The problem with this argument is that it's easily possible to write data to a CD-R with no errors whatsoever. I won't get into the problems associated with aging CD-Rs, and I don't know enough about the differences between how data is written and read and how audio data is. But if CD-Rs didn't record the exact bits of data that you requested that they record, then they'd be useless, and I think enough of us have written real data to CD-Rs to know that that sort of thing just doesn't seem to happen, at least not on a regular basis.

Actually, now that I say that, I realize that I do know one significant difference. The ISO-9660 filesystem contains a good bit of error correction. So if you need to write immutable data, use a filesystem, not just audio tracks. And, assuming that that's the purpose of using CD-Rs to master your data, audio or not, one would wonder why you wouldn't do it that way anyway. We, as a community, have been developing ways to extract audio data off of CDs better and better for years now. But we know how to extract data from a filesystem in an optimal manner, so why not just record it that way anyway? Of course, you'll need to eventually record it down to an audio master if you plan on pressing CDs (I assume), but premature opimization has been a problem since the beginning of computers. I guess we're getting to the point where even laypeople need to know about that.