I like such approach.
It also worth to test overhead for small chunks.
Someone is playing with such test on comnet now.
I posted screenshot with disk usage in that topic.
6 Likes
Also though for us we have chunks from SE These are the immutable data and they are already compressed, client-side. These chunks will number mostly in the 1Mb size and not tiny files. With tiny files, they are part of a dir and the dir is sent via SE thereby compressing and encrypting them to.
3 Likes
Do not completely understand what are you saying.
If you upload directory with lots of small files via single files put command, they will be packed into small amount of chunks?
Then how it will be possible to retrieve single file from such directory? By downloading whole chunk? Or maybe all chunks in case of compression?
1 Like
They will be in a container and go through self encryption. You will get back a data_map for that data. If it’s private then you keep the data map encrypted somewhere (likely as part of your own private filesystem). If the data is public then the container data_map can be public. But all files will go through self encryption.
With testnets right now this is a short cut to just make all data maps public as though there are no containers and it’s not how the network will work in reality.
Imagine a file must live in a dir/container (as they must) then 10,000 files in 1 container will give us a single data map for that container. Read that, rebuild the container and you can access the data map for any file and read that file.
10 Likes
Guys, I first heard about SafeNetwork about 6 years ago. And those development news always looks like “wait a year and you will get something usable” while in reality there are multiple testnets, a lot of research, and nothing that developers can use in the real world.
Maybe you should eventually select a subset of features to make them good enough to let developers like me use them in production and get valuable real-world feedback? Even if it is something really small. My major fear about Safenetwork is that you may just lose the grasp of reality and get stuck in building “spherical horses” in a vacuum.
1 Like
its do it right and not do it fast.
the goal of the project is to have data freedom which is no small task nor does it have a dev timeline.
once the network is solid you will be thankfull these devs didnt cut corners!
5 Likes
It is already subset.
No payments for example in testnets.
This is called optimism.
While in reality there are lots of problems, which requires lots of rewrite.
I think SN code already was rewritten several times.
I see just frequent changes of direction.
Maybe they are justified, maybe not.
6 Likes
Thank you for the heavy work team MaidSafe! I add the translations in the first post 
5 Likes