After claiming that most programmers just can’t program, and actually addressing most of the problems to the lack of passion of people who decide to start a career as a programmer, I would also like to express my point of view on a tightly related subject: what can be done to improve the situation? The problem that I was trying to bring up in the spotlights, is that a lot of people just start (or wish to start) a career in the IT for no particular reasons. Those are the ones who don’t love and don’t loathe programming, and they just see it as something that pays their bills. Well, maybe the first question that I should address, actually is: why is this bad? Sure there are so many jobs which don’t require passion at all, and people just do them because a job is just a job, and don’t really care. In my opinion, being a programmer is different.
There are many people, especially the ones who sit high in the hierarchy of a company, who see programmers as the last and least important step of a ladder. They often think that programming is quite of an automated and repetitive task, and it could basically be done by anyone, with just a little training. Unsurprisingly, this seems to be the opinion of most common people, who ignore what programming really is. I wouldn’t want to discriminate among different types of programming, or different programming languages, but it’s obvious to me that programming, to some extent, actually can become an automated an repetitive task. That’s quite the minority of cases, though, so I will simply ignore them, and focus on the rest.
As anybody who’s a programmer knows, programming is a highly creative task, that requires good imagination and great problem solving skills. Everybody else might just see it as “typing stuff on a computer”, and believe me, there’s a whole lot of educated people who think that programming is a monkey matter. Hence the term “code monkey”. This term has historically been abused a lot, by even programmers themselves. A “code monkey” is said to perform a programming task so easy that even a monkey could do, as the image suggests. There are two truths about this phenomenon: first of all, luckily, programming requires far more skills than it’s usually believed; secondly, and sadly, the majority of people just ignore it.
The problem with lousy programmer is kind of similar to a medal: it’s double faced. You could actually call it a dog trying to bite its own tail: as programming is believed to be an easier and easier task, more programmers are needed; as more and more programmers are needed, more people will jump on the field; as more and more people try to become programmers, the lousier the average quality of programmers gets. Unfortunately, what average non-programming people miss to understand is that although it doesn’t really take a hard training to become a lousy programmer, it takes a damn hard one to excel in the art of programming. Moreover, most people just lack the innate logic mechanisms that make you a potential programmers. Such mechanisms are developed in your mind when you’re very young, and it’s quite rare to develop them after your twenty-somethings. With this, though, I’m not denying that there are a lot of people who actually do develop those mechanisms in advanced age. I’m just trying to think of the big numbers, here.
So, getting to the point, what went wrong and how can it be fixed? I don’t think it would be wise to say that what’s wrong is that there’s too much need of programmers, ergo the average quality was inevitably doomed to lower and lower over the time. I rather think that the problem is with education. Of course I can’t speak for all the universities and colleges in the world, but I can at least try and speak for the one I’ve known personally, or through people who have studied there. It seems that, as more and more people apply to Computer Science or related departments, the easier it gets to get in (sorry for the pun), and to get through with it, i.e. to graduate.
I know this happens most likely in any other faculties, but seeing that there are people who have been studying CS for three or more years, and still can’t get through the most simple concepts, just doesn’t seem right to me. Yesterday night, I was sitting in an IRC channel about the C programming language, when somebody joined in and asked:
“I just started studying structures in C, and I don’t get them. Can anyone explain to me what’s the use for them?”
Ok, I don’t really think there’s anything wrong in not getting the point of C structures right away, but after a little chatting, it turned out that the guy was in his second year of Computer Science, and this was the second time he took the C class. Still that wouldn’t be a reason of hatred, of course (not that I have any hatred), but after another small while it turned out that the guy didn’t like programming at all, but he just got himself into it because he applied to CS since he liked to “fiddle around with computers”.
What’s really needed, in my opinion, is a harder and less tolerant educational system, that would be more selective, rather than pushing everyone forward. People that find out to be really not made for it, should just give up and move their focus on something less.
I’m actually very well aware that a lot of programming work, nowadays, is not really rocket science, still this doesn’t mean that it should be done by completely unqualified people. If what Jeff Atwood says in his post about programmers who can’t program is true, and that is that 199 out of 200 applicants (not programmers, applicants) can’t write any code whatsoever, than it obviously means that something is wrong. Looking at the numbers provided by Joel Spolsky, it looks like a lot of these basically incompetent people are going to end up working on an actual programming job, and maybe their code will end up on The Daily WTF (Paula, are you there?).
Unfortunately, the education is not the only one to blame. No matter how much education will improve, there will always be unqualified people who are going to apply for jobs that require a lot of skills, and in the end the odds will help them, so they’ll manage to get a job as a programmer. Is it so bad, considering that it’s most likely not going to be any critical position, and the only ones that will be damaged will be the owners of the company that hired them? Well, the point is that this is not true. There’s someone else who gets damaged, in this scenario. I’m talking about the community out there, the good programmers, who find themselves competing with newbies who’re happy to earn peanuts. The salaries keep going down, and customers are not able to distinguish a good job from a good one.
In a comment on the previous post of mine about this subject, Hoowie Goodell really gets a great point with this paragraph:
“There has been a great effort to industrialize programming, too. Again, there are many good features, and it’s a field I’m interested in. Building a large program requires a structured approach. Language design, libraries, programming frameworks and IDEs can and should incorporate as much existing human knowledge as possible: computer science, domain knowledge, solid pre-written code and human interface principles. (Check out Thomas Greene’s “Cognitive Dimensions of Notations” for some of the latter: I think of how programming tools fail to use them on a daily basis!)”
In a way, this suggests that the whole system is not ready yet, as it’s indeed years and years behind several other engineering fields, and that’s a good reason, probably, to explain why it’s so easy to fail at being a good programmer. Let’s just try to get some insightful inspection points, in order to build better generation of programmers:
-
Better education. The whole higher educational system should be improved in several way. Worldwide. Nowadays, it looks to me that in many countries graduation is just a direct consequence of applying to an University. Unfortunately, this kind of problem must be addressed on a country-basis, to properly identify the specific issues, but still the options that I would like to consider are worth mentioning. It all comes down to a single point: there should be less tolerance towards people that don’t learn. The thresholds for succeeding in a course should be raised to greater difficulty. Current models of testing should be seriously revised, so to ensure that students that really didn’t understand the subject are not going to make it.
-
Better tools. Are we trying to make programming just like a factory chain or are we not? If we are, as it seems nowadays, then the tools are not ready yet to second our intentions. Programming is too error prone and too time-consuming.
-
Better process. Software process that doesn’t conform to some standards, say ISO-9000 (sorry if it’s inappropriate, I’m not an expert on this kind of standards), shouldn’t be allowed to sell. Quality insurance committees should be taken more seriously as being part of the process. This might be against all principles of liberalism, I know, as bad software, you may say, will not sell anyway. I know many bad software that did sell well, for greatly different reasons than its (non) good quality.
-
Better judgment when hiring. I’m not going to try to teach you how to run your company, nor how to hire your crew. But sometimes really crazy thing happen (again, is Paula around?). A very interesting post by Joel Spolsky (I’m sorry, I can’t find it anymore: does anybody know the link?) talks about only hiring “A”-people, where “A” means top class. If you’re ever hiring a “B”-person, he’s quite likely to hire a “C”-person, someday. After that, it’s chaos. I recommend anyone not to lower their canons of perfections. Here’s another great article by Joel, about hiring good developers, I recommend it.
Concluding, improving the quality of programmers seems really to be a tough issue, and the whole thing depends on so many factors that tracking a precise problem is impossible. Cultural and technical difficulties arise all the time, and getting clues is hard. I’ve tried to get around the problem and give some insightful opinions: what do you people think?