Looks like my post yesterday struck a chord with a lot of people, since there's been much commentary on it from people whose opinions I respect, like Allan Grohe, Rob Conley, and the Greyhawk Grognard. I'm pleased by this, because I think it's a topic worth discussing, but I fear that, once again, my "thinking out loud" has been invested with more authority than I ever intended. I did not say that publishing a "true" megadungeon is impossible, only that it cannot be done easily. I said this for several reasons, some of which I elucidated in the original post and some of which I didn't. Let me be clearer now.
First, I think the hobby has changed a lot since the early days. In particular, I think there's a much greater desire on the part of gamers for "complete" products. That's why exhaustive campaign settings, adventure paths, and monstrously large rulebooks seem to rule the day rather than the more skeletonic offerings of the Golden Age. Game companies, naturally, recognize this and cater to that desire. While there's probably a market for a proper megadungeon product, I suspect that market is small, or at the very least, I suspect it's perceived to be small.
That's a purely "logistical" issue and only explains why we haven't seen a published old school megadungeon product, not why it'd be difficult to do. The more substantive issue, I think, is that creating a proper megadungeon is more art than craft. That makes it hard to "commoditize" in the way that most game publishers like these days. It might be possible to produce a "megadungeon construction kit" product, which brings together a bunch of maps, random tables, and "how-to" essays on the creation and maintenance of a megadungeon, but I'm skeptical. Like good refereeing, I see megadungeon creation as something that you can't learn from a book. You learn by doing.
I realize this opinion will be controversial and I fully expect comments ranging from calling me an elitist to pointing to this or that product as a superb example of teaching good refereeing in book form. Once again, let me be clear: I don't think it's impossible, only exceedingly difficult and I say this because I have never seen an example of the kind of "how-to" product we're imagining and I've seen a lot of gaming products over the years. I've seen some that purport to teach the principles of refereeing or adventure design and none of them ever struck me as being all that successful in their intended purpose.
I come to this perspective from my own experience. I learned to become a good referee after years of simply muddling through and by watching others referee their own games. I've often noted that, in the old days, experienced referees frequently "mentored" their less experienced colleagues, helping them to learn the ropes of this rather unique role. For good and for ill, that's how I learned how to referee. The same is true of dungeon design. It's only through having read, created, and played hundreds of dungeons that I figured out how to make ones that are not only enjoyable but have what it takes to hold the interest of my players for months at a stretch.
Now, maybe, everything I've learned over 30 years of doing could be distilled into a book or other product. If so, I'd be very interested in seeing it. But my gut tells me that most such products would wind up being either a chaotic jumble calling itself a "toolkit" or else a bunch of vapid advice of use only to the inexperienced would-be dungeon designers. There's utility in both those types of products, I think, but neither would be a megadungeon product. They'd be products for making your own megadungeon. And, as I noted yesterday, that's really the crux of the matter: megadungeons are made, not bought. I don't discount the possibility that someone could create products to better support the creation of one's own megadungeon, but that's a far cry from asking for Castle Greyhawk in a box, something that I think that's neither desirable nor likely, but I'd be happy to be proven wrong.