I just finished reading the book Coders at Work by Peter Seibel. This wonderful book is filled with interviews with prominent programmers: Joe Armstrong, Simon Peyton Jones, Donald Knuth, among others. My review isn't going to be the standard book review. Instead I'm going to talk about some of the specific ideas I picked up while reading the book and discuss those ideas, rather than the book itself. For those of you who are looking for more of a standard review, all I can really say is that Coders at Work is well worth the time for a programmer or a computer historian/hobbyist. Unfortunately for the casual non-computer-field reader, the interview topics (and the interviewees) assume a certain level of prior knowledge. What makes this book so good is that different people will be able to walk away from it with different ideas, depending on their areas of expertise and interest. As the industry grows and changes, we definitely do things based on what the people before us have already discovered - Seibel’s interview subjects are the fascinating programmers whose ideas and hard work make possible all the cool things we get to play with today.
In the interview with Douglas Crockford there was a lot of talk about code readings. If you’re not sure what those are (I wasn’t), from the way Douglas describes it they are regular meetings where programmers read the code base together. In my mind it sounds like someone is giving a presentation but instead of it being PowerPoint slides, it’s code. I picture the presenter going line by line and describing what is going on and why he or she decided to do things a particular way. These sound like an excellent idea to me. In my experience so far, the code review process seems to have been relegated to using Reviewboard or other similar website-based tools. I'm not trying to insult these tools, but they remove the human factor of the code review and hinder the discussion that can happen over a particular piece of code. I understand that programmer time is expensive and managers should be careful how much of that time is spent in meetings. But I think these meetings could benefit everyone and will have a return on investment (that is the time spent) in ways that cannot be quickly seen. One benefit is that everyone who attends will come out with a basic idea of what the code looks like within a given project, even if they’re not working on that project directly. So if things shift around (as they inevitably do) and a new person is moved into that project, he or she won’t be starting from scratch. Even just a little bit of knowledge might help a programmer get over the initial shock of having to dive into a new project.
Programmers need Empathy:
In the interview with Joshua Bloch there was a quick chat about programmer personalities. There were a couple of choice quotes regarding programmer empathy for the users, like, “...intelligence is not a scalar quantity; it’s a vector quantity. And if you lack empathy or emotional intelligence, then you shouldn’t be designing APIs or GUIs or languages.” and “What we’re doing is an aesthetic pursuit. It involves craftsmanship as well as mathematics and it involves people skills and prose skills—all of these things that we don’t necessarily think of as engineering but without which I don’t think you’ll ever be a really good engineer.”
I bring these quotes up because I agree with them, and because they speak of issues that are largely ignored in the engineering community (again, in my experience). I've sure we’ve all run into programmers that only care about the technical aspects of the project or the correctness of their code and do not concern themselves for how the user might interact with the system. For example, the company I work for designed a web-app that included separate “Kill” and “Terminate” buttons. This was a complete UI fail that could have been avoided, had the programmer (we’ll call him/her “Terry”) had empathy for the user. As a programmer, I understand the difference between SIGKILL and SIGTERM, but those buttons would be confusing to even the more-advanced-than-average user (and go far beyond failing the grandmother test). When this was brought up as a ticket for “Terry” to fix, “Terry” argued that there was nothing technically incorrect about the problem and closed the ticket “Won't Fix.”
One thing I noticed in a lot of the interviews was that Peter asked the interviewee what he or she looked for in a potential new hire. What I found fascinating is that none of them said, “I like to ask really hard questions.” or “I ask really obscure questions about the programming language they use to see how well they know it.” They mostly talked about looking for someone who had passion for the craft.
As someone who is frequently on both sides of the hiring table, I know that the current hiring process in Silicon Valley doesn't normally ask questions about WHO is being hired. Questions about hobbies or interests are rare. It's mostly just technical questions, as if the person who is applying for the job is another computer whose sole purpose is to solve a company's problems. All the emphasis seems to be focused on the technical skills of the interviewee. While it's important to know if someone can “do the job,” tech skills alone do not make a person a great programmer - a good one, yes, but I don’t get the impression that companies are settling for “good” anymore. Passion is what gives someone the tenacity to work through the obstacles and become a great programmer, and if we as interviewers insist on focusing only on hard skills, we risk cheating ourselves out of the truly great engineers.
Peter also asked every single interviewee about Donald Knuth's books and his “literate programming” style of writing code. I will admit that I have not read his books, but I want to. I'm not sure what this literate Programming style is, but I'm curious enough about it to learn, and see if there are some aspects of it I can incorporate into my own coding and documentation style. Next step, pick up an electronic version of the book, “Literate Programming” and give it a good read.
I fully enjoyed Coders at Work. I took away much more from the interviews than I talked about here, and I’m sure there’s even more to learn. I'm planning on rereading it again in a couple of years; as I grow as a programmer I will have a different perspective on the craft, and might find a whole new or different set of lessons from what I got this time. If not, at the very least I’ll have the pleasure of reading it again.