Irreal

Many of you know that I’m a big fan of AWK. After Lisp and C, it’s my go to language, especially for scripts that involve reading a bunch of lines and performing some sort of regex-mediated action on them. Its language is basically C-like but there are some anomalies that can seem mysterious.

Volodymyr Gubarkov has a project to provide an InteliJ plug in for AWK and part of that requires parsing AWK syntax so he has a good handle on those anomalies and why they exist. He discusses those peculiarities in his post, AWK technical notes.

It’s an interesting post that explains many things that you probably didn’t know about AWK. In a way, it’s getting into the weeds but it also increases your understanding of the way things work and allows you to write better AWK.

For example, one thing I didn’t know was that although it’s illegal to have a space between a user defined function and its argument list, it is legal to have a space between a built in function and its argument list. The reason for that involves string concatenation. Take a look at Gubarkov’s post for the details.

The post mentions some other interesting anomalies in the AWK syntax and the reasons for them. If you use AWK, it’s definitely worth reading the post if only to be aware of the gotchas waiting for the unwary.

Even before I learned Lisp, I really loved coding with recursive algorithms. When I decided to learn Lisp, I did so through Paul Graham’s ANSI Common Lisp, which meant, of course, that recursive programming became part of my programming DNA.

A significant part of my Lisp journey involved Scheme where recursion is the go to mechanism for most looping constructs. That’s because the Scheme standard mandates tail call optimization so that recursion is no more expensive in time or memory than any other looping mechanism. Not all languages or even all compilers for a given language support tail call optimization so a prejudice against recursion has emerged. The problem, according to this prejudice, is that recursion can lead to stack space exhaustion.

Tim Bradshaw examines this prejudice and dismisses it for the silliness it is. According to Bradshaw, there’s a deep-seated belief among programmers that stack space is a scarce, valuable resource while heap space is essentially free and unlimited. This leads to a preference for implementing inherently recursive procedures in an interactive way that involves managing the stack explicitly in the heap rather than using recursion and managing the stack implicitly.

As Bradshaw points out, the stack and the heap are both just memory and neither is more expensive or scarce than the other. There’s no reason to restrict stack space while allowing large heap spaces other than convention and it’s time for us to reexamine this prejudice. It’s hard to argue; Bradshaw has a point.

There’s a substantial body of lore associated with Dennis Ritchie and his work. One of the minor stories involves his PhD or lack of it. The most common story is that Harvard refused to award Ritchie his degree because Ritchie objected to paying a thesis binding fee. I’ve always accepted that story but it never made sense to me. Having gone through the process and pain of writing a PhD thesis myself, I can’t imagine giving up the prize over a triviality. But, of course, Ritchie was unique in many ways.

That story—right or wrong—is the least of the mysteries around his thesis. A recent paper by Brailsford, Kernighan, and William A. Ritchie examines the mystery of how Ritchie “typeset” his thesis. The scare quotes on typeset are because this was in 1968 and the only realistic way producing hard copy for a thesis was with a typewriter. But the extant copies of the thesis, which was highly mathematical, has aspects which seem impossible to produce with a typewriter. It is in any event a tribute to Ritchie’s attention to detail.

Brailsford, of course, has a video about the mystery but to really understand what an astounding feat accomplishing the typesetting was, you need to read the paper. Note, especially, how Ritchie centered the Roman numerals on his lists. Brailsford and Kernighan are experts on this sort of thing and even they can’t figure it out.

William Ritchie is Dennis’s brother and has a website on the mysteries of his brother’s thesis. It’s also worth looking at.

To really understand what’s going on and why the physical presentation of the thesis is so astounding, you definitely should read the paper. It’s 10 pages but very interesting. The paper and the mysteries it examines tickles the intersection of several of my interests and I really enjoyed it. If you’re interested in this sort of thing, perhaps you will too.