Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Aug 11 2021 20:52
    @RubenVerborgh banned @mikeadams1
  • Jan 04 2021 20:23
    @RubenVerborgh banned @WebCivics_twitter
  • Jan 04 2021 20:18
    @RubenVerborgh banned @SailingDigital_twitter
  • May 27 2019 06:08
    User @Mitzi-Laszlo unbanned @in1t3r
  • May 23 2019 06:49
    @Mitzi-Laszlo banned @in1t3r
  • May 16 2019 09:49
    @Mitzi-Laszlo banned @mediaprophet
  • Feb 01 2019 22:04
    User @melvincarvalho unbanned @namedgraph_twitter
  • Feb 01 2019 21:49
    @melvincarvalho banned @namedgraph_twitter
Mark Hughes (happybeing/theWebalyst)
@happybeing
@retog the browser implementation is under review though due to performance and other shortcomings of using Beaker.
@melvincarvalho thanks for those links. Should keep me quiet for a while. Dinner now though - have a good evening. I shall be hanging in later for the MaidSafe Dev update though (every Thursday :smile: )
Melvin Carvalho
@melvincarvalho
@theWebalyst thanks! Im interested in understanding the decentralized DNS.
Reto Gmür
@retog
Twee-fi now runs on its own domain: https://twee.fi/, you can now also view all your reviews.
Melvin Carvalho
@melvincarvalho
@retog cool!!
Mark Hughes (happybeing/theWebalyst)
@happybeing
@melvincarvalho haven't missed your question about dec'd DNS - will get back to you when i get a moment or three
Nicola
@nicola
dropping this project here: peerpad.net
Dmitri Zagidulin
@dmitrizagidulin
oh very cool! @nicola are you working on it?
Mark Hughes (happybeing/theWebalyst)
@happybeing
@melvincarvalho quick response on your SAFE questions above: decentralised DNS and GETDNS first...
Nicola
@nicola
@dmitrizagidulin nope, but it’s built by the IPFS team
Mark Hughes (happybeing/theWebalyst)
@happybeing
@melvincarvalho The following will be wrong on some details, but I hope will answer your question to a reasonable level (so don't take it as hand wavy and not definitive!). The network can be regarded as a big data store, currently for immutable data and mutable data. Immutable data exists in its own address space - with a 1MB 'chunk' of encrypted data at each 256bit address, with the address being equal to the hash of the chunk (the address is in xor space, actual location/IP unknown to anyone and changes anyway as nodes/vaults leave and rejoin). So if you know the addresses of the chunks of a file, you can access them and then decrypt them. To do that you need the 'data map' for the file, which is itself a file chunked and stored on the network. So that's immutable data.
@melvincarvalho mutable data consists of 1MB chunks that live in a different address space - still a hash - but based on the a 'name' you give the chunk rather than the hash of its content (since that can change). Mutable data is versioned, and its content is a key/value store where each key is also versioned. So you can add keys, write and overwrite their values. Like immutable data this is also encrypted but has ownership and other rights controlled by encryption keys, ownership transfer etc.
Mark Hughes (happybeing/theWebalyst)
@happybeing
@melvincarvalho so finally to DNS... a DNS entry is just mutable data (MD) that implements an NFS like file space. So your 'domain' and any services you create are stored in the keys of this MD and the value of a key points to the content for that service. So for a 'www' service, it points to a directory (ie an MD with file entries - as immutable data), where we expect to find index.html. Actually DNS is one type of MD, each with its own address space. Anyone can create (claim) any domain, but once done you own it and nobody else can claim it. The DNS therefore is just a particular way of interpreting the fundamental data types stored on the network, which could be used for may things, not just DNS, but all sharing the same fundamental properties of decentralised, secure anonymous storage on a truly autonomous network (where each user has an un-moderated relationship with the network and complete control of their data and who can access it).
@melvincarvalho So to answer the GET question, the browser is just an app which uses the API to take a safe://www.thewebalyst type URL and interpret it, using the API to access the NFS style file structure that the DNS entry for 'thewebalyst' points to, to access index.html (and any related files) and render it. So there is no separate GET/PUT/DELETE API at the moment (though we did have a RESTful API implementation in the previous iteration of the API so the network provides something along those lines, if not currently an actual RESTful interface).
@melvincarvalho hope that helps! Thanks again for the links you provided. Happy to answer further - here or offline as appropriate.
Mark Hughes (happybeing/theWebalyst)
@happybeing
@melvincarvalho Am I correct in thinking that these are the current active libraries for working with SOLID/Linked Data? https://github.com/linkeddata/rdflib.js and https://github.com/solid/solid-client
Tim Berners-Lee
@timbl
"Anyone can create (claim) any domain, but once done you own it and nobody else can claim it. “ of a new DNS — how do you prevent automated name grabs and NS speculation?
