LinLog

Author's note: This one from the archives was written back in March 2008. This is something I've seen a couple of more times since then. A new developer, usually fairly young, comes in and wants to rewrite everything using whatever technology or method he's most familiar with. It seems to be not so much a problem with learning the new codebase as it is developer hubris. "My chosen way is obviously the right way, so clearly that's how we should be doing things." The people who do this don't seem to stop and think about whether their chosen method or tool might not be the best choice in the current situation, or whether implementing it might just be more trouble than it's worth. Perhaps it's just a lack of experience - they haven't had enough rewrites go south to realize that this isn't always a good idea. Either way, enjoy the rant!

Here's a little tid-bit for you young, aspiring developers out there: When you start a new job, it's usually a good idea to wait until your second week before you suggest rewriting the entire existing codebase.

I say this because we recently hired a new developer. He's obviously very bright and knows his stuff, but his first suggestion, on his first day, was that we should rewrite our entire codebase from a custom PHP framework to Symfony. After all, it would probably only take a couple of months.

Now, I don't know much about Symfony, but I have nothing against against it. And, in all fairness, our current framework is a bit of a mess. So I'm not dead-set against the very notion of switching. However, the current framework has been tailored to our needs and it is perfectly adequate. I'm sure that Symfony is very nice, but it's not immediately obvious what compelling advantages it would offer. After all, we already have an application framework that does what we need and which all three of the established developers are very familiar with. Given that we already have a full plate as far as projects go (hence hiring a new developer), it's hard to imagine how you could make a business case for spending 2 or 3 months learning Symfony and converting the whole site to use it. And if the new guy things that we're going to convert in parallel with adding features to the existing codebase (which he actually mentioned), he's out of his mind.

The thing that really bothers me about this, though, is that the new guy suggested this on his very first day. As far as I know, he hasn't even written a line of code yet and he's already writing up the business case for switching to Symfony. Even though he's barely looked at our codebase and so couldn't possibly have any idea what kind of effort this would entail.

So, to all the developers out there, next time you start a job on an established development team, at least take the time to learn what technology they're using and why before you suggest replacing it. Do a project or two with the current stack. Talk to the other developers about the history of the codebase and the company culture. Pushing to replace major technologies before you've even written a single line of code just makes you look like a tool.

Author's note: In yet another installment of "From the Archives", here is a little entry that I wrote up back in September of 2007, but never bothered to publish. It's a little dated, but sadly it still has the ring of truth to it. Of course, there has been some improvement. A lot of hardware doesn't come with a CD anymore - they just give you a pamphlet with a link to the crappy software instead. But anyway, I thought this entry was based on an interesting quote, so here it is.

On an episode of Hanselminutes a month ago, Scott and Carl spoke with Jeff Atwood of Coding Horror fame about the "ultimate developer rig" he built for Scott. During the course of discussion, Jeff uttered what is, undoubtedly, the best quote about software developers that I've ever heard.

Nobody hates software more than software developers.

Not only does it sum up the feeling of true software geeks everywhere, but it also offers a sad commentary on the state of the computer hardware business.

This quote was uttered in the context of a discussion about "mouse software," i.e. those useless CDs full of crap-ware that get thrown in the package with even fairly standard mice. They typically include drivers, a control panel applet, some kind of tool that runs in the system tray, and probably a few other things.

As Jeff and Carl rightly pointed out, the reason they, myself, and many other programmers hate such software is that it is, quite simply, useless. Nobody needs to install the software that comes with a standard wheel mouse because their operating system (no matter what it is) already supports it. In fact, Carl went so far as to say (half-kidding, I hope) that if he caught any of his employees installing "mouse software," they'd be fired on the spot. Any half-competent programmer or IT person should know better.

Sadly, this phenomenon isn't just limited to mice. USB media devices seem to be the worst offenders. Cameras and MP3 players in particular always seem to come with some kind of software that's supposed to help you move files onto or off of the device. Of course, as it's always third-rate crap-ware, as the vendor makes their money on the hardware and just throws in the software as an after-thought. But the worst part is that there's no reason you should even need it, because any decent MP3 player or camera really ought to "just work" as a USB mass storage device (with the notable exception being the iPod, which is decent, if overrated).

So really, it's not that software developers hate software, it's that we hate crappy, unnecessary software. That's probably because we know what we need and can easily judge the value of a program. We can tell when all a program does is the digital equivalent of busy-work. Being handed a CD of code that doesn't do anything useful feels like being condescended to. That angers us.