I just had to watch a security training video at work. The work MaidSafe is doing may be the most important work of our time. SAFE Net solves nearly every (if not every) problem shown in that video. I am anxiously awaiting the day when one of these dev updates announces the release of SAFE Net. Meanwhile back to work on my own SAFE App.
=D did i mention that a simple messenger as signal-replacement for ‘difficult countries’ would be super-awesome? … since the telegram android client is very good-looking and a nice to use app … maybe some java-developer feels motivated to adapt the app to use SAFE as backbone …? =D
You mean forking this code and switching out the connection layer with SAFENetwork?
(probably this adventure would end up being more work than hoped … as it always does … but would be an awesome project i would think)
Hey @dirvine I’m doing a bit of homework on things mentioned in the update and a lot to digest but one thing that I stumbled across was strong eventual consistency vs weak eventual consistency. The latter not having a safety guarantee that any two nodes that have received the same (unordered) set of updates will be in the same state. So main question, is this implementation weak or strong?The state of nodes regards section splits and merged in this case correct? Not to infer that the weak implementation is a lesser one, either. A bit out of my depth here but hungry to know more.
Reading up on liveness is an interesting topic to dig into.
There is an astonishing amount of innovative work being done. I hope the academic work finds an audience.
For some reason this reminded me of the video in this post about the gossip protocol which has a great quote:
[54:37] I actually don’t know of many many cases where gossip in practice gives you a very very good solution
I think people will be impressed by the work maidsafe has done and the functioning implementation of these algorithms in the safe network.
It is subtle, but we are and have been considering both for different areas. Group consensus can be weak for instance, but for splits etc. it helps a lot to have strong consistency, however, the holy grail is async BFT consensus which has a ton of papers proving it cannot be done, but now has a few algorithms that provably do it So in the discussions and research we have considered both types, but have not settled concretely on any impl just yet. It is not stopping integration though as we hope that the API we use now for a working algorithm can be matched by any alternatives being considered.
This sounds as if those earlier proofs were worthless, so it’s probably a good idea to add that actually the opposite is true - those impossibility proofs showed which paths of research would be fruitless and made people direct their efforts towards ideas that cleverly work around known limitations. Those ideas eventually gave rise to working ABFT algorithms that exist now
Just to check… C# bindings… means I’ll be able to use .Net and unity to make apps powered by the SAFENet on the back end?
Short answer, yes! I am not a .net expert, but that is essentially what it means for java (native android apps, spring boot apps, etc)
Spectacular endurance on an amazingly complex project!
Via Con Dios
This keeps popping in to my mind. It’s a massive accomplishment.
Adoption would be unstoppable.
First dev use would trickle down, with all the benefits of no downtime, not threat of your server being closed, no repeating payments… safe assets… for some initial basic applications not requiring users to log in… just create one account and store all your assets there.
Perhaps information repositories for oppressed people across the world, that can not be taken down by bullying server and service providers…
Over time general users, will start to register with SAFENetwork to access more advanced app features…
End users wont even know they are using the SAFENetwork on simple apps.
Then the next step a little more interaction, they will get a whole load of “SAFENetwork Account Login Compatible” applications and wonder how their personal data will be available and secure and visible only to them across.
Eventually it will click, that all of these apps are linked to something called “The SAFENetwork”… and for those that have not already figured it out already, they will discover their user accounts gives them access to Farming… real Farming, not that fake Farmvil farming…
Ah, the future is great.
@Zoki I really like your post, those are exactly all the reasons that got me started in learning how to create an app on this platform, I think the benefits are inmense compared to what we are used to, and I think those are just to begin with, as we move on we will be discovering many more for sure
Are the Milestones in jira still valid for alpha 3 or not any more?
Just hearing of something called fast and furious Byzantine fault tolerance being applied in a project called Algorand. Have you heard of such @dirvine? here’s a vid that is supposed to talk on it, going to watch right now. https://youtu.be/mqqEjzmQva8
Yes in fact, the coin flip we liked that @Antifragile linked was by one of that team, it was a pretty neat algorithm we could adopt for async purposes. I will catch the video later on though. Cheers for the link. In fact the presenter was the author of the TRIVIAL coin flip.
How does it work in terms of patent in such a case ?
I’m looking very forward to this. Can it be clarified that the GPL-v3 being used has a linking exception? I can’t imagine you’ve be able to have the safe_app rust bindings without it due to standard GPLv3 requiring every link to be under the same license.
Yes it has and for the purpose you mention.
Brilliant! Thank you for the straight answer.