Thursday, May 28, 2009

Beware Of Mumpsimus

In first year of university, I took a class that used the wonderful book Becoming A Master Student. The book had many stories that have stayed with me over the years.

One is the story of mumpsimus. You can read more at the link, but the gist is that a monk used a Latin word, mumpsimus, for decades before discovering it was bogus. Upon the revelation, the monk replied that he didn't care: he had been using it for 40 years and so it would it remain.

In my freshman year, I was an uncomfortable chemistry major. Little did I know that I would be writing about that story many years, later vis-a-vis computer science. (The story itself may go back centuries!)

I mentioned mumpsimus in the comments of the last post. I had speculated on using protected methods over private methods. The feedback was unified in its rejection of the idea, yet I mused that I would probably continue my style. Ouch. That is mumpsimus indeed: after seeking opinions, I launched a heroic denial of the responses and continued on my merry way. Nice.

Another example is the defeat of mumpsimus. Years ago, in C++, I would define class members like so:

class Person {
int m_id;
string m_name;
int m_age;

When I turned to Java, I held on to that style -- for about 1 day. When I saw what IDEs could do for automatically generating getters and setters, it became obvious that the prefix had to go. Thankfully, logic carried the day over mumpsimus.

The Upshot

The point here is that we have a name for a particular mindset, and a reminder that it is important to re-evaluate ideas with an honest understanding of our biases.

No comments: