechanism into their cognitive behavior of specifying information requirements. Even further, this research argues that effective self-asking questions should meet at least two criteria: First, they should promote the answers that can lead to correct requirement specifications. And second, they should be compatible to human cognition in specifying information requirements. Current practice of requirement specification is focused on the first criterion and tries to induce correct requirement specifications directly from the generic questions for the constructs of requirement specification techniques. However, due to the inadequacy of reasoning processes and knowledge organizations, novice information analysts are often unable to come up with right answers for the generic questions, resulting in incorrect requirement specifications. Therefore, the design strategies that meet the above two criteria for effective self-asking questions should be based on the cognitive principles that people follow in reading comprehension and problem solving. Specifically, this research proposes that the design of self-asking questions should follow the principles of basic concepts, coherence, and systematicity because they are the effective guides for human cognitive behavior in reading comprehension and conceptual modeling. The rest of this section will discuss the four design strategies for self-asking questions including the generic question approach, and the three approaches based on the principles of basic concepts, coherence, and systematicity. In this research, Data Flow Diagram is used as an example of requirement specification techniques for demonstrating the design of self-asking questions.

Generic Questions

Generic questions are the questions that directly map the constructs of the intended requirement specification technique onto the concepts in the problem statement. For example, Data Flow Diagram is a requirement specification technique with four constructs: external entity, data flow, process, and data store. The generic questions for specifying information requirements on the basis of Data Flow Diagram are like those in table 1.

By asking these generic questions, information analysts try to identify the relevant concepts in the problem statement and match them to the constructs of the intended requirement specification technique. Without rich experience and knowledge in requirement specification, novice information analysts basically use the definition of the constructs as shown in table 2 to guide them in identifying and mapping relevant concepts in the problem statement. Although this approach can focus novice information analysts on the relevant concepts, incorrect requirements or missing mappings may happen very often due to the inadequacy of novices' reasoning processes and knowledge organizations.

Basic Concepts

1. The principle

According to the research in concept formation (Komatsu, 1992; Rosch, et al., 1976), humans can recognize objects at different levels of abstraction. However, there is one level of abstraction at which the basic concepts of the objects are made. From the perspective of human cognition, the basic concepts are the concepts that carry the most information for human reasoning and are easiest to understand. For example, the concept of "chairs" can be viewed as "furniture" at a higher-level of abstraction or as "kitchen chairs" at a lower-level of abstraction. However, when we see a kitchen chair, most of us would recognize it as a chair rather than a piece of furniture or a kitchen chair because most of us view the concept of "chairs" as the basic concept. We recognize a kitchen chair as a chair rather than a kitchen chair because it is difficult to decide whether a chair is a kitchen chair or not due to few minor differences between kitchen chairs and other chairs. On the other hand, although we can easily classify a chair into furniture, we do not recognize a chair by the concept of furniture because the concept of chairs carries more information for human reasoning than furniture. The instances of furniture share fewer properties than those of chairs. Many objects like lamps can also be classified into furniture even though lamps are very different from chairs. As a result, there are fewer inferences can be made by the concept of furniture than by the concept of chairs

In order to manage infinite phenomena happening in the real world, we have to generate concepts to abstract the phenomena into groups, with the instances of each group sharing similar important properties. By abstraction, we can use a few concepts to handle infinite phenomena in the real world. However, by abstraction we also lose the information about the differences among the instances in the same group. Therefore, we recognize objects by the basic concepts because they are the concepts at the best level of abstraction that include the most phenomena but lose the least information. According to Komatsu (1992), the basic concepts are the most useful concepts for us because they are those formed during perception of the environment, those formed during our childhood, and those to be most codable, most coded, and most necessary in language.

2. The design strategy based on basic concepts

For analyzing business information systems, business concepts are the basic concepts for the business users and hence easier to identify and understand than the information concepts such as objects, and processes used by requirement analysis techniques. For example, customers as a business concept are more natural for users' cognition than objects in Object Model, or external entities in Data Flow Diagram because users can generate more inferences from the concept of customers than from that of objects, or external entities. Therefore, on the basis of the theories on concept formation, the self-asking questions that are focused on business concepts are easier to answer correctly by novice information analysts than those on information concepts that are used as constructs for requirement specification techniques.

