Those interested in computer security will no doubt have heard by now that “security is a process”. The idea is that security is not a thing you can buy and have sitting on you shelf forever. Rather it is a living thing that requires a constant stream of work and resources to exist. The second you stop feeding it, it starts dying.
My idea for starting this topic is that sanity, too, is a process. Left to their own devices, even communities of level-headed and rational individuals can devolve into cultish echo chambers. They fall prey to affective death spirals, stuck in a positive feedback loop of enthusiasm about the Best Thing Ever.
I’m afraid MaidSafe has great potential for being the Best Thing Ever. It’s ambitious, it’s disruptive, and it’s got a great mission, and if you’re reading this, you most likely want to see it succeed - perhaps on an emotional, as well as intellectual level. So what? Nothing wrong with that. Except that MaidSafe is not the Best Thing Ever.
This is not an attack against MaidSafe, mind you. MaidSafe is not the Best Thing Ever, because there is no such thing. No, MaidSafe is not without flaws, nor will it ever be. No, it is not without downsides. No, it won’t solve all your problems, or even all your Internet-related problems. It’ll most likely create some of its own.
But what’s with the negativity? “What are you doing here if you don’t like this project?”
Oh, but I do! I’ve put this topic in the Meta category, because it’s not primarily about MaidSafe. It’s about the community, about keeping a cool head, a putting in the effort necessary for sanity maintenance.
Criticism is not an attack to be defended against. Doubts do not require systematic elimination. The instant this is not true anymore in a community, you know its has succumbed to the affective death spiral. A healthy community welcomes criticism, and does not see its originators as traitors or adversaries.
Of course, I’m not saying that relevant refutations are not welcome here, quite the opposite. But I’m saying that we shouldn’t necessary look for one if none springs to mind. We’re not defending MaidSafe here, we’re defending the truth, and if the truth is that some doubts about MaidSafe are valid, then so be it. If you start looking for a refutation, then you’ve decided what your answer was going to be before you even started considering the question.
So here’s a topic dedicated to the flip side of the coin. Take all those nagging doubts of yours, and post them here, not necessarily in an attempt to see them refuted, but in the hope that unless there is a clear reason not to share them, people will end up sharing them as well and having a more accurate view of the project.
If these doubts spark discussion, then great! Let’s discuss the problem, look at its implications, the things it relates to, its consequences, and its possible solutions. We can fork specific issues into their own topics. Let’s keep our sanity together!
So I’ll start:
I’m really worried about network latency. RUDP packets will typically hop across several (say, 3-10?) nodes between their source and destination. Nodes are randomly distributed across the address space, which means that, assuming world-wide adoption, neighboring nodes could be on opposite sides of the globe. I’ve seen @dirvine mention many times that it made sense betting on the network over local HDDs if you considered that network bandwidth was increasing faster that HDD speed, but network latency will always be much higher than HDD latency, because both are bounded by the speed of light. You could pull magical super-resistant optic fiber cables through the earth to link nodes sitting across the globe, and still you’d be looking at >80ms ping between those nodes. Multiply this by, say, 5 hops on average, and you’re looking at ~400ms RUDP pings.
This makes MaidSafe unsuitable for a whole host of applications (real-time video games come to mind) for which data cannot be cached. VOIP is less sensitive to latency, but depending on real-world figures, it might be out as well. I’ve had a recent disastrous experience with a VOIP call between France and Australia, with latency as high as five seconds. Now imagine that even Troon - Troon calls could get even worse worst-case performance sometimes because messages could pass through Australia, Antartica and the ISS before coming back home.
Latency matters. It’s one of the biggest things one needs to get right in order to build a great user experience - hence the recent push for async everywhere. I’m just not sure that it can be hidden well enough in all cases not to hurt MaidSafe’s adoption despite all its awesome advantages.
So that’s it, that’s one doubt I have. Discuss, or post your own.