Readers may be interested in my approach to the problem, or rather, the problem that remains even after any terminology issues are settled.
Summary: What we identify as "qualia" is the encoding of memories that we cannot yet compare directly between people, to the extent we can't compare them. This incommensurability can easily arise among agents who are similar, but who self-modify in a way that does not place any priority on the ability to directly transfer memories to other agents.
In that case, their methods of storing memories are ad-hoc, and look like garbage to each other -- but with the right assumptions and interaction, they can achieve a limited ability to compare, and thereby have terminology like "red" that means something to all agents, even as it doesn't call up exactly the same idea for each one.
I was intrigued when I first read this when you last posted it, and I thought about it for a while. The problem with it, it seems to me, is that this is a good explanation for why qualia are ineffable, but it doesn't seem to be come any close to explaining what they are or how they arise.
So, I could imagine a world (it may even be this one!) where people's brains happen to be organized similarly enough that two people really could transfer qualia between them, but this still doesn't explain anything about them.
This is our monthly thread for collecting these little gems and pearls of wisdom, rationality-related quotes you've seen recently, or had stored in your quotesfile for ages, and which might be handy to link to in one of our discussions.