Assume that an information analyst is reading a problem statement and feeling that the information requirements are concerned with order processing. And if the information analyst has the domain knowledge about order processing as shown in Figure 1, she or he can ask questions based on business concepts during requirement specification as in table 3:

[FIGURE 1 OMITTED]

It is obvious that the above questions about business concepts are easier to answer than those about information concepts. However, business concepts are domain-specific concepts and will become difficult to identify and model if there is no pre-exist template of requirement specifications, or if the information analysts have little or no background knowledge.

Coherence

1. The principle

Coherence is well accepted as the goal of human cognition on reading comprehension. Specifically, reading comprehension is a process of building a coherent situation model of the text or problem statement being comprehended (Gernsbacher, 1990; Graesser, 1995; Kintsch, 1988; Ortony, 1978). A situation model is a mental representation of the concepts that are explicitly mentioned in the problem statement and that are derived by the reader's knowledge (Graesser, Singer, and Trabasso, 1994; Kintsch, 1988; Mckoon & Ratcliff, 1992). People construct situation models in order to be able to answer the questions related to the problem statement. If people can answer the questions related to the problem statement by searching the structure of the situation model, then people feel that they have understood the material in the problem statement (Gernsbacher, 1990; Graesser, 1995; Kintsch, 1988; Ortony, 1978). Coherence is "a state or situation in which all the parts or ideas fit together well so that they form a united whole" (Collins Cobuild English Dictionary, 1995). The coherence of a situation model for a problem statement determines how well the reader can answer questions about the problem statement, and remember, summarize, or verify the statement (Kintsch, 1988; Trabasso, 1989). In general, the more coherent the situation model of a problem statement, the better the comprehension and usage of the problem statement (Trabasso, 1989).

From the perspective of reading comprehension, the task of information analysts is to make goal-directed inferences to achieve coherent requirement specifications from problem statements constrained by the predetermined constructs and cognitive structures of requirement analysis techniques. While information analysts analyze a problem statement on the basis of a particular requirement specification technique, they may find that the problem statement can not be fitted into the conceptual model very well. Information analysts then perform intensive coherence inferences to resolve the discrepancies. Although correctness may be the goal for information requirement specification, information analysts believe that understanding is achieved if the requirement specifications are coherent.

2. The design strategy based on coherence

With coherence as the goal for requirement specification, we can say that we have specified requirements for a problem statement accurately only if we can construct a coherent requirement model for the problem statement. On the other hand, if there is still any disconnected concept, or called incoherent concept, in our requirement model, we will feel that we fail to specify the requirements completely and accurately. In this situation, self-questioning can be invoked to ask questions that will be answered by our domain knowledge or the experts who have domain knowledge about the incoherent concept. If the answers can provide coherent connections for the incoherent concept, then the self-questioning serves its purpose and we have better chance to model the requirements completely and accurately.

Therefore, from the perspective of coherence, the purpose of self-questioning is to derive answers that can provide coherent structures for a concept. Coherence structures can be at least divided into two levels: local coherence and global coherence (Long, Oppy & Seely, 1997; Mckoon & Ratcliff, 1992). First, local coherence can be defined as "a small set of adjacent sentences that makes sense on its own or in combination with easily available general knowledge" (Mckoon & Ratcliff, 1992, p.444). From the perspective of local coherence, the coherent structures in Data Flow Diagram are the acceptable connections through data flows as shown in table 4. Derived from the coherent structures in table 4, a list of the possible self-asking questions that can be used to derive the locally coherent connections is thus shown in table 5.

Second, global coherence involves constructing a requirement model that reflects a bigger unit of problem statement like a paragraph, or even the whole problem statement (Long, Oppy & Seely, 1997). Two examples of the generic coherent structures defined at the global level are process decomposition structure (Martin, 1989), and event partition structure (McMenamin & Palmer, 1984). The coherent structure of event partition model is shown in figure 2. The questions that can drive the concepts in the event partition structure are shown in table 6.

