At which point will it be possible, if ever, to create apps that use the SAFE network for more than just anonymous, incentive driven, decentralized storage ?
The main tasks of social platform like Facebook, Reddit and Discourse require computations to provide each users with filtered and sorted view of the information available. (list of friend, most popular posts, ect) They also need to enforce access control dynamically, allowing and restricting access to many aspect of the application.
Basic search capabilities and conditional actions. Will SAFE ever do that and when ?
When I see all the craze about rewriting every existing app into itās SAFE counterpart I canāt but wonder what am I missing? Ethereum has been focusing on decentralized computation from the start with a team of 30+ and barely have anything working that equate what some of you guys expect to achieve.
Is a āsimpleā functional decentralized storage network where every userās bandwidth, storage and latency requirements is met with proportional incentives possible without decentralized computation and a complete rewrite of every existing webapp possible ?
The challenge is to be able to write front-end applications instead of having all of the storage being held and computation being done on a server.
Even when decentralized computation is implemented, the scripts will still have to be only on the client side. Decentralized computation is not performing back-end options, but rather performing the front-end computation on behalf of the user. So itās still just front-end applications, but a dev would then be able to count on having more computational power than the userās local machine would be able to provide.
This is a very new paradigm, and one that is to be encouraged. See this blog post for a approximation of what weāre talking about.
Basically, yes, it is completely possible with a little ingenuity and lots and lots of trial and error - just like any other new technology.
EDIT: I get it, itās hard to imagine the internet without big datacenters and black-box back-end code running the applications housed therein, but this platform provides the perfect opportunity to flesh the ideas of one such as that out. While the end-goal may be functionality similar to Reddit/discourse/facebook/etc, the approach will be completely different - both in the sense of decentralization and front-end driven applications.
Actually similar to bitcoin, The nodes donāt do any computation from the ledger, and can still move money around based on signature and verification. Thatās enough to get started with Safe Contracts
I think I see what youāre getting at, I guess I should read your whitepaper soon. Ethereumās approach has shaped my concept of smart contracts too much I think.
Thereās work to do, and we have to think differently, but we are actively working in these things in parallel with MaidSafe getting the network in place.
Itās hard, extremely exciting - because itās all so new - and a great opportunity.
For those who remember web 1.0 (static HTML - not even client side javascript), it is not hard to believe that this will happen, even though it is hard to imagine how!
I recall the MD of an innovative software company he started up with my then boss saying he didnāt think the web would be more than an online brochure.
This guy was not an idiot - he and his partner created a whole niche of visualisation and analysis software and sold it throughout the world, and wrote the first version themselves in a back bedroom. But he still couldnāt imagine what people might do with servers. We missed the first dot com boom partly because of that, but before I left there we were part of that revolution.
Itās our job to imagine and create the next SAFE internet, without servers!
@seneca a better reddit is one where users own the content and the dev team still has food to eat, and a comfy chair!! I think youāre onto it. maybeā¦
I think of ethereum as just a fast blockchain, so you can now make an app with a fast blockchain, and so when you store all the lines of code, itās a transaction and costs ether each step of the way.
Itās like imagine if you put each line of code onto the bitcoin blockchain in a series of transactions and made a client read those lines one by one compile the software execute and give you the result of its execution; THat is inefficient.
Unlike ethereum all the activities weāll make I mean any application can do the same without the huge expense; since we use idle resources. The cost is just at the storage or creation.
iām as techy as i wish i was, but i am hearing it to be like this:
SAFE Net will just hold all the data (IDās, chunks, etc)
and all these future super-clients around the world just call data from it, and then do everything on the local machine? (and in the future, can pull additional CPU etc from othersā local machines?)
So the network is really efficient, and only stores data, unlike todayās internet, which stores and mutates data with PHP etc
clients are gonna become super amazing in the future! maybe new client-side languages will start to be developed?
This is how I understand it as well. The big sticking point for a lot of people is the server (or lack of).
Reprogramming how we think about the interaction with data is a big thing to get your head around with the new decentralized storage.
I tend to think of SAFE as a giant network drive. Data will be stored but interacting with the data will not be done through the old way of client/server interaction (pearl, php, etc.)
The client side will have to pull all the weightā¦it will be interesting indeed to see how this is tackledā¦but owning your own data and not having it stuck in some centralized server for sale or hacker bait is well worth it!!!
Keep an eye on IPFS ā this is one of those problems that lies within the overlap of SAFE and IPFSā¦ Most of their data is immutable and there are no servers per seā¦ IPFS is significantly further along so it is a bit more tangible to play with.
They use Git Style Merkel DAG is their architecture, and it would be adequate for many apps. The problem is that the whole chunk of related data has an address that changes, so you have to add a mutable layer to keep track of what the current version of āeverythingā isā¦ (They use ipns) I suspect that MAIDsafeās SD will be able to provide that kind of āmutable list of immutablesā serviceā¦ You still have ālockā issues where two users are updating the same tree at the same time - and that would need to be reconciled.
