Thats veering a bit off topic here, but Java and Javascript are totally unrelated aside from the first part Rhyming…
Stallman doesn’t like JavaScript much because it is often used to download payloads that on non GNU without the user’s explicit consent – but there is no real problem with JavaScripts licensing itself as far as I have heard. – and most any in-browser supported language is going to have the same superpowers…
Lawsuit risks are inherent in nearly every piece of life… It everything will be safer in SAFE.
I believe Decorum and probably most of the SAFE apps will use JavaScript to one extent or another. It is really the easiest way that computing can be done in the browser, and really your only choices are doing the computing in the Browser and doing the computing on the Server – For the most part P2P eliminates the later…
IMHO IPFS VS Maidsafe is all about scalability & performance.
IPFS’s using existing software platform & protocol is not significant advantage over Maidsafe since both system does not require hardware update.
In terms of data transfer ipfs or any non encrypted transport will be slightly faster than an encrypted one (or signed). SAFE does a bunch of checks on authority as well. Ultimately we will need to get running and measure. It never bothers me to much with the increase in b/w cpu etc. This is why we took the plunge to go for a fully secured route, any slow down tht a few % is worth it.
Caching etc. will also make a difference mind you so we will need to measure and of course always improve as any project will, IPFS also will be doing the same. I suspect we will learn form each other thre.
In related section of Wikipedia page, maidsafe is listed but not linked. Is this something that can’t be done yet or just hasn’t been? @frabrunelle thoughts?
an alternative network stack for building secure, decentralized and privacy-preserving distributed applications. Our goal is to replace the old insecure Internet protocol stack. Starting from an application for secure publication of files, it has grown to include all kinds of basic protocol components and applications towards the creation of a GNU internet.
I am surprised of all the buzz around IPFS. I have tested it and it works great, but the command line environment is so much less friendly than Safe. IPFS really lacks a good Gui.
Another point is content gets deleted if it’s not pinned, which is a pain.
A great thing on the other end is that every file is given a hash which can be used to reference the file and avoid duplication.
Will Safe incorporate a way to produce a public unique identifier like IPFS hash, for a file stored as public ?
This exists at the API level, but we need it at the user interface level (i.e. like a safe:// URL). I wrote an RFC about this ages ago but have not heard of any plans from @maidsafe to support it (or not). If not, it can be done as some kind of add-on app/service, but it would be better IMO if there was support for it in the API because this would make it easier for apps to implement, cheaper for users (I think, in terms of PUT costs) and more secure (e.g. we may be able to make it harder to spoof links).
Duplication is handled inherently by the SAFE code. Two exactly the same files will only have one copy stored on the network and the datamap will be the same. Thus people can upload files without needing to worry about duplicates filling up the storage.
So the need for a “hash” to identify duplicates is not needed from the network side since it handles duplicates automatically
The way SAFE avoids duplicates is that the chunks have an address derived form the hash of the chunk and if a chunk already exists at that address then its a duplicate and the chunk is not stored again. And no refunds (confirmed) is given for storing duplicate files. At this level it does not matter if the files are private or public and each person storing the file is given a datamap.
The nice thing about the IPFS hash is that you can easily see if you already have the file before trying to retrieve it from the network. I think IPFS checks local storage first, then network computers (?) before retrieving from network.
So a Chunk/file:s hash can be stored in differents users datamaps. I guess it will not be possible to delete entries in the datamap as it would affect others users that had uploaded an identical file?
Yes correct. Of course no knows who else has the chunk address(es) in their datamaps.
Correct. And because a copy of the datamap can be shared amongst friends we have an additional complication since the system cannot even keep a count or anything like that.
I am certain we would. When this current work to finalise at2/routing/crdt types is done and we are in Fleming. Then we will be taking a look at other running networks to see what we can share.