[FIGURE 2 OMITTED]

Systematicity

1. The principle

When selecting a base structure (e.g. a conceptual structure from a particular requirement specification technique) to be mapped onto a target structure (e.g. a conceptual structures from a problem statement), the base structure with higher-order relation will be more likely to be imported into the target structure than is that with an isolated relation or object-attribute. It is called the principle of systematicity (Gentner & Markman, 1997). The principle of systematicity is a structural expression of our tacit preference for coherence and deductive power from the mapping. Basically, there are four different types of similarity that support the mapping from the base concepts onto the target concepts and provide different levels of coherence to the resultant model (Gentner, 1983): First, literal similarity provides the highest coherence because it matches both relational structure and object-descriptions (e.g., The order processing system is like that of company X I analyzed last year.). Second, analogy provides the second highest coherence because it matches relational structure and disregards object descriptions (e.g., The order processing system is like the library system I analyzed two years ago.). Third, abstraction provides the third highest coherence because the base structure is an abstraction of the target structure (e.g., I want to use the base structure, A-structure "external entity, request, process", from Data Flow Diagram to match the target structure). Fourth and finally, surface similarity provides the lowest coherence because it matches some aspects of object descriptions and disregards relational structure (e.g., I want to model customer as external entity, and order as data store).

2. The design strategy based on systematicity

The principle of systematicity is not a principle for deriving self-asking questions, but rather a principle for selecting the best set of questions among several sets of answerable questions. When there are several sets of self-asking questions can be raised to induce answers for information requirements, systematicity will select the set of self-asking questions that can induce answers to lead to the most coherent requirement specifications. On the basis of systematicity, the self-asking questions in table 3 will have the highest priority because they are derived from the principle of basic objects and provide a mapping on the basis of literal similarity. The self-asking questions in table 6 will receive the second highest priority because they are derived from the principle of (global) coherence and provide mappings on the basis of abstractions. The self-asking questions in table 5 also provide mappings on the basis of abstraction. However, they will have lower priority than those in table 6 because the questions in table 5 are derived from local coherence with lower order relations. Finally, the questions in table 1 can get only the lowest priority because they are generic questions and provide only the mapping on the basis of surface similarity.

The self-asking questions that provide a mapping on the basis of analogy, once identifies, will have higher priority than those on the basis of abstraction, or surface similarity. However, it is not easy to identify analogical relations between a base structure and a target structure at the first place (Maiden & Sutcliffe, 1992). Therefore, we do not actively pursue a set of self-asking questions for analogical mappings at the very early stage of requirement specification.

AN EXAMPLE OF SELF-QUESTIONING FOR DATA FLOW DIAGRAM MODELING

In this section, we will assume a requirement sentence, "The customer first sends an order to John, the order clerk", in a problem statement of an order processing system as an example to illustrate how self-questioning leads to correct requirement specifications. First, the novice information analysts will read the requirement sentence. Then guided by the principle of systematicity, they will raise different set of self-asking questions on the basis of their domain and modeling knowledge as discussed below:

Situation 1: Follow the Principle of Basic Concepts

1. The question-answering

When the information analyst has the prior experience on order processing system, he or she may raise the self-asking questions in table 3. The answers for the questions are in table 7 as below:

2. The outcome

After confirming the existence of an order file and a customer file in further reading of the problem statement, the information analyst can achieve a validated data flow diagram like in figure 1.

Situation 2: Follow Analogical Reasoning

1. The question-answering

When the information analyst has the prior experience on information systems such as library information systems that are similar to the order processing system under investigation, he or she may raise the self-asking questions derived from analogy. The answers for the questions are in table 8 as below:

2. The outcome

The information analyst can also achieve a validated data flow diagram like in situation 1. However, the information analyst may not pursue the analogical mappings because the analogical similarities are difficult to identify (Maiden & Sutcliffe, 1992). Situation 3: Follow the Principle of Coherence at the Global Level

1. The question-answering

When the information analyst has no prior experience on information systems similar to the order processing system under investigation, he or she may raise the self-asking questions derived from global coherence after fully understanding the intention of the whole problem statement. The answers for the questions are in table 9 as below:

2. The outcome

The information analyst can achieve a validated data flow diagram like in situation 1. However, the information analyst needs first to be able to understand the requirements of the problem statement at a global level.

