I suppose there is a certain amount of 'pride' that makes us want to keep Daphne as a separate project, but the real problem is the incompatible licenses of the code bases (GPL vs. strict non-commercial.) We can share techniques, but not the code itself, which increases the amount of work to move things from one platform to the other. MAME may go GPL in the future, though, which would help matters.
Also, much of the 'magic' of Daphne lies in the high-performance MPEG2 playback and searching, but the MAME devs have said they do not want an MPEG-based solution in MAME due to patent concerns. (The fear is that the MPEG consortium that owns the patents will some day start demanding license fees, just like Fraunhoffer has now done with MP3.)
Until the MAME devs can find a patent-free, high-performance, cross-platform video codec, lasergames won't make it into the official builds. If they (or we) can obtain such a codec, I'm not opposed to working on game drivers for MAME. Since we've already worked out a lot of details about the game hardware and LD player emulation, I'd like to see our names in the MAME dev credits list...
Oh, and re: the target data for MACH3, MAME's solution is to pre-decode the audio data track to a binary .dat file, and use that directly during emulation, rather than decoding it in realtime. Perhaps this isn't as 'pure' as using the raw audio stream, but MACH3 was the only game that used such a system, so a generic approach isn't really that useful anyway.
The main thing that's keeping us from adding MACH3 (and DL2!) to Daphne is the lack of an integrated 8086 CPU core. There is at least one GPL'd core out there, but so far no one has done the work to hook it into Daphne. Any volunteers?
Oh, and thanks to all for your continued interest and support of the Daphne project. There's been a lot of buzz about the unofficial MACH3 driver for MAME... I hope people don't forget Daphne's accomplishments too quickly....
-Warren.