I am wondering myself the same. I grew up with NCO and stuff, but then how else do you know which are popular, bad, and critical moves in a theoretical line. Even if you use databases, you see percentages, which often can be useless if the critical moves are not the playedest moves. It can also be the case that top players might start choosing the rarer moves.

Even if you have Informator and NIC YB< they both consider current theory. But if you need to know which are the main lines, and which are sidelines, I still find it difficult to make difference without something like NCO. So unless someone tells you which are the main lines and which are the sidelines, you might end up guessing based on incorrect database assumptions. A current repertoire bok would probably tell you, but if there are no current repertoire boks, then it is kind of like wandering through a forest half-blind.

Actually I dug up this review from when the book came out, which is a pretty interesting read:

http://www.theweekinchess.com/john-watson-reviews/nco-a-preliminary-look

The basic complaint is that some lines are wrong based on already published theory, and that perhaps the authors had only used database compilation and not looked at existing books and articles. His verdict is that the book is good for a club player that does not have access to a database program.

I haven't looked at my copy of NCO for a while. I opened it last night and was browsing. I was surprised by how outdated most of the lines are (its like 15 years old, so I don't know why that surprised me). But I would say at least 30-40% of the lines just look wrong, based on my knowledge of what's currently accepted as "best". 

I don't think the book was ever updated (since 1999 or whatever it was). Is there any alternatives? Or have databases and chesspub basically replaced all encompassing reference books?