53 Years of Relational Databases

If it weren't for him, there would be no modern databases. It is the 53st anniversary of the publication of his article that changed the rules of the game. Edgar Frank Codd is one of the fathers of SQL, and his name is one of the first to be mentioned in the history of databases. Or maybe even the history of computer development in general? See what we owe to Dr. Codd and why his work is still important.

Pssst! At the bottom of this article, you will find an interesting promotion for our SQL courses!

It was June 1970. Things happened then. The world looked at two brave Russian astronauts who spent 18 days aboard the Soyuz 9, while The Beatles and Ray Stevens reigned in the charts.

An issue of the Communications of the ACM, a noble journal of the Association for Computing Machinery, had been handed over to mathematicians and scientists. What could they read about? Among the excellent articles on mathematics, statistics, and computer science in vol. 13 no. 6, one could learn about using graphs for scheduling meetings to reduce conflicts (Joseph E. Grimes), methods for accurate floating-point summation (Peter Linz), or even interesting thoughts on automatic parsing for content analysis (Frederick J. Damerau).

On page 377, "A Relational Model of Data for Large Shared Data Banks" appeared. It was a collection of bold and innovative theses. The author of this new approach, which consequently changed the world of databases, was Edgar F. Codd of IBM Research Laboratory in San Jose, California.

Let’s find out what was groundbreaking about Codd’s article and why we still consider it important 53 years later.

Source: seas.upenn.edu

Who Was Edgar Frank Codd?

Edgar Frank (Ted) Codd was an English mathematician and computer scientist. During World War II, he was a pilot in the Royal Air Force. After the fighting ended, he returned to Oxford, where he finished studying mathematics at the prestigious university.

Interestingly, he first tried his hand at chemistry. Had he remained true to his original dreams, the IT world would look completely different today.

Codd came to the US in 1959 and settled in New York, where he worked as a programmer at IBM. He participated, among others, in the work on the Selective Sequence Electronic Calculator, the world's first electromechanical computer from IBM.

After a few minor projects, the STRETCH project brought him more success. It was the foundation of the IBM 7090 mainframe technology. As a programmer, Codd wanted more. He received a PhD in computer science from the University of Michigan.

In 1967, Codd transferred to the IBM Research Laboratory in San Jose, California. There, he got serious about the concept of relational databases. Fun fact: not everyone in his company supported his thesis. His thinking was so different from what was then dominant in IBM, so much so that the first company to use his model was ... Oracle!

Codd won many awards for his work. The most important was the Turing Award, which he received in 1981. This prize, often called the Nobel Prize for Computing, is the highest distinction and recognition in the world of computer science.

But let's go back to SQL, databases, and the article.

Databases in the 1960s

To understand why Codd’s article was so influential, it’s important to see what databases looked like at that time. The dominant database models in the 1960s were the network database model and the hierarchical database model.

In a network database, the records are linked to one another forming a graph. In a hierarchical database, the records are linked forming a tree.

To retrieve data from a network or hierarchical database, the whole graph or tree needs to be traversed starting from the root node. To get data out of this database, you need to write a program that traverses the network of records on the disk.

Essentially, you need to be a programmer to work with network and hierarchical databases and navigate down the tree/graph of records to find the data needed. This navigation metaphor was very common in the 1960s. Charles Bachman, a pioneer of the network model, even gave a speech titled "The Programmer as Navigator" when he received a Turing Award for his work in 1973.

The problem with the network and hierarchical models was that they were very inflexible. Any change in the structure of data required changes in the programs that got the data out. Yet these databases were incredibly successful in the 1960s.

The Road to Relational Databases

Codd saw this inflexibility as a problem. He proposed a new database model called relational model, based on the mathematical notion of relation.

In today’s terminology, Codd’s relations are called tables. A table has a name and a set of named columns holding data of the same type. The data in tables are stored in rows, with a value for each column of the table. It is relatively easy to modify a database in this model, be it adding a new table, adding a new column to a table, removing a table, or removing a column.

An important idea proposed in Codd’s model was the query language suggested in his article, which is now called relational algebra. He came up with a set of operations on tables that would extract data out of a relational database. These operations could be composed to build more complex expressions and do more complex computations. This was not SQL that we use today; however, if you read Codd’s paper, you will see that all the operations we have in SQL are already there, just using a different syntax.

If you want to know even more about the mathematical foundations of Codd’s model, check out "Theory and History: Why are Relational Databases ‘Relational’?"

What had been pages of code in a network or hierarchical database became a simple expression in Codd’s query language. Donald Chamberlin, one of the creators of SQL recalls:

[...] exposure to the relational data model at Codd’s research symposium was a revelation. For the first time, we could see how a query that would require a complex program in the DBTG language could be reduced to a few simple lines using one of Codd’s relational languages.

As we mentioned before, Codd’s ideas were not immediately successful. His paper circulated internally within IBM for a few months before he finally published it in June 1970. Even then, IBM wasn’t immediately interested in working on Codd’s ideas. At the time, they had IMS, a very successful hierarchical database. They didn't want to undermine their revenue from IMS by building a competing product. IMS is still in use today, which shows how successful it was back then.

In 1974, IBM started System R, a research project to explore Codd’s ideas in practice. For some reason, Codd himself was not on the System R team. Many design decisions for System R, such as transaction design and certain algorithm choices, influenced many relational databases down the road. The project also created SQL, a query language loosely based on Codd’s ideas and with the syntax we know today. The two people responsible for creating SQL were Don Chamberlin and Ray Boyce.

Relational Databases in the Following Years

Relational databases quickly became a commercial success. DB2 became the flagship database product for IBM. Oracle database was released in 1979 by Relational Software, now Oracle Corporation. Ingres, the predecessor of the PostgreSQL database, followed.

Along with the development of databases, SQL (Structured Query Language) also gained importance. Engineers and computer scientists needed a simple language to deal with databases – something that would help you extract data, process it, change it, and perform other operations.

The commercial success of databases meant that their users were not only engineers but also those who were less advanced. The query language had to be simple. And it has remained so until today.

SQL became the standard query language for relational databases and the ISO/ANSI standard in late 1980s. Many years have passed since then, and yet it is still the language for working with databases. As most of the market research shows, SQL has a dominant position and is not going anywhere!

Even the most modern and technologically advanced companies in the IT industry still use SQL. It is a language worth learning and having on your resume. Why? It just pays off. Here, you will find Types of Database Jobs. Choose one and start a new, better, and better-paid career. Not sure if SQL is for you? Find out in Who Should Learn SQL?

Relational databases dominated the world in the 1980s and the 1990s and continue to be an important data technology today. This would not have been possible had it not been for the foundations set by Codd and his article.

It is safe to say that if it were not for him, there would be no PostgreSQL, MS SQL Server, or other SQL dialects today. There would be no relational databases and thus a huge part of the modern IT and business.

Codd's Anniversary Promotion

Codd's article is just 11 pages long. What's more, it's written in plain language that is understandable even if you are not a mathematician or computer scientist. Read his paper and see how to change the world with words.

This article you are reading right now probably will not change the face of modern IT. But we really have wise experts and scientists at LearnSQL.com. And this article can change your life, if we can convince you to learn SQL!

It is a chance for a better job and higher earnings. Just browse the job offers and see if there is any demand for people who know SQL. Spoiler alert: it's huge!

How can you get our Codd’s Anniversary Promotion?

1. Create a free LearnSQL.com account or log in if you already have one.
2. Save $410 on our All Forever Package.
3. The promotion will be valid until July 6.

Celebrate 53 years of relational databases with us! Join the world of SQL today!