SQL

From APIDesign

Revision as of 19:00, 23 August 2010

SQL is a DSL used to manipulate and especially query relational databases. The mathematical roots of SQL are based on relational algebra. SQL seems to be one of the most often used programming languages in software engineering, althrough these days it may be slightly over its zenith.

History

The SQL was on rise in the seventies and eighties driven by the (for example Sybase) vision to unify all data available and represent them in one big database. These data were collected from various other (non-relational) databases which many businesses used to have at that time (and very likely still continue to maintain even now). Over the night a snapshot of such data was converted into various SQL tables. Next morning, the management could then effectively perform various queries and found out more precisely than ever what is the state of the business.

In this scenario the SQL database is not the primary source of the data. It is more or less reporting tool. There is nothing wrong on that! Compared to previous age where one had to ask a programmer to write code to traverse various graphs and nets of data to find out an answer to a question. Then one had to wait for a few weeks until the question is coded and executed to get answer when it did not matter at all. With SQL queries everything is much more simplified. All the data are available to a simple SELECT. Coding of a query does not require orchestration of a various languages and systems. Cluelessness par excellence! Even non-programmers can perform or modify queries and they can see the results almost instantly (sometimes), much faster than in the previous era.

SQL is well suited for this reporting job. I can imagine companies using it becoming much more effective. The increase of effectiveness was so visible, that SQL become synonymy for a database. Since then people started to use SQL databases more and more. So finally the SQL database become not just a view over primary data, but the data itself.

All the data are stored in SQL, instantly available, ready for various selects. One does not have to wait 24 hours for a report, one can get the report immediately. Heaven isn't it? Or did we trade this flexibility for something essential?

Internet Age

TBD: Why SQL is not well suited for the internet age...

<comments> I'd go with the idea that MDX are far superior for actual data mining and data-model construction. SQL does not natively represent complex data structures well. That said, SQL will probably remain around for quite some time -- it is easy, known, and fast. </comments>