October 31, 2011, 12:15 PM — Last week’s blog “Your technology skills have a two year half-life and 6 ways to stay current” was heavily read and hotly debated. Thank you to all and here is my reply.
Thanks for all your emails and blog comments on last week’s column. I love people’s feedback to my columns as it helps me look at things from different points of view. Quite frankly, I agreed with almost all that you have written.
To begin, the original question asked by the reader was related to the customization of a software package, rather than programming in a base technology, like C and PHP, which has a different set of factors.
As a software package example, in the late 1980s I was a DBA working on Oracle Version 3. Whereas Oracle is obviously very alive and well, none of the database tools or processes I worked on then still exist. This was pre PL/SQL and query optimization had to be done by hand because the SQL engine didn’t work all that well. That said, Oracle as a software package then, is totally different than Oracle the software package now. Sure, data normalization is the same, but virtually all of my exact Oracle 3.0 skill set at that time is now outdated.
That said, is the knowledge I had then, transferable to the latest version of Oracle now? Yes, I believe it is. But to the original point, my Oracle Version 3 skills were much more marketable in 1988 than they are today. Potential employers looking for a DBA would consider my knowledge to be old, and thus less desirable than someone with more current knowledge.
Even C, however, evolved into C++. Imagine if today you only know K&R C and didn’t understand function overloading, structured programming, base classes, and other related advances such as ODBC, XML, and multi-threading.
Also, C now has more competition than it did then, same as COBOL which I also programmed professionally. There are less companies today programming in C and COBOL because of Java. I’m not saying that Java is better or worse than C or COBOL, I’m just saying that they were used more widely in 1990 than they are today, thus further reducing the marketability of older technologies. If fact, I loved Borland Turbo-C, which of course today is far less marketable than the Microsoft .NET tool set.