ERIS @ Datenspuren 2021 [Log-Down Light]

@emery presented ERIS at Datenspuren 2021, the annual event of <<</>>.

This slide shared by Astro shows the relation between ERIS and the original Xanadu project by Ted Nelson, whom we cherish a lot here.

I would love to see the whole deck @emery :slight_smile: Here is the video:

1 Like

Yes, Datenspuren needed talks so I though ERIS was an appropriate submission. I was a bit nervous and worried about time, so some of it was a bit muddled. @pukkamustard should have gotten more credit as well.

The talk is a bit dry, so if you want to make to it more exciting, take a drink for each of the following:

  • exponentiation (two to the power of…)
  • use of the word “garbage”
  • wikipedia citations
  • spelling errors
2 Likes

We actually just finished watching it and we did not need any drinks. It was quite insightful, and the conclusion about “billability” is worth a good laughter.

Your Janastu project seems really cool! I will definitely look into it.

Also I saw you demonstrating ERIS uploading files into qtox, I’m curious how you did that.

Besides, and more importantly, I think you make a great case for ERIS by introducing all the various approaches before that, and you also clarify very well the advantages and limitations of this approach. There are a lot of questions asked and it will be great to follow the evolution of ERIS.

One aspect of the limitations that falls straight into my actuality is related to surveys and the possibility to ‘pre-hash’ all answers to lift the protective effect of encryption. It’s an actual issue because Liberaforms just submitted a grant application including some research around distributed poll storage to avoid, e.g., health data leaks, using ERIS and DMC. Their effort remains valid IMO but they should be careful with the case you mention, where answers’ hashes can be pre-computed.

The mention of libgen confirmed that indeed it’s a very good approach to share the load of storing stuff like sci-hub or libgen. Or a collection of books or music albums. Anything really that once published does not change. Unless we’re using DMC…

I’m still a bit confused about immutable identifiers and content-addressed identifiers…

:+1:

To say that I work on the Papad application is an exaggeration (or a simplification), I’m in the Janastu meetings now and then, and I wrote the HTTP gateway for them, but I haven’t written any of the Papad code.

At the moment Papad is using Syncthing to distribute recordings, if we commit to using ERIS the migration path will probably be to continue using Syncthing, but use ERIS URNs within the application to refer to files replicated by Syncthing, and then move to native ERIS storage.

On the forms, it was the best example I could come up with for fill-in-the-information. The easy fix is to put a nonce into every submission, but of course sometime people forget these things.

The qtox upload demo was using a bot that I wrote a while ago with an ERIS store patched in for the talk. I wrote the HTTP server first and I was using PUT to upload files, but using a chat protocol solves authentication and works well visually.

1 Like