Posted by timboudreau
on February 8, 2009 at 10:03 PM PST
I was chatting with a friend who works for RedHat today. You don't hear much about them anymore. Remember when RedHat Linux was all the rage? I've still got that self-congratulatory book with the guy with the silly red fedora hat on the cover. For that matter, remember when Sun was all the rage? Maybe this is the same story...
I was chatting with a friend who works for RedHat today. You don't hear much about them anymore. Remember when RedHat Linux was all the rage? I've still got that self-congratulatory book with the black-and-white guy with the silly red fedora hat photoshop'd onto his head on the cover. For that matter, remember when Sun was all the rage? Maybe this is the same story...
I remember trying out some video chat software from the NetBeans office in Prague in October, 1999. My "office" was a corner of a room, and it happened that there was some space in the corner, so all of these empty workstation boxes got piled there - we'd just been acquired by Sun, and had gotten a pile of computers for the trouble.
My friend was agog - "Is that really a giant pile of SUN WORKSTATIONS behind you?!?!" (I had to say, sorry, no, just cardboard boxes - but they're somewhere around here).
But here's the critical thing - nobody would say that today. Why is that?
So, enough biting the hand that feeds me - let's go back to my friend at RedHat. I had used Linux distros since the mid 90's ( Slackware still holds a special place in my heart for the sheer simplicity of its installer and how easy it was to end up with...something that didn't work - but Gentoo is the only easy way today to put together a similarly bare-bones (no gui) linux install). But at the end of the 90s, along came RedHat. And Netscape cum Mozilla. OMG! Companies are doing open source! And if you wanted a Free OS, but wanted it to work, you were using RedHat. Everybody was using RedHat Linux. It was the Linux.
Today, I have two machines with Linux installed. My coding machine runs Ubuntu, my server machine Gentoo. Anybody I know who knows what Linux is is running Ubuntu unless they're a sysadmin, in which case they're running Gentoo. What happened?
Well, at some point, I don't actually remember when, RedHat decided that this whole business of doing a consumer OS was costing a more than it returned. The money was in the enterprise. So eventually the whole consumer OS piece of it got sort of spun off into this Fedora-somethingorother thing (don't get me started about product naming), and progressively lost steam, but who cares - because they're at warp speed on the enterprise.
Back to Sun for a moment. In the dot-com heyday of the late 90's thru 2001, everybody bought Sun hardware. I remember one of the stupidest parties I've ever been to, in early 2000, at CeBit in Hannover, Germany (Hannover is nice, but avoid CeBit like the plague unless you like getting up at 5AM to spend 3 hours to drive 5Km to an exibition hall you won't get out of until 11 at night). You had to have a special pin being handed out at the show to get in. It was a Star Trek communicator pin with the logo "Dot Com The Enterprise." I think that was the moment I knew the bubble had to burst.
So what really happened to Sun? Well, sometime a few years before I arrived, Sun was making so much money on enterprise that, well, that whole workstation business didn't make the same kind of profits. Best to get out of it. I don't know if it was MBAs trained in the mantra of "focus on your core competency", malevolence, the fact that any large organization of human beings goes insane very quickly, or just pure stupidity (one day "focus on your core competence" will be studied curiously by anthropologists as a cult-like phenomenon - ask yourself this - do you have a core competency? Just one? You're only allowed one...) "Business" as it is studied today is a soft science like literary criticism with about as much credibility as Derrida deserved with deconstrution (i.e. none). It just scientififies itself by dressing unjustifiable assumptions up in pretty numbers.
There's just one problem with that whole "core competency" disease: Those unprofitable workstations were your marketing channel. They were your way to show the world that you do good engineering - to show that to the people who would be the decision makers at all those dot-coms to-be. To think the people who are deciding to buy your stuff today will be the buyers of tomorrow - that you've got a "permanent majority" isn't even the height of hubris, it's the height of stupidity. You can't even call it self-serving behavior because it's not even that! At best it's a malignant form of laziness.
For most any company in the enterprise market, what you put into *consumers'* hands is the determinant of your future success. If consumers - particularly college students - don't know who you are, the enterprise of the future likely won't either.
The big enterprise money is a distraction par excellance. Here I am watching two companies that both made their run to the top on consumers, and then forgot about them when the big bucks started coming in. Your loss-leaders that get young people using - and loving - your technology are the decision-makers of the dot.coms of the future - you live and die by those people's decisions.
I once sat in for a week at my friend's ISP - they'd fired their support company and had a week with no coverage before the new support contractor came in. She begged and pleaded and paid me to come down to connecticut, sleep on the couch and be the tech support department for that week. I asked for big bucks, but I did it (it was hell - never again). The CEO had this huge pile of resumes for the support job. I watched him rapidly go through them, and it was amazing how fast some went into the trash. I asked him, "How do you know so fast to throw away those resumes?" He replied, "Oh, those were MCSEs (Microsoft Certified Support Engineer). This did not surprise me - a year earlier I had dinner with a friend from high school that I had not seen in years. He'd just passed his MCSE test. Over dinner he asked me this question: "So, I understand what HTTP is. And I know what HTML is. But what's 'the browser'?" Which goes to show just what technical certifications of that sort are worth - they're only useful if you're interviewing for a company large and dumb enough that the interview process is checking off formal training and qualifications, not whether you actually know anything.
One of the best examples of how to do the opposite comes from when I was in college. I wanted to take electronic music - they had all these 1960's Moog synths I wanted to play with. I got really lucky, and Everett Hafner was the guy who decided if I got in or not. My music reading skills were and are almost non-existent. So he tested me on intervals - play them - could I tell a fourth from a fifth - he actually tested what I intuitively *knew*, not my knowledge of classical music theory. And that's the ideal job interview - you don't want to find out what someone can regurgitate by memory, you want to find out how they think on their feet. That was a revelation - that there are people in the world who will look at your skills not your paper qualifications. I'd suggest that any company focused on paper qualifications is not one you want to work for. If they can't invest the time to find out who you are, they never will, and that doesn't lead to happiness.
So, back to RedHat, or my perception of what's going on there: it's sad to see the same sad story repeated. The enterprise bandwagon is very seductive. It lures you to think you're so comfortable that you're comfortable permanently. And if you're not careful, that's exactly what you'll be.