Nothing wrong with the languages you mentioned; they do what they do best. COBOL and Fortran are both fantastic and nothing yet truly replaces them. COBOL is accurate out to 38 digits, which is amazing for such an "old" language. Banks still like their COBOL and it just works. No stupid "libraries and/or frameworks of the month" to worry about, it compiles cleanly and every time, and it's easy to write and maintain. Modern doesn't always mean best. I still write tons of stuff in sh. If it's under 100 lines, it's shell, awk, or similar. COBOL and Fortran still have tons of life left in them. Every time someone undertakes a massive project to replace COBOL in the financial sector (usually driven by a latte-sipping hipster and his recently graduated ilk), it goes pear shaped. We still use hammers. They work.
From my admittedly cursory look at D, I find nothing persuasive enough to get me to take it on and learn it when the chances of my ever using it are slim. I guess I'm just old school. It seems that every week there is a new language, framework, etc., and most of them are really not doing anything radically different. As a back end guy, I see most of the churn happens in the web dev world. I prefer old, stable, and very little churn, hence my continued love for COBOL in particular. I also like C++ and Python, but there is something fantastic about COBOL, sh, and awk, which are oft-used. I'm getting less and less enthusiastic about systems stuff now that I'm getting older and more into writing useful tools to help my guys. Call it an easy exit...
