Monthly Archives: May 2010

I’ve found a great post on Alan’s blog

I found this great post on Alan Skorkin’s blog. Click here to go to the original post on his blog, or read it below! 🙂

The post is called:

 

Did Your Boss Thank You For Coding Yourself to Death?

Programmers love to work long hours! There I said it, c’mon admit it, your job/boss doesn’t make you do it, we do it to ourselves. Alright, I’ll concede, maybe not all programmers love long hours, but surely with the amount of overtime that is prevalent in this industry at least half of us must love it. Right?

I can hear the excuses already. "No, no that’s not it, we just love working with cool tech and don’t want to leave a problem unsolved. It is actually a good thing it’s what makes us awesome!"

I say – you’re not seeing the forest for the trees. Here is some perspective, you’re not doing this for yourself, you’re doing it for "the man". Admittedly he might be a nice man, but you don’t owe him slavish commitment. Here is even more perspective, how often are you actually playing with interesting problems and cool tech and how many times are you churning out code desperately trying to get something delivered and meet some arbitrary deadline that someone has assigned to you? But hey, you’re a business savvy developer, you’re helping the company succeed, your manager has explained the financial situation to you – it has to be done, we’re relying on you. Well, unless that same manager is right there with you, entertaining you with amusing anecdotes at 2 am, his words are worthless.

Let me tell you a story that a friend once told me. It is about a brilliant developer – lets call him John.

John was a superstar, a one in a million programmer. He had an uncanny ability to understand and write code and was 20 times more productive than anyone else. One day the company got a big contract that needed a fast turn-around. The client sent a massive spec document – to everyone’s dismay. John came to the rescue, he took the spec home and noone heard from him for 3 days. When he came back to work, he looked like hell, but he had gone through the whole spec and had an outline of the solution already finished. Except for one bit which was impossible to implement, though the spec said otherwise – even the client didn’t realise this, but John picked it up. Amazing!

When I first heard that story, I was pretty impressed, my first question was, "So, where is this guy now?". To which my friend replied – "He is dead, too much hard living!". Too much hard coding would be more like it. Kinda takes the wind out of that story a little bit – John was in his early 30s.

Programmers take a perverse pleasure from sharing death-march war stories. Even when we do it with disgust, it is a disgust tinged with pride – daring our peers to do "better". But it is a bit like those guys who wear their pants so low you can see ALL of their underwear or the people who take up smoking for the "trendy image". They and their friends think it’s cool – everybody else thinks it’s stupid.

Making A Bad Situation Worse

I can see the necessity of occasionally putting in some extra effort and burning the midnight oil at work for a day or two. But when "occasionally" turns to "often", when your boss stops thanking you profusely for your efforts and just treats it as norm, this is when we’re all in trouble. It sets bad expectations, not just for you, for the whole industry. Humans are like dogs, we’re eminently susceptible to positive and negative reinforcement. And this whole industry has been conditioned by years of death-marches to the point where it even rewards this behaviour. Every time we give-in to the long hours argument, we continue to negatively reinforce this trend.

It doesn’t help that we’re herd animals, you only need to get one person and everyone else wants to conform. Guilt comes into the equation – "we can’t leave our mates by themselves to do the hard-yards, we gotta help them". The more people conform, the more pressure on the rest of the herd to do so until the whole team is chugging coke and eating pizza at midnight. But how do they suck even one person in, where is that famed programmer independence. We’re happy to "stick it to the man" and do whatever we want in school, but as soon as we’re in the workforce all bets are off. It is puzzling.

Interestingly, sometimes these gargantuan efforts aren’t even tracked properly, as it would make the project look bad. So they "cook the books", as far as the client is concerned everybody is doing 40 hours a week (i.e. they get billed for 40 hours) and the project is coming in on schedule (maybe), never mind the other 40 hours that everyone on the team puts in. OK, maybe they’ll track the real effort in a "second set of books". Accountants go to jail for these kinds of shenanigans, but our industry expects it – nay almost demands it.

The Sustainable Pace Effort

Most Agile processes talk about sustainable development pace. But, I’ve seen even self-confessed agile teams knuckle under and put in the hours, you know, for the greater good and all. They were still agile though, and don’t you dare say otherwise.

When I think about this stuff I am always reminded of lawyers. You come in as a new lawyer and you put in massive amounts of effort and time, it is the accepted way to get ahead in that industry. No developer wants to be compared to lawyers, but often the situation is similar except you’re not going to get ahead by doing a lot of overtime as a developer (unless you’re working for a big 4 consulting company and then you might as well be a lawyer :)). So, lawyer vs programmer, which one is the chump?

Studies about productivity declines when working more than 40 hours a week surface with disturbing regularity. As a developer your creativity declines, you make more mistakes, you miss existing issue etc., to the point where you’re doing more harm than good. Should I even mention the health concerns when you spend that much time engaged in the same activity (they even had rules about spending too much time at work in the Soviet Union, and those guys were all about putting in the time for the good of the people). What about diet, you can only survive on coke for so long – poor John couldn’t even make it to 40.

Can you tell that I am against long hours and death marches yet :)? Maybe one of these days I’ll tell you how I got my wake-up call, it is an interesting story. Herding cats is easy compared to getting developers to make a concerted effort in the same direction, it is something I both love and hate about our people (programmers) :). But I do wish that once in a while all the smart developers just took a stand to eliminate at least one of the truly crappy and counter-productive trends in our industry. As far as I am concerned, smart programmers don’t like to work long hours and won’t be pressured into it – there is more to life.

Advertisements

At last!… >:)

Hey guys! These were truly some days when evil was hovering over me!… or at least, so I thought!

I’ve been trying since last Saturday to deploy a Win CE application I’ve written in C# for a new embedded board I’ve received from Microsoft as a prize, but I kept getting this annoying errors, like “the Compact version used is too old” or “the application is already running” (and of course, it wasn’t running) and so on.

Since I’ve got that first error, I believed that there was a problem with the WinCE image I’ve built. So, just to be 100% sure that the problem was the image (which I admit, didn’t find to have build-logic problems), I’ve tried deploying the image on another older board, with another image of course, also on an x86 architecture, an image that I knew was perfectly capable of running anything I had in mind.

But what about this? I plug in the cross-over cable, boot the image, run Coreman for deploying the app I’ve written and boom: “Deployment and/or registration failed with error: 0x8973190e. Error writing file ‘\Windows\System_SR_enu.cab’. Error 0x8007274c: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.” I try the Coreman again, run the debug again, and then I get “the process is already running bla-bla-bla”.

So what to do next? Well, first I was thinking of writing an e-mail for support, but whom to write to? And what to write? Writing an e-mail with “nothing works” as subject, and the posted error in the body of the e-mail would have probably been ignored… And then it came to me: what if the network board is broken (to be read as f***** up”)?

A short history lesson: back in December I had to buy a new network board for my laptop because I’ve accidentally broken a pin of the original on-board Intel gigabit network board. And since it isn’t possible to change the on-board network module, I’ve decided to buy the best PCMCIA network board on the market: the D-Link DGE-660TD Gigabit board.

Enough with the history lesson. So here I am, unplugging my D-Link network board. I dock the laptop in the Lenovo docking station, plug the cable into the docking station network socket, change the IP addresses as required, start Coreman all over again, press deploy and taa-daam: deploy succeeded.

Don’t know if any of you ever had similar problems, but googling any of the errors I’ve got during these last 4 days I had absolutely didn’t help at all, so…

Ok, that’s it for today!

Alex