by Giselle N
The article I read this week was about the importance of data modeling. The author says that because there are a lot of DBMS tools available for easily creating tables, many developers skip the modeling process and go straight to building them, which could be done instantly. However, because some of these tables are created fast and sometimes sloppy, they usually result in not being sufficient enough for big “industrial strength solutions.” Many developers do not have a problem going back to the tables they created and taking time to check their work for errors. The author points out that many users become confused and do not see a need for modeling when developers can easily just redo large sections of code. Still, the author says that usually not everything is as clear as the developer thinks it is and “ the meaning of each data item is important enough to document” beforehand as it would be done in a data model. Also, if not enough time is spent in data modeling, more time will be spent in post implementation, with programmers going back to fix mistakes that probably could have been avoided given a proper design of the database in the first place. The author goes on to say that in many steady businesses “data reuse in new and enhanced applications should become something modeled effortlessly” and that is why these businesses see no need for data modeling and develop this bad habit. However, when new data is introduced, modeling can take a lot of time seeing as it has not been practiced and therefore to clearly understand it, the meanings and use of the data should be documented. In the end, the author says the only time when it is probably appropriate not to create a data model and therefore skip ahead to make a table is when only one person is developing it, no data is being shared, and no other application is going to be using it, other than that a data model should always be created.
I agree with the author and believe data modeling is a very important step in designing a database. For example, someone wouldn’t build a house without first having a blueprint. It’s important to know and understand the context of the data and I believe that could only be done by first examining it through a model instead of jumping ahead to the next step. Like the author says, that is why a lot of time is wasted on fixing errors because developers do not clearly know the meanings and use of the data and will never get that opportunity unless they analyze it through first creating a model.
Schraml, T. (2009). Yes, data modeling matters. Database Trends and Applications, 23(4), 31-31. Retrieved from http://search.proquest.com/docview/199361104?accountid=10357