Being a hacker doesn't protect you from being outdated

Let's start with a few well-known facts :

• Emacs (or XEmacs) is a slow and bloated piece of junk, a text editor has no business being a mail and news reader.
• Garbage Collection (GC) is for people who find it too difficult to call free(). It is always possible to have nice matching malloc()/free() pairs. If you can't, your design is broken (or use static arrays, or quit programming altogether).
• Object Oriented Programming (OOP) sucks, it makes things more complicated, harder to debug, and produces bloated programs. Plain C can always do the job just fine.
• C++ is an inherently slow and bloated language. Multiple inheritance is wrong, you can always do with single inheritance, and if you really need it, Java interfaces are the right solution. The C++ Standard Template Library (STL) is totally useless, the GLib does exactly the same job and is much lighter. Templates are a bad idea anyway, you can achieve the same functionality with void pointers (or virtual functions if you really have to).
• Lisp is just a language to teach about linked lists and try to do Artificial Intelligence. You can't possibly do anything useful with it.
• Goto's are evil and should never be used. Using a goto means you're too lazy to write a clean, structured program.
• Memory Protection is needed only because people are too lazy to fully debug their code and are writing bad programs.

At this point, chances are that at least a couple of these statements above ring very true to your ears, while you'll find the others totally ludicrous.

Note that none of these are from me directly. They were all picked from mailing-lists, newsgroups, or slashdot discussions, all programming related. While the later two aren't renowed for their high signal-to-noise ratio [1], mailing-lists are usually frequented by rather clueful people.

But still, all these "facts" share two other more important common points :

• They are all false.
• Considering them true means that you're missing a crucial point, that a part of your mind has closed shut, and prevents you from seeing any further, and expanding your technical skills.

What this document DOES NOT talk about

First and foremost : why (or if) any of these facts is true or false. I'm not interested in discussing any of them, please refrain from emailing me about any of them.

Another issue I'm leaving aside is of course personal taste. One can choose to dimiss a technology because he simply doesn't like it, or is not comfortable using it. While this can be a problem in the overall course of a programming career, it's a much less serious one than blind rejection on false grounds.

What this document DOES talk about

The point of this article is to attempt to detail the reasons which can lead a programmer on this most traveled path, and to show that being a hacker (here used in a somewhat wider sense of the term than usual, "programming nerd" would probably be more appropriate) in no way prevents you from following it. Quite the contrary, it can be an incitement.

The Technician's Obsolescence Problem

Anyone who has worked in the IT industry is aware of the "technician's obsolescence" problem. At some point in his career, a technician's skills (and thus himself) become irrelevant. Most, if not all, of what he knows is outdated and has been supplanted by new techniques, usually more efficient.

Usually, the Dilbert Principle has kicked in before this happens, and that technician has become a manager. The kind which you try not to discuss technical issues with, because you know he just won't understand.

The Most Usual Cause : growing old

As you would expect, the main cause of obsolescence is of course age. It's hard to keep up with the furious pace of Information Technology, and having to learn a new thing to replace one you've mastered (or, even worse, haven't even mastered yet) is always tedious. This does not necessarily mean "old age", you can be caught by this phenomenon from time to time, on a given subject, while still being a beginner [2]. The "Learning muscle" requires training to stay in shape, just as any other one.

Adding to this problem is the fact that, among the monthly crop of IT novelties, it can get pretty hard to know the weed from the chaff, but you can safely assume that 1% is worth a look and the remaining 99% is short-lasting buzz. Hence the understandable tendancy to dismiss it altogether, if only to wait and see what will actually last.

One would think that, in the case of a hacker, the sheer passion for everything technical discards the lack of learning will. This is indeed true most of the time. A hacker won't have a problem with picking up something new, because he likes it and he's good at it.

Other Causes

There are several other factors, among which :

• Being content with one's skills : You've reached the level of competence which you were aiming at, consciously or not, and are happily resting on your (maybe years old) laurels. You just aren't interested in learning anything new.
• Not having any external challenges : A variant of the previous one. Having acquired some sort of good reputation among your peers (be they work colleagues or fellow hackers), you're convinced that you're good. Nobody around you is better than you are, or has skills you don't have.
• Being a few against many : Thinking that you know of have understood something that everybody else is missing (and bragging about it). This is most common among users of a "minority OS" (e.g. anything but Windows), and even more in the hacker culture as opposed to the software industry (although such divergence among hackers are quite common). Using an alternative technology often gives you a feeling of pride, because you feel being part of an enlightened minority, hence an elite. The fewer people agree with you, the truer your opinion feels to you. The old cliche that the greater number tends to gravitate toward what sucks is a most supportive argument.
• Clinging to a tradition : "This is how we've been doing it all the time". In the case of the industry, this is a derivative of the first factor, e.g. age. Working in a Hi-Tech environment doesn't prevent you from getting hard to shake habits. In Hackerdom, the "tradition" feeling is much stronger, because you care for what you're doing and how you are doing it (especially as opposed to how the software industry is doing it, see previous point). The craft is actually a part of the culture.

Hackers can become outdated too

All these traps are very easy to fall in, and it should be obvious that hackers are ideal targets for them, much more than people for whom programming is just a job. The fact that among the hacker's traits are ego, and the preference of elegant, and usually less "easy", solutions over dumb ones should also be considered as potential weaknesses rather than strengths, or as an hindrance to keep an open mind and improve one's skills, rather than an enticement to do so.

Experience shows that the only winning strategy in development is pragmatism : Use what works best. Linus Torvalds has directed the Linux kernel the exact same way so far, choosing a solution on proven technical merits rather than on an epidermic feeling. A more far-fetched example, but still relevant in a way, is Microsoft. The company got big in a large part because even when they had invested a huge budget on some project and the market wouldn't follow, they never had any second thought at scrapping it all, do an about face and follow the market (e.g. MSN vs. the Internet, Blackbird vs. Java - please, I'm not interested in discussing about other methods they may be using :-).

Of course, always doing so is next to impossible. We all have pre-conceived ideas, and admitting that you're wrong is always difficult. Also, as noted before, personal taste does play an important part. When it's about using a tool or a language, you will always be more efficient using a technique you know and like but which, strictly speaking, is less effective than another one which you don't know and dislike. Always falling for the new buzzword is even worse.

However, I hope that this short discussion may hint some people into trying to revise some too radical opinions they may have, to check if they aren't in fact "missing something", and being too quick to leave aside some tools which they might find invaluable further down the road. In particular, young Linux users migrating from Windows should keep in mind that the very impulse which drove them out of the Wintel rut can nail them dead on in another one.

[1] : the term "slashdot reader" has more or less acquired a derogatory and demeaning sense, see The Slashdot Phenomenon by Elliot Lee, and the the relevant bits of Eric Raymond's "Understand my job, please".

[2] : On the other hand, I've had the pleasure to work with an IBMer who, at the age of 50, was still curious and interested in new stuff, and eager to use it whenever he had the chance. There is no fatality here.