dms3 draft codec request

[yozgatsi]

please consider this request to register a public codec for an early stage project. \thank you

[rvagg]

Would you be able to provide a little bit more information for the description field? And are you expecting to be needing a lot more entries around this or can I suggest you move it down among the other numbers just to avoid stretching the far end of the spectrum unnecessarily. Or is there something about this number that is a nice fit for your project?

[yozgatsi]

DMS3 = Decentralized Multi-State-Space Search (IR) platform on a p2p network. Nodes will leverage the technology, but form a separate network from that of IPFS/Libp2p. I am fine fine to move the number to compact the codec space. I don't expect to need a large block of numbers, and it's difficult to anticipate all my needs at this time. If you allowed block reservations, an order of 10 may be appropriate.

[rvagg]

Suggested change

	dms3, multiaddr, 0xd03000, draft, dms3
	dms3, multiaddr, 0xd03000, draft, Decentralized Multi-State-Space Search

[rvagg]

There's a ton of space above 0x100000, whole blocks of the first digit in those are missing, there's no good reason to jump to d, you could even just take 0x100000 if you want. I'd just prefer we not stretch the space out further if there's not a good reason to. If it spelled something or the number had some relevance for you then that'd be enough, otherwise, just choose a lower number and you're good. If you pick a number with some space after it then you're welcome to come back later and allocate more. Particularly in this end of the number space, people are reluctant to bump right up against another un-related entry so you should be fine if you choose a block in that end of the table.

[yozgatsi]

Hi Rod, thank you for the suggestion, works for me. Please let me know if you can apply the following changes. \Tavit

dms3, multiaddr, 0x100000, draft, Decentralized Multi-State-Space Search
dms3-key, ipld, 0x100001, draft, DMS3 Public Key

[rvagg]

@yozgatsi please update this branch with the changes or, if you have trouble doing that feel free to close this and open a new pull request with it.

[yozgatsi]

seem I got it right on the second try to modify the PR...

[rvagg]

Suggested change

	dms3-key, ipld, 0x100001, draft, DMS3 Public Key
	dms3-key, key, 0x100001, draft, Decentralized Multi-State-Space Search Public Key

unfortunately you can't use ipld if this isn't a codec that can decode bytes into something that can contain links; if it's a codec that doesn't involve links then it's a serialization, but given you're saying it's a public key then I suspect it's not even that.

I'd suggest it's probably just a key, but you also may want to consider the name dms3-pub to match the other public key types. But this also raises the question: is this actually a key type? Can you just use one (or more) of the existing key types for this?

[rvagg]

this objection still stands, can't be ipld here

[yozgatsi]

It is an alternative to, and modeled according to the entry:
libp2p-key, ipld, 0x72, permanent, Libp2p Public Key
If you believe this is problematic or needs more definition, then I can remove that entry now and keep only the multiaddr entry. we can revisit the former at a later time. Please advise, and thanks again for your help.

[yozgatsi]

Hi Rod, Is there any update or timeline for this?

[yozgatsi]

Hi Rod,

A global, not just IPFS/LIBP2P specific, multicodec namespace registry can lead to conflict-free co-existence of multiple p2p networks in the wild.

The administrative role for such a registry would serve a function similar to the IANA.

If maintaining a global registry is outside the scope of this repository's mission, then say so and let's close this pull request.

I have a working p2p network that assigns the two codec values suggested above form the private-use range of values. My preference is to be transparent with all stakeholders by asking the maintainers of this repository to assign known codec values as discussed above.

Thanks
Tavit

[rvagg]

@yozgatsi I already made it fairly clear what the next steps are, please see my review above saying that you can't use ipld for this, you'll need to change the label type before we can proceed and I also suggested a more descriptive entry on the line to make it a bit clearer what it is. #366 (comment)