Safe drive + Wikipedia on Safe - technology preview

Hi guys,

This is more a pool than anything else. I want to play with the network (I know it’s full now), but time is scarce so I would like to know if anyone is interested in using these kinds of apps (answer with +1 and the project number if so):

PROJECT 1 - safe-vfs: a virtual filesystem for Linux (“safe drive”)

  • Phase 1: Safe as a remote disk
    Linux only. It mounts 2 folders: “/private” and “/public” which can be accessed by any application. With this, you can use the OS file manager to manage your files. It can be useful in development to sync files with the network.

  • Phase 2: DNS management support
    Implements a “/dns” folder, containing the user domains. Domains are displayed as symbolic links pointing to the “/public” folder and their names will be in the format <service>.<long_name>. It will be possible to create new dns by creating new symbolic links in this folder. That means you can manage your domains and create new sites, all using “ln” and “mkdir” commands.

  • Phase 3: public sharing
    See others “/public” files. It implements a “/others” folder, containing symbolic links to publicly available dns. It mounts others /public subfolders in read-only mode.

  • Phase 4: private sharing
    Add members to share your private folders in two modes: read-only, read-write. This can be achieved by creating private/public keys and sharing it between members.

  • Phase 5 and beyond:
    Windows / Mac support; Web interface.

PROJECT 2 - Searchable Wikipedia:

  • Phase 1: Simple English, read-only mode
    The entire “Simple English” Wikipedia, in read-only mode, with no images. All 225k articles. A script can be set to sync the “outer world” Wikipedia with this version.
    In my testings, all those articles can be stored in only 178 immutable and 100 sd objects. My idea is to not use the filesystem submodule. No need to bloat the network to make this functional.

  • Phase 2 - Search mechanism:
    My mechanism requires one GET per keyword, so it’s ultra fast. If you are looking for articles containing 3 specific words, the app will perform 3 GETs to show you a list of articles.
    For Simple English it will be necessary between 2k to 3k immutables to store the whole search index.

  • Phase 3+:
    Write permissions can be added later, as well the ability to follow moderators.


PS: I know, MaidSafe will probably develop these things in the future. This is not intended to be an official version of these apps, just some kind of technology preview .

PS II: I do not guarantee that I’ll develop them, I can but only if is there a combination of real interest and time.

22 Likes

totally suing you for copyright

lol jk jk idc

5 Likes

This was one of the deliverables of the early project and as far as I know is still going to be done. A drive that one can mount in the OS file system which holds your public and private directories and thus usable by any program.

4 Likes

These are both awesome projects IMO - so nice to see them spec’d out and researched :slight_smile:.

I find it hard to choose but if I have to, I would say drive has more reach and early value to the project so higher priority, while demonstrating a searchable Wikipedia would both be useful and show how to build something people will be inspired by (including myself - scripting to copy existing websites is something I’ve pondered but not thought deeply about, and Wikipedia is an obvious choice for me too, though I wouldn’t do there :wink:).

@neo is right that MaidSafe had plans for drive, but I know they would rather someone else built it to free them up to do I other things (at least that was their position a while ago). Also, as he points out, @whiteoutmashups is doing something similar, but AFAIK not this - Will, correct me if I’m wrong, but you are planning a web app, rather than a mountable system drive? Both are useful, and so I don’t think either negates the value of the other. I want both!

So overall:

+1.1 SAFE drive
+1.0 SAFE Wikipedia

But if you built either it would be awesome! :slight_smile:

12 Likes

Great stuff @loureirorg

Both +1

Project 2, then Project 1, then Project 3 a Safenet virtualization

As Project 2 may draw more interest and benefit Project 1

Project 1 is definitely going to blow top off use so I wud push for financial support -

So you may or may not do it. Likes on this forum is’nt enuf. What is the motivator for you?

4 Likes

+1 for Project 1
more chars here

2 Likes

Project 1 +1
<padding>The Jocund Mavis and Valdo Pons - stars of stage screen and toilet walls </padding>

1 Like

I started working on something like project 1. It uses fuse to mount safe net volumes as virtual drives. Current version is read only and is bound to old REST API. I planned to pick it up again when new API was finalised.

Project components are here:

13 Likes

You got the point. Actually, I think we have a niche market here: the community is avid for the network, we can clearly see this in form of frustration of some. Maidsafe team is a perfectionist team, they are a talented but small team, things will not happen soon and there’s a lot of people willing to consume this product. Why not develop products to these people? It would be a win-win situation. Give the people what the people want.

Long, long time ago I remember that emule had a lot of modified/hacked versions with improvements (and lots of bugs) - they were branded as “emule plus” and things like that. They were independent, non-official versions. Some paid, some free. In the same sense, why not do something like that?

Would people pay to have a non-official Android SafeBrowser? Or to access a private test network? Or pay for extra PUTs?

Not intending to be a replacement for official development, just to offer products for an eager to consume market.

I’m derailing the topic, but this can be an interesting thing to think. Maybe we have a whole niche market here.

PS: back to the topic, my idea is to develop these things as a hobby, in my spare time.

7 Likes

So, we discovered your real identity. You’re a dog.

I did some FUSE virtual filesystems in the past too, so yeah, it will be a FUSE-based filesystem.

2 Likes

Oh no, the secret is out! :slight_smile:

2 Likes

@loureirorg is this still a possibility or was it one or the other only?

3 Likes

uuuuuh i missed the part with wikipedia here :smiley: xD @Savage your memory is superb =) aaaand i’m once again surprised how far other peoples thoughts already are on that topic :open_mouth:

ps: https://github.com/loureirorg/wikirofs
[he probably didn’t completely forget about wikipedia :smiley: just an app for now and not a website]

3 Likes

How feasible is it to build something like a working Wikipedia?

If it could be done, not just as a read only tech demo, it could be used to host other wikis, think decentralized Wikia.

Would it be possible to build a decentralized Wikia on SafeNet and have it work as well as Wikia does on the web today, with search and everything?

The reason Wikia exists today is to provide “free” hosting for wikis in return for filling them with ads. Is it actually feasible to build an decentralized competitor to Wikia on SafeNet?

1 Like

Top story on hacker news now is Uncensorable Wikipedia on IPFS.

https://news.ycombinator.com/item?id=14292396

Someone in the discussion on hacker news also mentioned a project called Lunyr, that seems to be about getting a Wiki up on Ethereum.

Are those two protocols uncensorable?

From the discussion it would seem like IPFS is more censorable than SAFEnet at least

https://news.ycombinator.com/item?id=14293144

I haven’t looked into any details about Lunyr, but they have a whitepaper on github.

2 Likes

I haven’t thought about it before, but apart from all the obvious reasons why SAFEnetwork will blow all these other projects out of the water if it hits its goals, the crucial factor for projects of this type will be the ability to survive in the face of governments that turn against them (which is increasingly looking like all governments as they become ever more authoritarian and oppressive).

I don’t know of any other project that takes this issue head on and might achieve it, although as I said, it isn’t something I’ve gone into in any depth.

3 Likes

The last time I looked at this, it looked relatively easy to build a Wikipedia mirror. You just needed plenty of bandwidth to upload it.

IIRC, you they essentially store the data in mysql and then generate static content for fast hosting. This is perfect for safe net, as we don’t have to worry about how to make it dynamic - it should just be static html.

2 Likes

What could go wrong?

3 Likes