Wildland – open data management protocol | Hacker News Reader

Wildland – open data management protocol

wildland.io

128 pointsomnibrain47 comments

Apart from just "blockchhain", does anyone have context on what this means and/or what problem it's solving?

> Part of each payment will be converted into GLM to generate a Proof-of-Usage token that will give you certain decision rights within Wildland’s unique governance system.

I found this while researching if GLM was their own token, and it actually appears that it isn't their own token, but it's possible Wildland was created to show off GLM: https://old.reddit.com/r/GolemProject/comments/i4qwng/wildla...

Related to that, this is a lot less marketing-speak than their website: https://old.reddit.com/r/GolemProject/comments/nxfufh/golem_... but I still don't follow the relationship of "payment" to "decision rights" or "governance"

From the practical introduction, it says, "Wildland is a collection of protocols, conventions, software, and (soon) a marketplace for leasing storage and in the future compute infrastructure. All these pieces work together with one goal in mind: to decouple the user's data from the underlying infrastructure."

This would seem to make a user not dependent on any one platform to hold their data. So it improves data availability, platform lock-in risk and some censorship-resistance.

Thank you for your reply

So, I pay Wildland USD$100, they spend USD$80 on Google Cloud to provide some "s3" storage, possibly instances or cloud function execution or whatever, and then they give me back GLM$20(??) because ... they want GLM to flourish?

If "a user [were] not dependent upon any one platform," then why would anyone pay Wildland to be in their alleged marketplace? Why wouldn't I buy GLM on coinbase (or wherever folks do that kind of thing, since AIUI PoU means I can't mine them) and then save myself the hassle of paying Wildland for only some percentage of GLM refunded to me?

And all of this talk about our brave new semantic d-web 3.0 future, but until Google Photos or my insurance company lets me "grant" them access to my photos or medical documents or whatever, it's just college dorm talk, IMHO -- who is the target audience for this webpage?

Have a look at https://wildland.io/2021/06/11/introducing-client-v0.1.html and examples which talk a bit more about the client usage. GLM seems to be a completely optional add-on feature, not something wildland is based on.

The client they released yesterday has no built in payment mechanism, and the site says explicitly that "If you have access to a suitable infrastructure you can use [Wildland] for free".

The relationship is explained in detail in the Wildland paper -https://golem.foundation/resources/documents/wildland-w2h.pd...

There's also a talk on this very subject available at https://vimeo.com/473452625

Great, thank you for those resources

For followers on, the answer is in section 3.3.2 of the whitepaper (currently page 31) and boils down to:

> A very important feature of the PoU token is that it is designed to not become a speculative crypto asset. As it is permanently locked with the user’s public key, it cannot be transferred and hence traded. While we envision delegation of voting power along the lines of liquid democracy principles, this delegation can always be revoked by the user

so the PoU tokens granted back to the user is actually not related to the monetary relationship between the user and Wildland, but is a side-effect of merely participating in the experiment. Those tokens appear to grant the user voting rights on the product roadmap of Wildland, although it's not obvious how one would "spend" those PoU tokens given that they can't be transferred back to Wildland in order to buy a feature

They talk about it in terms of voting based on the accumulation of PoU, which I think understand, but if Jeff Bezos got to vote 54B times for President and did not have to _spend_ any one of those 54B assets to do so, the country will perpetually do whatever Jeff wants without any real risk of overtaking that king of the hill status

This project would benefit from someone with some marketing skills being clearer about not what it does, but who would use it and why.

I couldn't understand what it does and I'm working on something not too dissimilar (I think)

I don't even get enough grasp of what it does to decide whether I want to spend more time to figure out if it's useful to me. It can't be that hard to explain the thing your project solves/makes possible for me in 5 to 10 lines right?

Ditto. Nor do I know how it addresses any of the top-5 issues when it comes to data (consistency, speed of access, scaling, etc)

I don't believe it addresses many of those things, but it might address scaling given that its very concept is federation

But, AIUI, *the problem* it is solving is that [conceptually] instead of the user leasing S3 space from AWS, and putting their photos on Someone Else's Computer, but then being wholly responsible for the ACLs required to grant or revoke access to said photos, in this model the user can instead give out a lease of access to the photos to dedicated groups, and the storage comes from The Whole Internet (ala Filecoin)

Again, AIUI, this is much closer to a virtual-host model: I can create joshspankit.photos.example.com/timeline/2022/02/30/Trip to Antarctica/ which only you can access, and only the photos I put in that container, and I can revoke that at will by deleting joshspankit.photos.example.com at which point your lease evaporates

Plus, blockchain something something

Exactly. Tell me what problems it solves, not what features it has.

