I recently spent several months in a very-SAS shop (see my recent Meditation on the automated recovery of program understanding), and I was quite amazed at what they were able to make SAS do. I also noticed how SAS Institute had in-effect woven a procedural programming language right into their heretofore “PROC and DATA-step oriented” system. It quickly occurred to me that they were using a lot of quasi-SAS code to do what a little bit of Perl code could do, and that the resulting (Perl) code might be a good bit easier to maintain (take vastly fewer resources, etc). My position was heard and understood, and, technically speaking, was not denied. The decision was made to stick with the Devil that they already knew, and I did not press the point further.
Nor did the point need to be pressed. Their objective was to get $work done, and, in their own peculiar, sassy way, they did so. Thankfully, I did not hear much “Perl bashing,” because most of the people I was talking to had worked somewhat with Perl in their distant past. Also, I noticed that several of them were surprised at what Perl could do now, versus the last time that they had encountered it. So, those seeds I planted might yet bear some fruit.