Situation 4: Follow the Principle of Coherence at the Local Level

1. The question-answering

If the information analyst has no prior experience on information systems similar to the order processing system under investigation, and does not have an adequate understanding of the intention of the whole problem statement, he or she may pursue the answers for the self-asking questions from the principle of coherence at the local level. The information analyst first assumes that the incoherent concept "Customer" is an external entity. He or she then may try to validate his or her assumption by answering the questions in table 10 as below:

2. The outcome

On the basis of local coherence, the information analyst is focused on the concepts in the requirement sentence without referring to the rest part of the requirement statement. The information analyst can thus achieve a validated partial data flow diagram like in Figure 3.

[FIGURE 3 OMITTED]

Situation 5: Answer Generic Questions

1. The question-answering

If the information analyst has no domain knowledge on information systems similar to the order processing system under investigation, and does not have adequate modeling knowledge on Data Flow Diagram, he or she may try to specify the information requirements on the basis of the definitions of the constructs as in table 2. Therefore, the answers to the generic questions as in table 1 are like in table 11 as below:

2. The outcome

On the basis of the definitions of the constructs of Data Flow Diagram, the information analyst is focused on the surface meaning of the concepts in the requirement sentence and tries to categorize the concepts into different constructs. Some errors may happen due to the lack of validation from the coherent structure of the requirement specification.

CONCLUSION

Traditionally, the research studies for improving the modeling performance of novice information analysts have been focused on providing access to domain and modeling knowledge through computer-aided software engineering (CASE) tools. However, this research is concerned with how to improve novices' cognitive abilities for requirement specification. On the basis of the principles of basic concepts, coherence, and systematicity, this research has proposed design strategies for effective self-asking questions. Guided by the effective self-asking questions, novice information analysts can specify more correct information requirements.

In sum, the design of self-questioning mechanism of requirement specification contributes to the research in information requirement specification in the following three aspects. First, this design can be used as the theoretical model for testing hypotheses about the impact of self-questioning mechanism on the performance of information requirement specification. Second, this design provides a theoretical basis for training novice information analysts to use self-questioning to improve their performance of requirement specification. In addition, the self-questioning mechanism can also facilitate the transition of information analysts from novice to expert. Third and finally, this design provides a theoretical basis for the development of computer-aided software engineering (CASE) tools. On the basis of this research, how to guide novice information analysts to ask right questions to themselves at local and global levels of problem statements is an important issue for the research in CASE tools.

REFERENCES

Adelson, B. & Soloway, E. (1985). The role of domain experience in software design," IEEE Transactions On Software Engineering, se-11 (11), 1351- 1360.

Agarwal, R., & Tanniru, M. R. (1990). Knowledge acquisition using structured interviewing: an empirical investigation. Journal of Management Information Systems, 7(1), 123-140.

Allwood, C. M. (1986). Novices on the computer: a review of the literature. International Journal of Man-Machine Studies, 25, 633-658.

Batra, D. & Davis, J. G. (1992). Conceptual data modeling in database design: similarities and differences between expert and novice designers. International Journal of Man-Machine Studies, 37, 83-101.

Batra, D. & Sein, M. K. (1994). Improving conceptual database design through feedback. International Journal of Human-Computer Studies, 40, 653-676.

Chin, C. & Chia, L. G. (2004). Problem-based learning: Using students' questions to drive knowledge construction. Science Education, 88(5), 707-727.

Collins Cobuild English Dictionary (1995). Harper Collins Publishers. Couger, J. D., Colter, M. A. & Knapp, R. W. (1982). Advanced System Development/ Feasibility Techniques. New York: Wiley.

