Feb 242010
No Gravatar

MacGyver in the 80′s and Gregory House today are in fact, two characters built on the same basic premise. Of a maverick genius who thinks outside the box, solves the puzzle and saves lives. That is dramatic enough for television but in fact maverick geniuses in the real world are more likely to come up with ideas like google’s revolutionary “let’s be the site where people spend the least possible amount of time – but the one they visit a thousand times a day”.
Not exactly television dramatics, but in fact the core point is the same. IT can consistently trace it’s greatest successes to somebody with an idea that broke every expected “rule” about the industry using the available resources in novel ways for a new and unexpected outcome- which changes the game. When you change the game, your company makes money because you immediately one-up every competitor.
I’ll bet a lot of CEO’s out there are thinking “ooh, yeah, that’s me- the game-changing MacGyver of the industry”… and with a few exceptions – sorry, you’re all wrong. The exceptions are those engineers who made a gamechanging creation in a startup and became a CEO without ever having studied an MBA. Again Google springs to mind.

Where both shows get it wrong is that they portray these maverick geniuses as rare gems of humanity that hardly exist. In fact… I’m not at all sure that’s true- the evidence suggest otherwise. Consider that in a fairly random sampling of the population (like, say, your office) 10% of the people ought to be in the top-10% IQ range in the world… that’s a lot of very smart people and a few truly rare geniuses too. But the news gets better, most people of “average” and below intelligence don’t work for software companies, they work for MacDonalds. So that means you staff is really only selected from the top 50% – and that means 20% of them are in the top-10% and a full 4% should be in the top 2%.
This number varies – the better your company is at attracting the best of the best (a matter of payscales, benefits and culture) the higher those values will get as you push the selection from which your sample comes to an ever higher subsection of the total.
So it seems… the genius part isn’t what’s rare… it must be the maverick part. The great out-of-the-box idea that no manager would ever have approved. Gregory House and Angus MacGyver both rely on bosses who value their results over methodology (and occasionally to enforce just a bit of sanity on them by stopping them before they get too risky… we rarely hear Pete Thornton say “thats’ an order MacGyver” – but when he does, it’s because MacGyver was about to risk his life a bit too much). They get given the freedom to break the rules, because the results are worth it.

That’s where google and opensource (I don’t say free software here because I’m talking about the development methodology, not the ethical movement) mostly gets it right. They go out of their way to reduce the redtape and checks and balances for their engineers. You can test and QA later, in beta. First- get the ideas, ideas make money. Google’s greatest invention wasn’t map-reduce, it wasn’t pagerank, it wasn’t contextual advertising… it was 20%-time. Give each employee a day a week to work on any project they want, with or without help – to try ideas without risk of rebuke if it fails – but with great rewards (massive bonuses) if it works.
It was 20%-time that gave google products like gmail and google-talk and the chrome-browser. That was google’s biggest game-changer. The best thing is that this idea not only got them great products, it made them the most sought-after employer in the world. Every software engineer is a little jealous of the people at google… where staff get gaming systems for stress relief and company sponsored massages for an hour a week… but above all – 20%-time… time to be creative and innovative without restriction.
This popularity has allowed google to headhunt the best of the best engineers in the world – a cycle that then feeds back into itself as these top engineers are attracted to the freedom they get to create, and what they create made google rich.
Open-source gets gamechanging creativity by making it possible to start a software project with very little expense – without lawyers or managers approving it – and engineers do it because it gives us a chance to really be creative and innovative.
But most software companies have no gamechangers, just the same products and their easy to predict follow-ups. Because in a system of rules, regulations, controls, checks and balances – nobody innovates, you just keep your head down and hope to not get blamed. When the system is like that, you can’t dare take risks – without risk there can be no creation. This is why most software projects are late. This is why most of them suck. This is why a good software company hardly ever becomes a great one.
A great software company must open the door to risk if they are to open the door to reward.

So throw away the rulebook, get rid of the dresscode (seriously guys, what does Larry Ellingson, Steve Jobs, Sergei Brinn and Sandy Lehrman have in common ? None of them have ever been seen in a suit !), give your superstars the freedom to shine… and give everybody on your payroll the chance to be a superstar. Most of them – are not going to let you down.

Dec 092009
No Gravatar

I won’t get into the concerns about whether google chrome is proper free software right now, mostly because I’ve started a discussion on it with the gnu/linux-libre group, which is a coalition of free distro developers where we collaborate and discuss these things together – and I don’t want to push anything until that conversation is done. Instead, here’s my review of the google-chrome browser’s official GNU/Linux beta release as I found it in my testing – with a mostly technical focus.

