[penagwin]

I mean once we account for all the different types of DNS records - regardless of its original intent, isn't it essentially just a networked, hierarchical key store? For example the TXT field is "dns".

This project is still doing key -> value. It just fetches the value from Wikipedia first, much like your normal dns servers have to fetch non-cached keys from their sources (other dns servers normally)?

[sfoley]

Just because it's doing key-value pairs does not mean it's DNS. If I can't `dig` it, it's not DNS. This is simply doing HTTP redirects and works with no other protocols.

[capableweb]

Hm, there are plenty of DNS records (not to mention all the custom ones all around the world) that you won't be able to `dig` but most people would still call DNS.

[icedchai]

Can you provide an example of one that you can't 'dig'? I have my doubts.

[capableweb]

One example: I can't seem to get dig to work with URI records (but I might be missing some flag). Doing `dig URI _kerberos.hasvickygoneonholiday.com` returns "no servers could be reached" while doing `kdig URI _kerberos.hasvickygoneonholiday.com` returns the proper records.

So seems to be a proper DNS record, but I can't make it work with dig.

[icedchai]

Plain old "dig" works for me! I suspect it may be an older version of dig you're using? This is DiG 9.11.5-P4-5.1ubuntu2.1-Ubuntu on Ubuntu 19.10 ...

[capableweb]

Strange, but thanks for letting me know! I'm on my Ubuntu laptop now, so `DiG 9.11.3-1ubuntu1.11-Ubuntu` and it works too! But initially tried on my Arch Linux desktop, where it didn't work, and I would expect my desktop to run a more recent version than my laptop. Very strange, will take another look.

[capableweb]

Yeah, agree! I'd take it a step further and say it doesn't even have to be "networked" (in the technical sense) but could be local or even done person-to-person, which would also work albeit be slow.

Let's call that sneaksolving (from the infamous sneakernet/FloppyNet)