Dardenne, A., Lamsweerde, A. V. & Fickas, S. (1993). Goal-directed requirements acquisition. Science of Computer Programming, 20, 3-50.

Davis, A. M. (1988). A comparison of techniques for the specification of external system behavior. Communications of The ACM, 31 (9), 1098-1115.

Doerr, H.M. & Tripp, J.S. (1999). Understanding how students develop mathematical models. Mathematical Thinking and Learning, 1(3), 231-254.

Dorfman, M. (1990). System and software requirements engineering. In R. H. Thayer & M. Dorfman (Eds.), System and Software Requirements Engineering (pp. 4-16). Los Alamitos, CA: IEEE Computer Society Press.

Fraser, M. D., Kumar, K. & Vaishnavi, V. K. (1991). Informal and formal requirements specification languages: bridging the gap. IEEE Transaction On Software Engineering, 17 (5), 454-465.

Gernsbacher, M. A. (1990). Language Comprehension As Structure Building. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Publishers.

Gentner, D. (1983).Structure-Mapping: A Theoretical Framework for Analogy, Cognitive Science, (7), 155-170.

Gentner, D.& Markman, A. B.(1997). Structure Mapping in Analogy and Similarity, American Psychologist, 52(1), 45-56.

Glaubman, R. & Ofir L. (1997). Effects of self-directed learning, story comprehension, and self-questioning in Kindergarten, The Journal of Educational Research, 90(6), 361-374

Graesser, A. C. (1995). Inference generation and the construction of situation models. In Weaver, C. A., Mannes, S., and Hetcher , C. R. (eds.). Discourse Comprehension. Hillsdale, New Jersey: Lawrence Erbaum Associates, Publishers. 117-139.

Graesser, A. C. & Person, N.K. (1994). Question asking during tutoring. American Educational Research Journal. 31(1), 104-137.

Graesser, A,, Singer, M. & Trabasso, T. (1994). Constructing inferences during text comprehension. Psychological Review, 101 (3), 371-395.

Greenspan, S. J. & Mylopoulos, J. (1982). Capturing More World Knowledge in the Requirements Specification, Proceedings of International Conference in Software Engineering, IEEE Computer Society Press, 225-234.

Guindon, R. & Curtis, B. (1988). Control of Cognitive Process During Software Design: What Tools Are Needed? Proceedings of CHI'88, 263-268.

Guindon, R., Krasner, H. & Curtis, B. (1987). Cognitive process in software design: activities In early, upstream design, in

Human-computer Interaction- INTERACT'97, Bullinger, H. J., and Shackel, B. (eds), North-Holland, 383-388.

Hsia, P., Davis, A. M. & Kung, D. C. (1993). Status report: requirements engineering. IEEE Software, 75-79.

Huang, I. & Burns, J. R. (2000). A cognitive comparison of modeling behavior between novice and expert information analysts. Proceedings of AMCIS Conference, 1316-1322.

Huang, I. & Burns, J. R. (1998). A cognitive model of information requirement analysis on the basis of structure building theory of language. Proceedings of AMCIS conference, 659-661.

King, A. (1992). Comparison of self-questioning, summarizing, and notetaking-review as strategies for learning from lectures, American Educational Research Journal, 29(2), 303-323.

King, A. (1989). Effects of self-questioning training on college students' comprehension of lectures, Contemporary educational Psychology, 14, 366-381.

Kintsch, W. (1988). The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review, 95 (2), 163-182.

Kintsch, W. (1974). The Representation of Meaning in Memory. Hillsdale, New Jersey: Lawrence Erlbaum Associates, Publishers.

Kletzien, S.B. & Bednar, M.R. (1988). A framework for reader autonomy: an integrated perspective. Journal of Reading, 32(1), 30-33.

Komatsu, L. K. (1992) Recent views of conceptual structure. Psychological Bulletin, 112(3), 500-526.

Koubek, R. J., Salvendy, G., Dunsmor, H. E. & Lebold, W. K. (1989). Cognitive issues in the process of software development: review and reappraisal. International Journal of Man-Machine Studies, 30, 171-191.

