August 1998        Issue: 4

Journal of Conceptual Modeling
www.inconcept.com/jcm

Is it still a requirement if the subject matter expert
 didn't tell the analyst?
by Dr. John K. Sharp

Most information technology projects start with an incomplete and possibly inconsistent problem statement generated by the subject matter expert or an involved manager. The subject matter expert and the manger know a lot about the process(es) being analyzed, but they do not know what is needed by the information analyst to provide for the creation of the supporting application. Usually this knowledge is randomly transferred to the analyst through a series of sessions that may only serve to train the analyst in the general aspects of the subject area. The hope is that the analyst can use this general knowledge to create the needed design that supports the existing problem. If the end result of this effort is an information system that meets all of the requirements and is delivered on time and within budget, then everyone involved can share in the success. All to often, problems appear during development or even after the application has been delivered. Responsibility for the problem is hard to assign, especially when the analyst feels that the subject matter expert did not provide the involved requirements.

Assigning responsibility for the various levels of project failures (not meeting requirements, cost overruns, and/or schedule extensions) is a challenge. Unfortunately, accountability for the problem is now more likely to be a political decision and not a technical one. Since the information technology group provides support to groups that make money for the business, the information technology group is in a lower political position. If accountability is handled as a political decision the information technology group consistently looses. The corporate flagellation over the application failures often results in the threat or direct action to go outside for contractors to create applications because the internal staff cannot do it.

A solution to enable accountability to be a more technical issue is to use a deterministic procedure (such as Natural Language Modeling (NLM)) to structure the analysis. The NLM procedure ensures that the requirements gathering effort is focused and that all pertinent questions will be asked of the subject matter expert. Responsibility is shared by both the subject matter expert through providing answers to simple questions and by the analyst through using a deterministic analysis procedure.

Example problem:

The example problem involves a movie marquee. The challenge is to be able to properly specify all of the business rules that are to be enforced about this information.

Figure 1: Movie marquee from military theater.

The NLM procedure starts with the subject matter expert generating a sentence from the example. In this case an initial sentence may be:

The movie Jaws is showing in theater 1 at 1000 hour.

An identifier for each object is now established by evaluating an existence sentence for each variable.

Object: Movie

The movie Jaws exists.

The movie <MovieTitle> exists.

The matrix question is: "Given that the true sentence "The movie Jaws exists." is it allowed for another movie title [such as Mad Max] to exist such that the sentence "The movie Mad Max exists." is also true?"

Does the movie title Jaws exactly identify a movie? Yes

Object: Theater

Theater 1 exists.

Theater <TheaterNumber> exists.

Does the theater number 1 exactly identify a theater? Yes

Object: Time

1000 hour exists.

<Hour> hour exists.

Does the 1000 hour exactly identify a time? No

[A different movie can be shown in the same theater at 1000 hour on different days. Thus, 1000 hour does not identify a time.]

What other object is needed to identify time? Day

A new sentence is generated to identify Time that includes both hour and day.

Object: Time

1000 hour on Monday exists.

<Hour> hour on <DayOfWeek> exists.

Does the 1000 hour on Monday exactly identify a time? Yes

[Assuming that the marquee planning is only for a week at a time.]

Object: Day

Monday exists.

<DayOfWeek> exists.

Does the Monday day of week exactly identify a Day? Yes

Now that all of the objects have been properly identified the initial sentence is analyzed after it is extended to include day.

The movie Jaws is showing in theater 1 at 1000 hour on Monday.

The movie <MovieTitle> is showing in theater <TheaterNumber> at <Hour> hour on <DayOfWeek>.

The Yes and No answer vector for this sentence will yield two identifiers in this sentence by following the NLM procedure for a Mixed Yes and No vector. The results are presented in figure 2.

Figure 2: Table for the Movie marquee example.

Four identifiers have been found for objects in this sentence and two keys exist in the required table. Key number 1 enforces the rule that at a certain time a movie is shown only once. This constraint is the result of a business rule where the company chooses not to pay the additional lease fee to get a second copy of the same movie. Key number 2 enforces the rule that only one movie can be shown at a time in a theater.

The subject matter expert is only required to specify the initial sentence from a populated example. There is no need for the subject matter expert to independently define all of the rules for showing movies or provide valid instances of fact types that are to be managed. The NLM procedure provides for the proper identification of the objects and the population restrictions for the fact types.

Now assuming that the NLM procedure was not used and the resulting model did not contain the rule that only one copy of a movie is leased at a time, who is responsible for the missing requirement? The analyst should be responsible because direct questions could be asked of the subject matter expert that would establish this rule. The subject matter expert is not always responsible if a rule is not provided! Why not get the model right the first time with NLM and not re-work the application or try to assign blame?

Happy Modeling!

Dr. John Sharp is the founder and principal consultant for Sharp Informatics.Before starting Sharp Informatics in 1997 he was employed by Sandia National Laboratories in Albuquerque, NM for 18 years. While at Sandia he held staff and management positions in all areas of information technology, including analysis, design, implementation, maintenance, information architecture, data administration, and information technology research. He has worked closely with Prof. Shir Nijssen of The Netherlands to improve the NIAM analysis methodology. Dr. Sharp is the creator of the first information analysis procedure known to be mathematically precise.This procedure reformulates the usual (imprecise and inaccurate) statements and examples from a subject area into verified fact types. The output of this productivity enhancing process (a set of information requirements) is compatible with all the latest and most productive database application creation tools. John is the editor of the international standard for conceptual schemas. He has co-chaired two international conferences on natural language modeling and he has presented numerous papers and seminars at professional conferences.

Contact information:

Dr. John Sharp
Sharp Informatics
1604 Vassar SE
Albuquerque, NM 87106
sharp@sharp-informatics.com
505-243-1498
fax 505-248-0345
http://www.sharp-informatics.com

© Copyright, 1998-2004 InConcept (Information Conceptual Modeling, Inc.) All Rights Reserved. Privacy Statement.
ISSN: 1533-3825