25 May 2020 |
hellekin | This is how we got involved with NGI0 (along with CCT, the Center for Cultivation of Technology, in Germany) | 14:04:37 |
26 May 2020 |
@tg:disroot.org | In reply to @pdecrat:matrix.org But I'd be interesting in talking about your organizations and programs, I'm working on these topics as well what topics/context are you working on? | 19:53:45 |
Chris Gebhardt | In reply to@how:public.cat thanks pdecrat. Do you have a legal entity for IC? We were thinking about using ps to tap into the 200K figure, and dispatch to indidvual partners. Would you like it? I'm looking into starting a non-profit, but nothing currently exists. Help and advice is definitely welcome! | 20:41:12 |
Chris Gebhardt | (I'm in the US, so not sure what cross border options may exist..) | 20:43:42 |
Chris Gebhardt | In reply to@tg:disroot.org Chris Gebhardt: and what's your plan for turning these ideas into practice once the design proposal is ready? My general plan is to pursue simultaneous formal spec development and prototype implementation of the core Repository and DMF (Data Management Foundation). (Spec and dev work can help inform each other) IE, the user-space side, can be minimalistic for now, perhaps thin adaptors to existing UI / web frameworks. | 20:49:49 |
Chris Gebhardt | * My general plan is to pursue simultaneous formal spec development and prototype implementation of the core Repository and DMF (Data Management Foundation). (Spec and dev work can help inform each other) IE (Information Environment), the user-space side, can be minimalistic for now, perhaps thin adaptors to existing UI / web frameworks. | 20:50:39 |
@tg:disroot.org | In reply to @chrisgebhardt:matrix.org (I'm in the US, so not sure what cross border options may exist..) https://ngiatlantic.eu/ -- this for example but i dont know much about it, perhaps hellekin knows more | 20:54:35 |
hellekin | NGI Atlantic has about the same deadline as NGI Pointer. It requires two entities (one in the US and one in the EU), and it's aimed at startups as far as I understand. This could be a route to explore, but we need a legal entity in the US. | 20:56:57 |
@tg:disroot.org | hellekin: do you know if there are going to be more calls for ngi atlantic? | 21:05:22 |
@tg:disroot.org | i found the answer, there are going to be 5 calls, and as far as i see this is more for experimentation on testbeds like fed4fire.eu | 21:08:04 |
@tg:disroot.org | so this will be good for the experimentation/evaluation phase but first need to develop a prototype | 21:09:04 |
@tg:disroot.org | hellekin: can PS work with non-EU partners? | 21:09:57 |
hellekin | sure, PS is not nationalist :) | 21:30:37 |
@tg:disroot.org | then the pointer seems more suitable | 21:31:17 |
hellekin | we already worked with australians and brazilians among others | 21:31:55 |
28 May 2020 |
| robert.best joined the room. | 03:58:29 |
4 Jun 2020 |
Chris Gebhardt | Quick update: (which I should do more often!) Progress continues quickly on the design overhaul, but the changes are pretty heavy. I'm now working on first class support for collections in the Persistent Data Model, based on Merkle trees but not using path-based referencing or naming. I realized I need this collections support for aggregate targets of metadata, large data handling, and efficient ontology component management / gathering / import. This will likely not be used for "semantic graph level" collection support, as with membership statements. Everything in PDM is raw data management oriented, so collections here need to be about access and aggregation of data entities and their payloads. | 00:01:07 |
17 Jun 2020 |
@pukkamustard:matrix.org | Hi Chris Gebhardt and everyone! I've been working on a proposal on how to make RDF content-addressable (https://openengiadina.net/papers/content-addressable-rdf.html).. Work is very much inspired and influenced by InfoCentral, although it only covers a small subset of InfoCentral.
I'd be very happy for feedback and comments. We also have a JavaScript demo that shows how the scheme works: https://openengiadina.gitlab.io/js-eris/
| 09:33:50 |
7 Jul 2020 |
| @jodie2099:matrix.org joined the room. | 17:42:14 |
| @jodie2099:matrix.org left the room. | 21:55:56 |
8 Jul 2020 |
Chris Gebhardt | Sorry for late reply.. Riot never notified me of the new message here. :( | 17:15:44 |
Chris Gebhardt | pukkamustard: Cool! I'll take a look and compare my notes on similar. | 17:16:12 |
hellekin | Chris Gebhardt: actually you're timely, as we were notified yesterday that we got the funding to pursue TG's and pukkamustard's related research. | 17:19:00 |
Chris Gebhardt | nice. Congrats! | 17:50:05 |
9 Jul 2020 |
Chris Gebhardt | In the ERIS design, what is the motivation to always encrypt the Merkle tree (branch) nodes given that the data blocks (leaf nodes) are already encrypted? | 09:13:14 |
hellekin | I guess it avoids leaking information about the tree... Surely pukkamustard has a better explanation :) | 09:16:05 |
@pukkamustard:matrix.org | That's exactly the reason. The branch nodes should be indistinguishable from the leaf nodes. If this were not the case an adversary could learn the lenght of the encoded content. | 09:25:15 |
hellekin | it's all leaves :) | 09:25:51 |
@pukkamustard:matrix.org | :) | 09:26:23 |
@pukkamustard:matrix.org | ECRS from GNUNet (https://grothoff.org/christian/ecrs.pdf) also encrypts leaf nodes. They have a nice description of the threath model and what an adversary could do. | 09:27:21 |