ex tempore

The world, as we get used to know it, is changing fast, and we got ourselves into the middle of a crisis, a global one, that some dare to compare to the ‘great depression’. In short, hard times. Some things didn’t change, or didn’t changed (yet) enough, and one thing that didn’t change is the ’speed’ with which everybody ‘analysis’ everything. In this Internet age, people expect and demand answers and analysis in real time. And the answer the world got for the mess, in the high end financial world was that all could be summed in a bunch of buzzwords – ‘lack of regulation’, ‘executive pay’, ‘greed’. This kind of ‘reasoning’ may be good enough to please the masses, specially when baked together with heavy doses of demagogy, but long term, won’t do much to avoid new messes. This, because the real root of the problem lies somewhere else, and is much more general. Is that in a increasing complex world we tend to focus and specialize on the small things, while missing, ignoring, trusting (pick your choice) everything else that surrounds.

This is superbly pictured in the piece that Vanity Fair run about the failure of AIG and the systemic crash, on the whole financial system, that come after. People get so fascinated, so distracted, with their own ‘little’ things, that everything else stops mattering, other than following ‘trends’,’ blindly trust the ’system’. As a result the system, which, per definition, is the sum of all those ‘little’ things, starts malfunctioning because the checks and balances that should had naturally happened, didn’t, because all the different hands at play didn’t fell the urge to understand what was going on, and why, outside of their small ‘kingdom’.

It wasn’t a matter of just greed, it was worst – it was a ‘culture’, that goes way paste the financial system, that simply took people to just trust on whatever was going on. As the access to a plain calculator, shouldn’t be an excuse to anyone to not have minimal math knowledge, also the availability of powerful computers, powerful statistic and data analytics software, shouldn’t be an excuse for anyone to ignore or not understand the fundamental principles of what goes underneath around.

In a sense, this is all about laziness, about managing complexity and entropy levels, about not relying just in ‘trending’, about just going up when everybody goes up, and as now… just going way down when everybody else is going down. In a strange way this blind trust in the ’system’, is relieving, one hasn’t to doubt, one hasn’t to question, one hasn’t to criticize or run against the tide, one has just to be slightly better/faster than the competition, but in the exact same way as them. the ’system’ ends working as a kind of global ‘intelligence’, one that dumbs, and normalizes, everybody. In AIG, and elsewhere, things gone wrong because simply, and in plain words, no one had seen the urge to ‘think out of box’.

Now, lets turn to software. Software (stack wise) isn’t getting simpler, and yet it never stop amazing me about how much time, money and effort is spent in hiding and abstracting all those complexity, instead of getting things simple, clean, transparent, and … introspectable. Amazingly, some are preaching this trend as a business model (very much as the sub-prime was), as while all those abstraction wrappers somewhat appear to simplify life at top of the stack, it makes the core so insanely cryptic and complex, that it acts as ‘natural barrier’ to external change, and so preserving existing ‘grass fields’. That this happens in closed shops and ecosystems as Microsoft’s, is no wonder, that this happens in open-source stalwarts is worrying. It is worrying because it, in practice, replaces the need for real insight (about what goes on at application level, and the relations of that application with the stuff it depends on, and that depends on it) by a ‘brute force’ approach, since as one is doing what everybody else is, then it must be safe…

Yes, there will be a post-{rpm,deb} world… the real question is if that is going to happen because some had learned something with the woes elsewhere, in advance, or if we, Linux world, as an ecosystem, we’ll ‘need’ our own ‘Microsoft like’ woes, to move forward again… in the layer that will matter most. Assuming we want the age of monolithic stacks over, what will get deployed, to tons of places, is a pack of lots of small pieces, developed asynchronously, by totally unrelated teams. Managing the lifetime of those deployments, in a sane, scalable and simple way, installing, updating, erasing them, understanding their inter-relations, is what (should) matter most in the end. Sooner or later, the ‘operating system’ concept will evolve somewhat to turn to be the distribution layer, the one that manages and deploys, the lifetime of whatever runs, locally or in the cloud, as both sides will turn into an extension of the other. It’s just a matter of time, and it is fundamental to do it right, and that, is why i still bet in conary.