Lists Home |
Date Index |
Michael Champion wrote:
> I wonder if there are analogous cases in the history of technology
> (ahem, besides SGML?) where a set of good ideas got out of control.
How about SQL and the relational data model? Back in the old days we
had a variety of approaches to data storage -- however, today, most of the
new folk in the industry have never even seen anything that wasn't
relational with a SQL-like query syntax. We seem to have forgotten that
while you can do just about anything you want with a relational database, it
often makes sense to use a network or object database instead. Sometimes a
simple indexed file is best and other times, a random access record
structure worked fine (remember them? Access record #2...). The "good idea"
which was Codd's work on the relational model has "gotten out of control" in
that it has completely swept away most of the alternative (and sometimes
superior) models from consideration.
The monopoly of the relational model has often forced people to
twist their applications to fit the model rather than choosing the DBMS that
matches their need. In other cases, we're using complex relational systems
when much simpler methods would be more efficient.
Note: Before anyone flames me for being anti-relational, please
understand that I am no such thing.