Long, D. L., Oppy, B. J. & Seely, M. R. (1997). A "global-coherence" view of event comprehension: Inferential processing as question answering. In van Broek, P. W., Bauer, P. J. & Bourg, T. (eds.) Developmental Spans in Event Comprehension And Representation: Bridging fictional and actual events. Mahwah, New Jersey: Lawrence Erlbaum Associates, Publishers.

Maiden N.A.M. & Sutcliffe, A.G.(1992). Exploiting reusable specification through analogy. Communications of the ACM, 35(4), 55-64.

Martin, J. (1989). Information Engineering, Book I: Introduction. Prentice-Hall Corporation.

McKoon, G. & Ratcliff, R. (1992). Inference during reading. Psychological Review, 99 (3), 440-466.

McMenamin, S.M. & Palmer, J.F. (1984). Essential Systems Analysis. Yourdon Press.

Nolan, T.E. (1991). Self-questioning and prediction: Combining metacognitive strategies. Journal of reading, 35(2), 132-138.

Ortony, A. (1978). Remembering, understanding, and representation. Cognitive Science, 2, 53-69.

Osman, M.E. & Hannapin, M.J. (1994). Effects of advance questioning and prior knowledge on science learning. Journal of Educational Research, 88(1), 5-13.

Ozgungor, S. & Guthrie, J.T. (2004). Interactions among elaborative interrogation, knowledge, and interest in the process of constructing knowledge from text, Journal of Educational Psychology, 96(3),

Roman, G. (1985). A taxonomy of current issues in requirements engineering. IEEE Computer, 14-22.

Rosch, E., Mervis, C. B., Gary, W. D., Johnson, D. M. & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology, 8, 382-439.

Satzinger, J.W., Jackson, R. B. & Burd, S. D. (2002). Systems Analysis and Design. Second Edition, Course Technology.

Scharer, L. (1981). Pinpointing Requirements, Datamation, April, 17-22.

Schenk, K. D., Vitalari, N. P. & Davis, K. S. (1998). Differences between novice and expert systems analysts: what do we know and what do we do? Journal of Management Information Systems, 15 (1), 9-50.

Shemer, I. (1987). System analysis: a systematic analysis of a conceptual model. Communications of the ACM, 30 (6), 506-512.

Singer, H. & Donlan, D. (1982). Active comprehension: problem-solving schema with question generation for comprehension of complex short stories. Reading Research Quarterly 27, 166-186.

Sutcliffe, A. & Maiden, N. (1992). Analysing the novice analyst: cognitive model in software engineering. International Journal of Man-Machine Studies, 36, 719-740.

Swanson, P.N. (1998). Teaching effective comprehension strategies to students with learning and reading disabilities. Intervention in School of Clinic, 33(4), 209-219.

Trabasso, T. (1989). Causal representation of narratives. Reading Psychology, 10, 67-83.

Vessey, I. & Conger, S. A. (1993). Learning to specify information requirements: the relationship between application and methodology. Journal of Management Information Systems, 10 (2), 177-201.

Vessey, I. & Conger, S. A. (1994). Requirement specification: learning object, process, and data methodologies. Communications of The ACM, 37 (5), 102-113.

Vitalari, N. P. & Dickson, G. W. (1983). Problem solving for effective system analysis: an experimental exploration. Communications of the ACM, 26 (11), 948-956.

Wieringa, R. (1998). A survey of structured and object-oriented software specification methods and techniques. ACM Computing Survey, 30(4), 459-527.

Wong, B.Y.L. (1985). Self-questioning instructional research: a review. Review of Educational Research, 55(2), 227-268.

I-Lin Huang, Langston University

Table 1: The Generic Self-asking Questions for Data Flow Diagram

What are the external entities in the problem statement?

What are the data flows in the problem statement?

What are the processes in the problem statement?

What are the data stores in the problem statement?

