A good article in American Scientist about different cultures of coding. When i started (and I should confess, I'm a History/English/Writing major) in AeroE, I would say RPI was training me more as a computational scientist: modeling data and computational problems. Where I landed roughly ten years later was definitely the software development culture. Every now and then I run into a developer who tells me a variation of the following statement below - that we get lots of good software developers who know the languages, but that don't know their "basics". Admittedly, in my experience, we hire for the former, not the latter. Perhaps in part because they create copious amounts of business-requested code (features), even if it's not based entirely in the fundamentals. It's a bit of a conundrum and we seem to juggle it a bit by having a separate R&D group as well as some groups that balance on that line between software development and computational science and even computer science that can focus on the underlying efficiency and methodology rather than delivering external user features.
"Programmers today are intensely partisan in their choices of programming languages, yet interest in the underlying principles seems to have waned. Two years ago I attended a lunch-table talk by a young graduate student who had turned away from humanities and business studies to take up a new life designing software. She had fallen in love with coding, and she spoke eloquently of its attractions and rewards. But she also took a swipe at the traditional computer science curriculum. “No one cares much about LR(1) parsers anymore,” she said, referring to one of the classic tools of language processing. The remark saddened me because the theory of parsing is a thing of beauty. At the very least it is a historical landmark that no one should pass by without stopping to read the plaque. But, as Edith Wharton wrote, “Life has a way of overgrowing its achievements as well as its ruins.”"
No comments:
Post a Comment