Tagide

This post was prompted by a Facebook interaction regarding Dijktra’s famous “GO TO Statement Considered Harmful” article, a letter he sent to the editor of CACM back in 1968. Seen through the lens of currently accepted research reporting practices, Dijsktra’s article reads like a technical rant, something that we might find in today’s blogs, but that have been essentially abolished from peer-reviewed research publications. Here’s a person stating his strong opinion about a language construct that he dislikes, starting with a strong premise about the negative effects of that construct without presenting any data whatsoever supporting that premise! As far as anyone can tell, the premise was based entirely in his own experience and intuition — which happened to go against opinions of other famous Computer Scientists at the time like Knuth. Granted, this was just a letter to the editor, not a full-blown technical article. However, the influence of this letter is unquestionable! Following Dijsktra’s rage against GO TOs, practically all programming languages that came after it avoided that construct, and the ones that didn’t, work hard at hiding it from programmers, sort of like “oops! sorry! there’s goto, but try to avoid it, ok?” (Also, this letter was single-handedly responsible for starting the meme “X considered harmful” that has been going on in CS since then, although credit for that meme goes, not to Dijkstra, but to the CACM editor, who spiced the title a little bit from its original “A case against the GO TO statement.”)

Whether GO TO is harmful or not, is besides the point. In the process of writing my upcoming book, I spent a considerable amount of time over the past year looking through old CS literature. It’s fascinating! The topic of this post is the evolution of methods in Computer Science research for the past 60 years, and the changing ways by which ideas are considered well-argued for.

Continue reading →