Database Design for Mere Mortals: An Interview with Mike Hernandez
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
Ken Getz: What was your background as a developer before you wrote your book, Database Design for Mere Mortals?
Mike Hernandez: Well, let me begin by saying that my story begins back in the days of the PC database era, starting around 1986. I worked at EggHead Discount Software from 1986-1990, initially selling software and then doing technical training in DOS (remember DOS?), Word, Excel, XYWrite, Lotus 123, Quattro Pro, dBase, Paradox and R:BASE. I started doing some database development on the side and finally decided to open my own consulting firm in 1990. I specialized in doing R:BASE apps and eventually moved to Microsoft Access. I was doing quite well. Aside from writing articles for various trade publications and teaching database classes, I was a featured speaker at various technical conferences across the US and in England. This is where I got to know the real luminaries of the industry, such as you, John Viescas, Paul Litwin, Stephen Forte, Brian Randall, et al.
Ken: What made you think that a book about database design would be useful to the general developer public?
Mike: One thing folks should know about me is that I have only have a high-school education and diploma. I went to the university for about a year, but it was only to chase a young lady, hoping to win her heart. (Didn’t work out and I wasted all that time.) I believe one of my strengths has always been the ability to learn things easily and quickly. If you can show me how to do something just once, and you explain it clearly and thoroughly, I can then do it myself from that point on.
When I started getting into my database consulting practice, I really wanted to know how to design a proper, effective and efficient database structure so that I could then work on more complex projects (which would result in more money). So I went to every bookstore I could find in search of a book that would teach me how to design a database properly. I did find a lot of books on database design, but they were all very technical. Aside from the fact that the authors typically used database software for their examples that I was not using, they discussed concepts that were quite foreign to me such as normalization, functional and transitive dependencies, subsets, cardinality, optionality, and joins. I often thought, “Whoa! Isn’t there any database book that can describe the design process in plain English?!” There were no such books, unfortunately, so I just planted my feet firmly in the ground and decided to read what was available. I read works from authors such as C.J. Date, David Kroenke, Barbara Von Halle, Thomas Connolly, and Toby J. Teorey. I even read Dr. E.F. Codd’s book. (For those of you too young to know or remember, Dr. Codd was the father of the relational model.) It was a long, tedious process, but I learned quite a lot. I even found that I actually understood the concepts behind set theory and predicate logic. (Although, don’t ask me anything about it nowadays.) This surprised me because math was one of my weaker subjects back in high school.
I really didn’t mind going through these books (it was kind of fun, in a weird sort of way), but I still lamented the fact that there wasn’t a good book explaining the logical design process. I eventually got so good at what I was doing that I was asked to write articles on database design and to do presentations on this subject at various conferences. I also started teaching classes on design at a number of local companies. It was at the classes and conferences where I learned just how much people needed a good, simple and clear book on design. Here was an audience yearning and waiting for someone to write a book that could help them.
There was a “PC Database Summit” here in Bellevue back in 1995, and there was a young lady named Kathleen Tibbetts representing Addison-Wesley (now part of Pearson Education) that was searching for potential authors. As you might expect, she contacted the A-list developers and authors first, such as you, John Viescas, Paul Litwin, Roger Jennings, Brian Randall, Michael Groh, and Alan Simpson. All of you actually had book projects in progress at the time, so she asked if you knew anyone that might be a good prospect. You, John, Paul and Brian seemed to have the same answer. It went something like this (I’m paraphrasing): “You should go talk to Mike Hernandez. He’s definitely got a database book in him. Ask him about his idea for writing a book on database design. That’s all he talks about and he’s really good at it. Tell him it’s time to put up or shut up. And please let him do it! Perhaps he’ll stop whining so much about the fact there aren’t any good easy to understand design books out in the market.”
Kathleen did indeed come to see me and we went out to have lunch and discuss my ideas. (I learned about what you guys said at this meeting.) Well, she told me who she was and what she did for Addison-Wesley, and I then told her who I was and I pitched the book to her. After a three-and-a-half-hour lunch, she said she’d get in touch the following week so we could sign a book deal. And thus Database Design for Mere Mortals was born. This was to be a book that could help anyone learn how to design a properly structured database, regardless of their previous background. It seems to have achieved its goal, as it has been out now for 17 years and counting.