This also means there won’t be a kongoni port just yet – whether there will be one depends on the outcome of the aforementioned discussions.

I received a mail from google last night (which I’d requested) to inform me that the Linux beta for google-chrome is now officially available, followed the link and got greeted by a nice XKCD-esque comic about it. Followed another link and got the download page with the ugly EULA. Oh well. Packages were available for a few major distros – four in all, 32/64 bit RPM’s and 32/64 bit debs.

Kongoni can convert either to a usable format, and I’d previously done some chromium testing using the debs, but I opted for the RPM’s here. Doubt this makes much difference but just for interest. I grabbed the 64-bit one, ran rpm2tgz on it, and installed it. It created a /etc/cron.daily script which is meant to install the regular updates, currently of course, this won’t work, but if I end up supporting it, the kongoni version can easily enough replace it with it’s own. The RPM version’s script seems to use yum, presumably the deb version will use apt.

That out of the way, the next clencher was that it missed some libraries from mozilla-nss, which was odd since I have it installed. Double-check, the library names weren’t the same – close but it wanted additonal 1d and 0d extensions, a couple of symlinks sorted that out.

It came up, seemed to work – but wouldn’t render anything, checked the console output – lots of shm messages. Okay, I know that one from earlier experiments, set /dev/shm to world-writeable and retried. It imported my firefox settings, including saved passwords and bookmarks wonderfully – and suddenly, it works sweetly.

Okay, start playing… it’s fast, very fast. Faster than I remember from testing on windows… much faster. It’s slick, easy to use and just flows around the net. Played a bit with extensions – installing one for facebook and one for twitter – both worked instantly, without requiring a restart, and ran very nicely (though the buttons-next-to-the-address-bar choice may not be their best decision, that could get very cluttered fast for people with lots of extensions). Some googling around failed to find an addblocking extension just yet or anything for laconi.ca but I may just have not looked hard enough.

Still, I rather like it, it’s bleeding fast, beautifully rendered.. just about perfect in fact. The slight difficulties installing is probably because I didn’t built a proper kongoni port, and thus had to do manual effort to sort things out when converting a package built for another distro. All in all, I think google and their volunteer developers on chromium did and awesome job with the port. Well done, when my only gripe is a minor one of aesthetics – that says something.

Jul 082009
No Gravatar

I wasn’t planning to blog about this, I tend to skip whatever is trending in the blogosphere and besides – the views of one distro developer on another distro are likely be seen as somewhat skewed, but I was discussing the announcement of google’s chromeOS with a friend and explained how I see the news – and how I see it impacting on kongoni. The same could be said for almost any distribution really, but I spoke about kongoni because that’s the one I’m intimately familiar with.

She asked me to please blog it so she can share my view more widely, and I hereby oblige her. The below is just a cut-and-paste from the message I sent her so it’s not fully edited or perfected, but it does I think convey some of my feelings on this.

Now without further introductory ado:

As for chromeOS – been reading up … well it’s just another linux distro – the one difference is the google name – and the approach of basically building it to run nothing but a super-fast browser.

Well… the original netbook guys were on almost exactly the same page (remember the first EEEpc ? ) – it didn’t work out in the market though. People wanted netbooks to be cheap notebooks and expected features that those simplified interfaces didn’t have.

This led to linux taking a knock in the netbook market, despite having been first in. For what netbooks were, it was perfect, and XP sucks.
In fact, this effect ruined netbooks – what we now call netbooks really ARE just cheaper notebooks, and getting less cheap all the time because they have to keep making them more powerful.

The tiny device meant to be a mobile augmentation to a full computer didn’t take off as planned – bad marketing I think, ASUS didn’t have the capital left to market it the right way.

Google however, are masters at left-field marketing, masters at getting people to think in new ways and really good at stealing microsofts customers… this could actually be the linux that makes it big. I do hope so.

Kongoni would never have been that gnu/linux. Kongoni is a power-user’s system and in fact is probably more appealing to mac power users than windows power users (mac is built on a BSD base while Kongoni is heavily inspired by BSD – both thus giving you a BSD like OS underneath a pretty desktop – though the desktops are not very similar techs [but I think Kongoni's desktop is seriously sexy, have a look at the screenshots])

Either way – I hope chromeOS does take off – anything to reduce microsofts monopoly – and this isn’t trading one monopolist for another. ChromeOS is free software, anybody can improve it, rebuild it, modify it, create a derivative… slavery ended.

Of course, once people switch to a gnu/linux – ANY gnu/linux they do get better, and better, later they do want to try others out – want more power and less handholding… that’s when kongoni will be waiting ;)