The use of Clarion for this case study was entirely accidental. It resulted a number of years ago from a consulting contract with major insurance company where Whitemarsh was hired to assist in the data modeling and database management aspects of a business information system project. The project involved building a "data warehouse" from a large collection of stove-pipe-based databases and applications.
Whitemarsh's contracted work was to focus entirely on data model design. Nothing more. However that changed.
The first task was to build an overall data model for the data warehouse. This was done by building the data model into Clarion's data model creation tool that results in a database dictionary (DCT file). Once that was done, SQL scripts were generated. As the project progressed, the DCT-based tables, columns, and relationships evolved. So too did the continually regenerated SQL data model scripts.
What changed was caused by the inability of the data warehouse application development team to create and thoroughly validate the overall data warehouse application architecture model in advance of detailed design and coding.
The stove-pipe communities were insistent on validing and verifying that the data warehouse application development effort would both be cost effective and successful. They didn't just want to see whitepapers and diagrams. No, they wanted to see real operational demonstrations. Ah, like how dare they impose such a requirement....ahem.
As a long-standing Clarion user, I interjected a recommendation that Clarion could be used to directly assist in addressing that "unreasonable?" stove-pipe community requirement.
Before that was really allowed, however, I had to demonstrate that using this Clarion tool set from "Florida" would produce something of "actual" value.
So, I undertook a Friday through Monday effort to create an operational model of one stove pipe extraction, data transformation, and data loading into one of the functional areas of the data warehouse.
Now to Clarion developers, this is a simple, no-brainer task. To traditional main-frame developers not only is this--one weekend task--a mountain too high, it's a Mount Everest! Not suprising to Clarion readers, the operational model was demonstrated that next Tuesday.
Only after the prototype's generation was I allowed to explain just how Clarion would help with the overall data warehouse application model. Clarion provided two assists. The first was to create the individual data extraction application models from each of the "stove pipes." The second was the construction of the overall transformation and database loading models for the data warehouse.
Clarion was used to generate these two assists. Once generated and tuned to the first draft of the appropriate application models, they were presented--in execution form--to the user community from the "stove pipe" communities. Presented were real, executing models, not paper-driven and/or diagram based models.
Reaction from the user communities was immediate and powerefully userful. The Clarion-based DCT data models were adjusted. The Clarion-based application models were adjusted. Another round of presentations to the "stove pipe" communities was undertaken.
Each of these cycles took a few days to a week. What? a few days to a week? Yes, and that is because of Clarion's database and application generation capability. Once these cycles were completed, the Clarion-built models of both the database and the data warehouse data extraction, transformations and loading essentially became the architecture and design for the overall project.
Subsequent to these cycles, the overall Data Warehouse was constructed using main-frame tool sets. Debugging focused on coding errors not architecture and design errors. Not having architecture and design errors was--of course--the direct result of using Clarion. For that and for that alone, the procurement, deployment, and use of Clarion is worth many, many more times its cost.
Bottom line was that in this non-Clarion project, Clarion increased productivity, increased quality, lower risk and lower cost.