@theWebalyst
Mark Hughes (happybeing/theWebalyst)
@happybeing
@timbl that's a big topic (much discussed on the forum) - not finalised but to be addressed through payment, limits per account, and so on. I don't think there's a perfect solution to this one, so bright ideas are always welcome. for now the team are focussed on securing the network (data chains parts 1 & 2) but then we'll be introducing (proof of resource) 'farming' of Test Safecoin, and this will need to be in place before Test Safecoin can become actual Safecoin (the token used to pay for storage, and to reward resource providers).
I keep saying 'we' but I'm not part of MaidSafe, sorry :-). Just closely following and helping out for a while now.
Mark Hughes (happybeing/theWebalyst)
@happybeing
@timbl I understand from David Irvine of MaidSafe that he would like to talk to folks from your team to share ideas - for example about the best ways to handle private data, especially synchronization of shared private/public data (fast updates). Is this something that you'd be interested in?
Melvin Carvalho
@melvincarvalho
@theWebalyst thanks for the info, interesting!
Tim Berners-Lee
@timbl
@theWebalyst Yes, I’d certainly be interested in a chat David Irvine. Where is he based?
Sarven Capadisli
@csarven
I'm thinking that having a property for WebID can be useful. Loosely implying where a Person's POD may be. owl:sameAs could work too but it is a strong claim and there is no easy way to distinguish one resource from another as to which may be along the lines of "WebID sitting at a POD". Without this, there are many many many Person URIs out there. Signal:Noise :)
Melvin Carvalho
@melvincarvalho
@csarven aren't your WebID and your pod logically distinct? ie
  • A pod may host your files
  • A pod may authenticate you
  • A pod may host your personal profile document / WebID
Sarven Capadisli
@csarven
Like I said, "loosely implying where a Person's POD may be"
For existing profile descriptions out there, it is a matter of extending with foo:webid so that one can discover canonical WebID or POD may be.
Melvin Carvalho
@melvincarvalho
@csarven given that webid and pod are disjoint, dont we already have pim : storage for that? Would perhaps foo : pod be a more expressive name? Regarding canonical WebID, I think @timbl used preferredURI. Seems to me a few different conversations are going on.
Although I have been in favor of a : webid predicate before to indicate what your webid is.
Sarven Capadisli
@csarven
Sure, preferredURI is fine too. Does that exist under some ns?
tho for some reason I didnt work out it takes a literal string
I think we do need a webid predicate still tho
Sarven Capadisli
@csarven
Aside: we acknowledge that a URI of a Person could potentially be their WebID, but it doesn't have the same semantics as a WebID that someone can write to per se. Many Person URIs exist in the wild and most are not writeable by the people that the URI is about. So, a unique WebID relation can signal that I think.
Mark Hughes (happybeing/theWebalyst)
@happybeing
@timbl David is in Ayr, Scotland. I gather he's snowed with planning the alpha3 work for the next few weeks so around end of November is more feasible. Would that suit you and do you know where are you likely to be then? Videoconf is possible wherever of course but if you are not too far away maybe face to face would be possible.
Melvin Carvalho
@melvincarvalho
@csarven agreed, tho you can find out from the header if it's writable using the MS-Author-Via
which might be a better place for that information
Sarven Capadisli
@csarven
It is also a good place to look up for their storage
ie:
example:Melvin foo:webid <https://melvincarvalho.com/#me> . <https://melvincarvalho.com/#me> pim:storage bar:storage .
Melvin Carvalho
@melvincarvalho
  • A WebID is defined as a URI that denotes a person
  • You can point to your storage (pod) using pim : storage
  • You can point to your oidc auth (pod) using solid : oidcIssuer
What's missing?
btw I like foo : webid
e.g. you could use this in a git config file to tell it your WebID
I think it's also useful to have a webid predicate which can be used in a bnode
Mark Hughes (happybeing/theWebalyst)
@happybeing
@timbl BTW do you do podcast interviews? One of our community guys (John Ferguson) does them on SAFE related and similar technology and he's very good as well as being a really decent guy. He calls himself "the simpleton of SAFE" and does a good job of translating from geek to everyperson! :-) I know he would love to interview you!
Mark Hughes (happybeing/theWebalyst)
@happybeing
I made a short twitter thread summarising the SOLID Spec Overview: https://twitter.com/safepress/status/925390948565241856
Melvin Carvalho
@melvincarvalho
nice, liked! :)
Sarven Capadisli
@csarven
@melvincarvalho What do you think of acl:owner or maybe a distant acl:delegates even?
To fulfil the role of foo:webId