by Joe C
The normal way of using Entity Relationship Model is to classify objects by personally evaluating them, then assigning them as entities or attributes. Therefore, because of this manual task, creating an ERM, even by professionals, will take a considerable amount of time. In addition, things in the field constantly change but the data modelings usually won’t have enough time to interview field workers for updated information. People have tried to solve this by creating automated systems to assign entities and attributes, but many times these systems do not function perfectly or efficiently. To fix that, the paper proposes a new solution with various improvements. It asks for an input in natural language form, in which the system will only look for key entities while filtering out redundant phrases. In addition, while CASE and ERWin have been useful in the past, this new system believes there are still flaws. In the end, the new process is defined into four steps. 1: Have the worker write out a Standard Business Description, 2: Schematize the SBD into an initial Bill of Materials, 3: Refine the initial BOM, 4: Convert the refined BOM into the Entity Relationship Model
Basically this article talks a lot about the current tools people use in building ERMs and why they are uneffective, then it goes on to propose its own solutions. In essence, the solution sounds good – make it easier for regular workers to use, make the database less redundant, make tasks more automated, etc. but it’s just the fact of putting the theory to the test that might be the restricting factor. I mean, improvements in every field without sacrifices sounds great! It just has to be actually created first or else it’s just all talk. In addition, this article contains a lot of specific examples which I do not understand yet, but they may solve a lot of small problems. Personally I think that as someone goes through with this project, many new problems will arise and so it may take a long time to finish. Of course, the benefits are great, we just don’t know if this is a project worth realizing/if it’s possible or if it will be abadoned mid-development.
This article relates a lot on about what we are currently studying, such as on Entity-Relationship Modeling, entities, attributes, etc. I personally think that it’s a good idea in theory, but it’s one that will never become used in actual practice. It may be one of those spur of the moment ideas where a thinker/scientist gets an idea he thinks is amazing, then provides a general solution along with many small fixes he thinks of on the spot for small errors. Basically a nice fantasy to make an inventive run on, but then will later disappear.
Kim, N., Lee, D., & Moon, S. (2007). Behavior-inductive data modeling for enterprise information systems. The Journal of Computer Information Systems, 48(1), 105-105-116. Retrieved from http://search.proquest.com/docview/232574060?accountid=10357