This is the problem with nearly every crypto project. Extreme focus on how it works and the tech behind it rather than use cases and who they benefit

This actually reminds me of an idea I once had. Rather than service providers holding onto your data, there could be something equivalent to custodians but instead of having lockboxes for your valuables, it'd be for your data. Service providers would then "borrow" your data from these custodians to provide their services.

This looks sort of like that but marketed from the tech angle using docker as an analogy rather than a custodian. I'll have to take a deeper dive into this!

Very similar to Solid! Worth checking out. It’s a Tim Berners-Lee project

Or Hyper Hyper Space:

https://www.hyperhyperspace.org

Or the now defunct tent protocol https://tent.io https://en.m.wikipedia.org/wiki/Tent_(protocol) from the team that made the also now defunct Flynn https://Flynn.io

I've had some thoughts about it too. In addition, I think, entities which want a piece of your data, should be able to somehow define "queries" and the "custodian" (in your terms) should be able either to tell yes or no (if you accept to answer to the query).

I imagine it this way, you have some kind of device, wanna drink in a bar? The bartender can "query" whether you are over 21 or not, you see the query, authenticate (with bio and pin/password) and your device confirms that you are over 21.

Wanna take a credit, the entity queries to check if your credit score is over 9000, and if you agree to let them know then your device just answers yes/no.

But there must be ways to protect the data, ideally the data should be encrypted at rest (wherever) and decrypted only in your device in some kind of secure enclave, so that your custodian can sign the result of the query being sure that it is not altered.

this is just a convoluted way to access information.

Imagine the use case you want to apply a filter to all your photos.

Either you give service provider access to all your photos, or they give you access to run the code out of their control.

There's no way to compromise this.

For this reason SOLID is pointless too. It is just a fancy way to saying "everyone will pay for their own S3 bucket and give access to random directories to random companies, oh and pay for the entire bandwidth"

There's no currently practical way to compromise. On theory we have homomorphic encryption which would be able to apply filters over opaque data. But it's not fast enough yet.

Sept 2020 seminar on Wildland, https://golem.foundation/2020/10/10/WildCon1.html

> Wildland is a highly ambitious project to decentralize the web, help individuals to take back control of their data, and free users from dependency on on-line service providers ... To achieve this goal Wildland completely redesigns data management, by switching from a service-oriented to the data-oriented paradigm, and giving individuals explicit control over the choice of data storage infrastructure ... Wildland architecture also allows for the creation of decentralized and user-controlled social networking apps.

Videos: https://vimeo.com/user124198022

First thing I looked for was whether there's content encryption, but couldn't find any mention on the landing page.

A deeper dig found it in the docs[0] but you may want to mention that up top.

[0] https://docs.wildland.io/user-guide/encryption-backend.html#...

How is discovery implemented? Is there a DHT or some gossipping?

I found only this:

> However, we believe it should be possible to bootstrap Wildland without any Wildland-specific infrastructure, such as supporting some custom network protocol

Does this mean that RnD haven't yet touched networking?

The way I understand it, it means that you can run a client on your manually defined data containers and not care about the networking aspect of wildland-specific environment.

https://docs.wildland.io/#multi-categorization is something I've always wanted, ever since I first started using computers. It seemed so strange to me that a file can only be in one directory structure. (Yes, links, I know, but that's a workaround IMO)

Many (most?) filesystems support hard links.

> Multiple hard links – that is, multiple directory entries to the same file – are supported by POSIX-compliant and partially POSIX-compliant operating systems, such as Linux, Android, macOS, and also Windows NT4[2] and later Windows NT operating systems.

https://en.wikipedia.org/wiki/Hard_link

Perhaps what I'm really missing is the UI support for these features. Also ability to hard-link directories. Remains to be seen if wildland will satisfy.

