I like this proposal the most. We should decide a name of which layer to use as the name for the whole thing in the case of two-layer HF. And then keep the same notation as we use to day for spec docs, e.g. Capella for CL, Shanghai for Engine API specs.
Using a mixed name is a good one but I don’t like it because everyone will have to keep three names in mind when talking about two-layer HF.
Note that we will run out of Devcon city names in case of >1 HF per year. Devcons to be happening in locations defined by EL’s HF names is one of potential solutions to this problem.
Please can we just stick with what we have for the execution and consensus layers? We already have a history of past names and a plan for future names, so any change will cause disruption and add to the technical debt of the ecosystem.
Naming things is hard, because we don’t know what the future brings so cannot pick something that will work well with future changes (Ethereum is a good example of this). But consistency brings its own benefits, and reduces the mental load when coming in to the Ethereum space. Taking the consensus layer as an example, we are now on a path “A,B,C,…” for each hard fork, which makes it very easy for someone coming in to the space to know the ordering of the hard forks. Execution layer requires a bit of tribal knowledge, but that is now a given and not going to change so the best we can do is not to increase their burden but having multiple systems that happen at different points in time.
As for having a user-facing name for the upgrades, I maintain that the best solution here is to talk about the execution layer name and not the consensus layer. Anything that is consensus layer-specific is likely to only matter to validators, and they are generally speaking more aware of the difference between the layers and understanding of what is activated when (the BLS to execution change operations are a good example of this). If we consider users to be those that may generate execution layer transactions, or benefit from execution layer transactions, then the thing they care about is upgrades in the execution layer so that it is what we should be talking about.
Kindly birthday note beacon chain, and the community!
conversations? of a posthumous naming scheme layer. Examples; Hawking,
Name is identity, whether we imagine a planet, city or else. In my mind, I compare updates/upgrades to historical artworks. “Pink. Election” draws pictures of our Consciousness. However; the term and the story behind delegate curiosity via different tenses.
Given that @protolambda is letting go of Duhbe (due to some people having difficulty spelling), it looks like Deneb might be the default winner for the D-starname.
Could we perhaps make our own images and get them standardized in unicode? This is partially a joke, and partially a serious proposition.
Alternatively, maybe we can get the Ethereum logo standardized in unicode, and then we can add new ZWJ emoji sequences that, say, combine the Ethereum logo and the owl/swan.
It feels like we’ve naturally converged towards this proposal, with people using Shapella (and now Dencun) more and more. Another property that’s nice for this is you can tell by looking at the name whether it’s a CL/EL-only upgrade (star or city) vs. a CL+EL upgrade (combined name).
Example: Ethereum X.1.0 “Shanghai Capella”
(I’m not sure what the X would be)
Things like gas repricings and the DAO fork would be patches, things like adding opcodes and new transaction types would be minor versions, and things like the merge (since it removed the original functionality of the DIFFICULTY opcode) and any SELFDESTRUCT changes would be major versions.
Please no - first, it’s no fun! Fun is a big part of Ethereum - this forum is literally called EthMagicians !
Second, too much confusion with “Ethereum 1.x”, “Ethereum 2.0”, etc. That said, once we’ve cycled through all of the A-Z stars, and have caught up to devcon(nect) cities + all major conferences, I’m fine letting that go. Ethereum should be boring infrastructure by then
Encodes a [CBOR] structure tagged with the data type [URTYPES], and is therefore self-describing.
Uses a dictionary of exactly 256 English words with a uniform word size of 4 letters.
Only two letters of each word (the first and last) are required to uniquely identify each byte value, making a minimal Bytewords encoding as efficient as hexadecimal (2 characters per byte) and yet less error prone.
Additionally, words can be uniquely identified by their first three letters or last three letters.
( for 
Caneb (Cancun + Deneb)
!