Software Development and the Bullshit of Big Ideas

=> The Cult of Microsoft | Mindset

I worked in private industry for more than fifteen years, and one of the things I learned early was all the big bullshit ideas by all the big companies roll downhill. At my first place, things were typically passed around via email. The CDO would send an email to all the developers, pass along what he was reading. Sometimes it was a link to Joel on Software. Sometimes it was Myers-Briggs (aka astrology for executives). The content changed, but the idea was always the same: read this, do this. Here's a better way to do things.

(As a side note, I've always come up as INFP.)

Some of it was good, almost commonsense - eg, "you should be able to run your builds with a single command". Some of it was not. There were whispers that all the developers were ranked against each other. Microsoft did it, so we were going do it. And the rumours, I found out later, were very much true - there was no culling of the lowest x%, because let's be honest, I don't live in Silicon Valley, and at the time, the market here was starting to turn in favour of developers. But they were used for things like annual raises. At the time, I made well less than half what I do now. Granted, I've had a lot of experience in the years since. But the company paid low and raised low. This was years before things like levels.fyi, even Glassdoor.

When I quit, one of the managers, who used to be the technical lead on a bunch of big projects I worked on, let slip, "I don't know why they always ranked you so low. I've worked with you. Your projects always succeed."

(Me neither, and yeah, they did.)

Talk about comments that hurt to hear, and were also good to hear. I took a lot of things from that job, but among them, a real dislike of Silicon Valley fuckery, and a chip on my shoulder as to what people thought of me. I've never been a flashy developer, but I've got good instincts, write clean code, rarely get sick. The tortoise, not the hare. Years later, after that company had been bought, and sucked dry, and collapsed, after I took a new job and kept writing my own code in the evenings and on weekends, I released my big side project, which I still regularly update to this day. And after that, as former managers were finding funding and starting new companies, I had a number of them inquire as to whether I'd be interested in joining their companies. My answer, always, was no. I always think of that first company as the development experience I needed in my 20s - I learned to work hard, to write good code under pressure in an environment that today would probably be called "bro-ish" - and I promised myself (and my partner) that I would never, ever work for a company like that again.

As I've gotten older, I've gotten tired of the games and the meta-discussion, the sort of things people think will inevitably lead to better software and continued market dominance. Meanwhile, the developers want what they always want: a lucid set of requirements; a good development machine with the right tools; the autonomy to suggest and implement improvements; the time and space to get everything done.

Zitron writes about how Satya Nadella, the CEO of Microsoft, has remade that company using notions of having a growth (rather than fixed) mindset. Based on the idea that people are capable of many things when they believe they can learn and improve, the idea is, apparently, central to Microsoft now. Whereas in the (bad) old days, you had Steve Ballmer sweating on stage, clapping his hands and yelling, "Developers! Developers! Developers!", now Microsoft is gripped by the happy notion of yes-we-can, which is, under the surface, a hardened insistence, something like yes,-you-must.

=> The Man Who Killed Google Search

Zitron goes into it in more detail, maybe too much - I like reading his work, but I'm not convinced that this topic needs anywhere near the amount of digital ink as his other work, particularly the excellent "The Man Who Killed Google Search". But I get it. There are word counts to hit, details to include in future podcasts, that sort of thing. I'm just not interested in the specifics of this particular pseudo-psychology. Rather, I'm interested in this as seen from above: as another point in a long line of big ideas. I've seen this before, and I'll see it again, not this-specifically, but some other thing, some great idea. Silicon Valley's accepted wisdom trickes down to those of us in the technological hinterlands months or years later. We ape it because if Google or Microsoft does it, we should, too.

If I seem jaded, maybe I am. I remember Joel Spolsky's blogs (the Joel Test, anyone?), Agile development ("hello and welcome to the web team's 74th biweekly sprint", words I've actually said). Stack-ranking employees. Test-Driven Development. MVPs and SLCs -

I remember when I was younger - mid-twenties, fresh out of dropping out of an advanced degree, starting my first job in software development - thinking, the first or second time: okay, this is great; surely this will fix us. And yet the weeks pass and the code ships and maybe it's a little better, or a little worse, hard to tell, but in either case, there's nothing revolutionary, just more process trowelled onto the hardened mortar of the old.

Satya Nadella won't be at Microsoft forever. Very few are. He'll be forced out, or, maybe more likely, announce that he's stepping aside to spend more time with his family. Someone else will come in, to great fanfare, and the trend du jour will change. But it'll become our problem, again. We'll see articles in the tech press, then by young enthusiasts and true believers in Hacker News, talking about how this, this will change things for the better. And life will go on. We'll get emails from CDOs, or directed to posts on the private intranet. The idea will be incorporated into quarterly checkins, annual reviews. And meanwhile, stickhandling their way around it, the lowly developers will get on with the real business, the one they're actually paid for, which is of course writing and shipping code.

=> gemlog