Ken: I know I get terrifically frustrated when I see people use Excel as a database, and then they stuff all sorts of data into columns in a seemingly random fashion. How do you explain to people that unless they manage their data in an organized fashion, they can’t retrieve the data they need?
Mike: I simply ask them to find all instances of a specific piece of data and then sort the results in a particular manner. That’s when the concepts of duplicate data and multivalued fields come into play in a very obvious manner. I then explain how a database could handle these tasks more quickly and efficiently, and how they can get even more information from their data if they were managing it in a real database.
Ken: Can you summarize, for a novice developer, what “database design” is, and why they need to know about it?
Mike: Database design is the process of creating and developing a structure that will store and manage data that is important and relevant to you or your organization. It’s important because it’s crucial to the consistency, integrity and accuracy of the data in your database, and enables you to create a structure that supports your information requirements or those of your organization.
Ken: It seems to me that there ought to be tools out there to help people create well-designed databases. Am I dreaming? Are there such tools? If so, what are they and how do I find them?
Mike: Actually, there are quite a few tools that developers can use to create structures and then produce them in a specific database software program. You can find a list of them here: http://www.sqlmag.com/content/content/96845/96845_WebTable_V22.pdf. [PDF File] (NOTE: This is from a 2007 SQL Mag article, but a majority of the companies are still around and still producing their products.)
The problem I have with these products (and one I’ve had since the very beginning of my database career) is that they mix logical design with physical implementation. This could be a very long discussion and debate, so let’s save it for another time and place. Suffice it to say that I believe there should be a clear distinction between the two. The main point to remember about these products is that they are merely the tools you use to create, develop and maintain the physical implementation of your database. You still need to know how to design the logical structure of the database in order to create it in the software program. If you also understand all that is involved in the logical design process, you’ll be able to take full advantage of the software program you’re using to implement your structure.
I’ve always believed you should design the logical structure of your database first and without regard to any RDBMS. By doing so, you’re more likely to design a sound structure because you’ll be focused on your or your organization’s information requirements. Once your design is complete, you can then clearly determine how you should implement the database (single-user application, client/server, web-based, and so on) and which RDBMS you should use to facilitate the implementation. Then you can take advantage of a tool like CA ERwin to create and develop the physical implementation of the database.
Ken: Is there a difference in the way you should design databases for use in client applications as opposed to services?
Mike: This is actually an implementation question, in my opinion. To me, the important thing is to ensure that the database structure supports stated information requirements. Again, once you have the completed structure in hand, you can decide how you need to implement it based on the ultimate business requirements.
Ken: Is there a way to document your database design in some standard way, so ten years down the road, a new developer could review the documentation and understand it?
Mike: I don’t know of any particular standard for documenting a database. There are many ways of doing so and I certainly promote one in my book. It really doesn’t matter what method you use, but it is absolutely essential that you take the time to do the documentation. It will certainly come in handy if something goes wrong and it definitely provides an audit trail for the decisions you or your team make during the design process.
Ken: Once people have mastered the concepts in your book, what then? Where should they go for more information?
Mike: They can start by tackling some of the books on the recommended reading list in my book and then take some database related courses at their local college, university or technical training center. There’s also a vast array of resources you can find on the web. You’ll always find more to learn, especially if you treat this as a true craft and you take pride in your work.
Ken: Is there some attribute that the two of us share that very few people know about? Do you want to share it with the public?
Mike: Actually, there are a few:
- We both came up in the PC database ranks of the 1980s, you working with Ashton Tate’s dBase and me working with arch-rival Microrim’s R:BASE.
- We both eventually moved to Microsoft Access and had a great run with it.
- We both were instructors for AppDev (known back then as Application Developer’s Training Company).
- We’re both musicians (Ken-piano, Mike-guitar).
- We both share the same birthday, although I’m a year older.