Here’s a dirty little about data management: It’s about art as well as science. In this post, I discuss how many people often mistakenly focus on the “science” of things while minimizing the art piece.
Considerations
Any good developer knows that there are many ways to skin a cat. For even something requiring high a degree of precision, there are often many options. Of course, some are better than others. When I develop data management tools, I often have a number of different alternatives with respect to moving and manipulating data, including the use of:
- Temp tables
- Queries
- Batch processes
- Export/import routines
While the MIKE2.0 framework provides for extensive best practices at a general level, there is a blizzard of individual decisions to make. Major development questions often include:
- Should a data transfer process occur automatically or is there a need for someone to approve an action mid-stream?
- What’s the right application for end users to enter and maintain data?
- Are there any audit or regulatory concerns?
- How technical are those being asked to maintain the data?
- What type of safeguards exist so my clients won’t have to call me with minor questions?
- How can I lock down the data–and the magic behind the scenes–so people can’t break things, dunintentionally or otherwise?
The answers to these questions drive my development efforts and basic philosophy for data management. For example, if I build an ETL tool for the IT department, it’s reasonable to assume that employees’ expertise will allow them to make some changes, especially if I document things well. I can probably automate many things and let SQL dance. However, if the same tool is built for gengerally less technical folks, there exists the very real danger that someone might break something. I typically err on the side of simplicity and more manual data management but it’s appropriate for that audience.
In the classic movie This is Spinal Tap, there’s an amazing line: There’s a fine line between clever and stupid. That quote need not be confined to the music industry. The same aphorism holds true for data management. Different folks in organizations have different levels of understanding and proficiency of all things data. I have seen repeatedly throughout my career the perils of overreaching; sometimes, really neat methods of data management are lost on end users, resulting in confusion, frustration, and dysfunction.
Feedback
What say you?
Read more at MIKE2.0: The Open Source Standard for Information Management