Joanna Rutkowska (QubesOS) is part of the team for this

Seems really cool, and I can tell that it probably is something I want to use... even locally, but the documentation assumes people know way too much already and only gives a couple strong examples. It appears to be an overlay FS for organizing and sharing data using possible blockchain solutions to then make that data accessible to other people, and vice versa. Am I getting that right?

Usage examples up front would be great for devs

this seems like a good starting point: https://gitlab.com/wildland/wildland-tutorials/-/tree/master...

Docs are not far away. Here's a practical introduction. https://docs.wildland.io/#a-practical-introduction-to-wildla...

seems to have some overlap (at least conceptually) with https://remotestorage.io/

I have no clue what are they doing. My "data", your "data"...it's not here, nor there, it's everywhere.

Kinda like a decentralized Dropbox except the data is actually stored in a virtual container on your machine. That makes it censorship-resistant and cuts away the single point of failure, like ones where outages of Google affect a significant part of the global infrastructure. These containers are encrypted by default and access to them can be shared (key exchange) or be made public. They want to decouple user data from the operating system it's stored on. The goal is making it possible to store different data in different containers (think of them as folders) and being able to clearly define the access to each container. It solves a lot of problems with IT security, since it allows a framework (and in the future a user-friendly interface) for everybody to store their valuable data in very secure containers only they have access to (operating system, secret documents, keys) while running pretty much the rest of their system in isolated containers. This way even if you were to download some malware only the container it's running on would be contaminated, unlike nowadays where the hacker gains access to pretty much your whole storage space. They also define another kind of containers called 'compute manifests' where instances with access can perform computations; as far as I know legacy cloud storages like Google Drive don't allow the user to run executables in them, so that's quite revolutionary.

What problem it is solving? I have gone thru the docs, but couldn't figure it out the purpose :(

As I commented above, from the practical introduction, it says, "Wildland is a collection of protocols, conventions, software, and (soon) a marketplace for leasing storage and in the future compute infrastructure. All these pieces work together with one goal in mind: to decouple the user's data from the underlying infrastructure."

This would seem to make a user not dependent on any one platform to hold their data. So it improves data availability, platform lock-in risk and some censorship-resistance.

When I see that description you quoted, it just reads like a giant string of jargon.

This probably dates me, but back in the day (1990ish), there used to be an application on Mac (System 7 or 8 I think) called the MBA Phrase Generator that would generate sentences like that.

From the website:

"We believe in the universal rights of data accessibility and data portability. We believe that users should be allowed to do what they want to with their data, to decide with whom they want to share it, and who can process it. They also should not be restricted in their ability to access their data, move it between different storage options, and interact with it with the tools of their choice. Vendor lock-ins, outside content moderation, and denial of service do not fit in with our view of an open, decentralized, user-controlled Internet.

This is why we are developing Wildland as a backend-agnostic, censorship-resistant, and open source protocol. Self-defined data containers that form the building blocks of Wildland can be stored anywhere you want. You can keep them on your hard drive, on your home NAS, or on one of the storage backends available through the Wildland marketplace. You can easily replicate your containers on several backends, each with distinct access policies defined, and move your data between different storage options without having to update their location in the file system. And with backend stacking employed you will be able to easily encrypt your data on the fly and turn any backend into end-to-end encrypted storage. In short, with Wildland, we are putting you in charge of your data."

https://wildland.io/2021/05/13/access-denied.html

There's also a long paper describing the rationale behind the project https://golem.foundation/resources/documents/wildland-w2h.pd...

Does anyone know of any other standards or libraries for self-describing directories (eg manifest files)?

I have a need for this in my own data archiving and it would be nice to not reinvent the wheel.

Hmm quite interesting. Might give it a spin.

What is the license?

It's open source.

Sure, but the license, which open source license?

https://gitlab.com/wildland/wildland-client is clearly marked with GNU GPLv3 license, and at the bottom of the README there's a licensing info that states: "By contributing to [the Wildland project], you agree that your contributions will be licensed under the GPLv3 License."

The client is currently GPLv3 https://gitlab.com/wildland/wildland-client/-/blob/master/CO... but I don't see any repo for their website, nor the standard, whitepaper, whatever else, and their bootstrap and tutorial repos are All Rights Reserved

Reply on news.ycombinator.com