I keep my eyes open for things that pertain to the data world. I jot them down in a list of possible blog topics. This morning I took a look at the ever-growing list and decided to take ten items and build a “what’s on my mind” list. It’s definitely a random mish-mash of topics. I hope you enjoy.
- Are relational databases going away soon?
Definitely not. VSAM is still going strong at almost fifty years. Big data and small data are hot topics but have their specific use. Relational databases will continue to carry the workload of day-to-day IT business applications for many years. Most certainly, it will give way to and evolve into a leaner, faster technology. - When will big data tide come to the masses?
Big data is mission critical for many web-centric corporations. Companies dabbling in complex analytics also have a need for big data. Many large to medium sized companies are dipping their toes in the water. Widespread adoption will speed up as the technology matures, enterprise strength solutions evolve, and the cost comes down. Small IT shops will benefit from the advance in technology resulting from big data adoption. It could be 2-5 years. - Is there a place for nulls in the database?
There certainly is. When you are counting or averaging a column, a null means no response. A blank can very well be another valid response. This is very true in analytics and business intelligence. I assume a value will be present. I have to have good justification and requirements to define a column as null. - Why is metadata still so hard to manage?
I was introduced to metadata management with my first warehouse in 1995. I found no solutions back then and am still waiting. This is a fundamentally difficult problem to solve. The database structures of tools and databases constantly change making metadata harvesting difficult. Metadata never seems to get the traction and respect it needs. Dedicated resources are needed to manage metadata. Although more automated today, it is far from the push of a button solution. - What’s the value of reverse engineering a data model?
I see value in reverse engineering databases into data models. These models give the blueprint of the data and visually show data and relationships. Companies may not have the resources to do full lifecycle modeling. Any data model is a good tool to help people understand their data. It is much easier to refactor a database and expand functionality with a data model of the database on hand. - Why don’t people understand data models?
Data modeling is meant to simplify the understanding of a database design. It still has a jargon of its own. Everyone approaches technical subject matter with a varying degree of comfort and knowledge. Boxes and line are just shapes to many. There is still a need for education to understand the nature of a data model. - Is abbreviating names an unnecessary relic from the past?
I do hate abbreviations. Years ago they were a necessary evil when field names were limited to 8-18 or so characters. Database technology has now enabled very long names. I abbreviate only when necessary. Analytics necessitate abbreviations when the column names can approach the length of a long sentence. I abbreviate according to the database abbreviation scheme when working with legacy databases. - Why isn’t there an easier way to compare data models?
Here’s my beef with my data modeling tool. They have enabled the comparison of just about all characteristics of the data model and database. The typical complete compare should be a basic comparison of table and column attributes and metadata. I would love out-of-the-box wizard and make the “compare the world” option just that—an option. - Will we ever reach the point when we have enough data?
Next question please. Data will exponentially grow. The management of this data will become more complex. The quality of the data will be even more important to know. The security of the data will be harder to manage. None of these facts should deter the storage of the data. I trust technology will evolve to manage these massive data stores. - Why won’t data architects use social media?
Change and unwillingness to change. I have been interviewing interns over the past few weeks. They gave me insight into this. All spoke of the power of the Internet to solve problems, find solutions, network with colleagues and grow their knowledge. Google, meet-ups and tech web sites are their go-to places. I do not see this in data architects. It’s not the world they grew up in and a place they are uncomfortable living in.
Tom Bilcze