Table 2: The Definitions of the Four Important Constructs in Data
Flow Diagrams (Satzinger, Jackson & Burd, 2002)

Construct Definition

External entity External entities are sources or destinations of
 data flows which are outside the system under
 development. They represents users or other systems

Process Processes transform inputs into outputs. They are
 the only active elements in data flow diagrams.

Data flow Data flows are composite data items flowing from
 an element to a process (input dataflow) or from a
 process to an element (output dataflow).

Data store Data at rest. Data are stored for later.
 Corresponding to entities in Entity-relationship
 Model. Often implemented by databases

Table 3: Self-asking Questions Based on the Basic Concepts from
Order Processing System

Are there people processing orders?

Is there an order file?

Is there a customer file?

Are there customers outside of this system?

Are there orders flowing in the system from the customers?

Table 4: Legal Conceptual Connections for Local Coherence

 Data Moved To

 External
 Data Flow Entity Process Data Store

Data External Entity A
Moving Process B C D
From Data Store E

Table 5: Local Coherence-driven Questions for Data Flow Diagram
Modeling

Construct Question

External A-structure: (external entity--request--process)
Entity What is the request of the external entity?
 To what process?
 B-structure: (process--report--external entity)
 What is the report for the external entity? Provided
 by what process?

Process A-structure: (external entity--request--process)
 What initiates the process? From which external entity?
 B-structure: (process--report--external entity)
 What report is generated by this process?
 To what external entity?
 C-structure: (process--intermediate--process)
 What immediate result is generated by this process?
 To what process?
 D-structure: (process--save--data store)
 What data store save the result of the process?
 E-structure: (data store--retrieve--process)
 What data store provides information for the process?
 What data are provided?

Data store D-structure: (process--save--data store)
 What data in the data store are saved?
 From what process?
 E-structure: (data store--retrieve--process)
 What data in this data store are retrieved?
 By what process?

data flow A-structure: (external entity--request--process)
 What external entity is the sender of the request?
 To what process?
 B-structure: (process--report--external entity)
 What external entity is the receiver of the report?
 From what process?
 C-structure: (process--intermediate result--process)
 What process generates the immediate result?
 What process will do the further processing?
 D-structure: (process--save--data store)
 What is the data store saving the data?
 What is the process generating the data?
 E-structure: (data store--retrieve--process)
 What is the process retrieving the data?
 From which data store?

Table 6: Global Coherence-driven Questions for Event Partition
Structure

Construct Question

Event What is the event?
Source What is the source?
Destination What are the destinations?
Trigger What is the trigger?
Response What are the responses?
Process What is the process?
Data Store What are the data stores?
Update Does the process update any data stores?
Use Does the process use any data stores?

Table 7: Answer the Questions Derived from the Principle of
Basic Concepts

Question Answer

Are there people performing an John, the order clerk
 activity of processing order?
Is there an order file? Need further investigation
Is there a customer file? Need further investigation
Are there customers outside Yes
 of this system?
Are there orders flowing in the system Yes
 from the customers?

Table 8: Answer the Questions Derived from Analogical Similarities

Question Answer

Is there an activity similar to Order processing by
 checking out books done by librarian? John, the order clerk
Is there a file similar to a Need further
 check-out books file? investigation
Is there a file similar to Need further
 student file? investigation
Are there people similar to students Yes, customers
 outside of this system?
Are there similar dataflows from Yes, orders
 the students?

Table 9: Answer the Questions Derived from Global Coherence

Question Answer

What is the event? Order entry
What is the source? Customer
What are the destinations? None
What is the trigger? Order
What are the responses? None
What is the process? Process order
What are the data stores? Customer file, Order file
Does the process update any data stores? Order file
Does the process use any data stores? Customer file

Table 10: Answer the Questions Derived from A-structure: (external
entity--request--process)

Question Answer

What is the request of the external Order
 entity?
To what process? Order processing by
 John, the order clerk

Table 11: Answer Generic Questions for Data Flow Diagram

What are the external entities in the Customer, Order Clerk
 problem statement?
What are the data flows in the problem Order
 statement?
What are the processes in the problem Process order
 statement?
What are the data stores in the problem ?
 statement?