lol. Clearly I’m thinking of another I saw recently on YouTube. There have been several on reflection including one with Herb talking. I’ll be on the lookout for the new one. Thanks for the correction.

No problem, I consider it part of my job of being on the conference staff. :)

You might have seen this one which was posted on his blog only a week ago. (I haven't even seen this yet, just found it): https://herbsutter.com/2025/08/19/my-c-on-sea-talk-video-posted-three-cool-things-in-c26/

I don't know much about reflection. I haven't used it in any other language before. But it seems to me that reflection is all about generating code (or other stuff like binary data) sort of like templates except that reflection can do it by analyzing existing code. But in the end you're generating C++ code that is current standard C++ code.
So why could you not just invent a language that can do that without reflecting? I can't imagine what such a language would look like. But it seems like what you're doing is analogous to using C++ language to generate assembly language. Can there be a "meta-language" that generates the C++ that you want without having to reflect on itself? (And if so, then just go directly to generating assembly.) This makes it seem like reflection in C++ is a kind of patch on C++ to get you there while holding on to backward compatibility. I may be misunderstanding reflection. But if there could be such a meta-language I think I'd like to go there.

BTW, I really like all the cpp2 syntax and I'd use it more for hobby projects except that I'm too addicted to Visual Studio intellisense and syntax color highlighting in normal C++, lol.

If you're familiar with C++ template metaprogramming, C++ reflection is just doing that in a sane, designed way, instead of being an abuse of a feature that happened to be powerful enough to use it this way.

And then the answer to "why not a separate language that generates C++" is the same as with template metaprogramming: You want bidirectional compatibility with the regular C++ code around it. You want your generic code to be able to refer to C++ entities in your program, and you want the C++ code in your program to be able and use the generic code.

@tattwamasi That talk hasn't happened yet. It is three weeks from today.

I see the talk is online now, at least from direct link. I watched and appreciated it. I liked the demos and Max's auto diff stuff was really cool (though beyond what I could follow in that presentation). Thanks for the heads-up.

On my original question, beyond reflecting upon reflection, I am still curious what cppfront's next steps are and what, if anything, would be useful in terms of contributions, testing, and just general support.