Its a difficult concept, AB, because there are so many differing opinions not only on performances but also what constitutes core repertoire. I know what you're saying but I rarely (with some exceptions) enjoy vocal works for example and complete operas. To others thee works are essential.
I take that point. My thought was that you -for example- wouldn't have mentioned Peter Grimes for Britten, but someone else might. Clearly, everyone's "Building a Library" contribution will be derived from their own experiences and/or preferences. But I think if you were somehow to review
that Tchaikovsky thread I mentioned, that worked quite well. Lots of different opinions expressed, which is fine, and someone mentioned Orchestral Suites as being a 'must' that others had not. It all seemed to work out fine in the end though!
So apart from disagreement on core repertoire you have problems with performance interpretations. I can think of quite a few recordings that polarise opinions (a certain performance of Beethoven's 9th, for example that is lauded by some here but described as "terrible" by myself and a certain well-known music critic :lol

. Its so difficult recommending anything to anyone.
Obviously so, I think. But if people had in mind, "What would I recommend to a novice, in such a way that it wouldn't befuddle him or her", I think that would help. The advanced nuances of performance practice can be left for a more skilled audience, probably. I quite like Hurwitz on Bach, for example, a month or so ago: yes, there are HIP performances; yes, there are "modern instrument" performances; both are fine, but today I'm going to do the modern performances, because they're valid too.
Sure. If people get precious about things, and protecting their turf, as it were, then I can see it wouldn't work. But, to take one example, if in a discussion of Bach we kind of agreed that the Goldbergs were essential listening, I wouldn't have a problem with someone recommending Gould. Despise it though I do, I quite understand that a well-rounded listener at least needs to have heard the Gould in order to form an opinion about it!
Even the old Penguin Rosettes were contentious for some. I used the old Penguin guides to start to build a collection but found that there were much more interesting recordings around. I actually felt that over those early years I wasted a fair amount of money listening to other people's recommendations.
Well, I think that's where things have improved a lot since then -because now you can listen to someone's recommendations on Spotify, for example, for nothing at all (at least, as I understand Spotify's free tier service). There's no drawback in having recommendations when it doesn't cost someone actual cash to follow up on them.
That said, obviously the idea is to 'channel' exploration to some extent, so I wouldn't want to see 58 recommendations for a work. That's just the same as having none at all! But a smattering of different suggestions seems a reasonable thing to do. On the Tchaikovksy, we got three symphony cycles recommended before someone piped up for Petrov (I think it was): I was committed by that point, so the extra recommendation fell on stony soil!
With the advent of streaming theres no need to take a shot in the dark any longer as you can listen to many of these on Spotify, etc. You csn sample on youtibe, etc for free and try out repertoire and find what you like.
I should probably read your whole reply before answering to bits of it! I see we are of the same opinon... but seem to be coming to different conclusions! The fact that one can experiment for free is a huge improvement on the days of having to either borrow a crappy, scratched copy from a library or spend up big and keep your fingers crossed. But I see that as not only a great opportunity, but a monstrous wall of choice, in the face of which people... simply refuse to choose. Without any information about what or why they should listen to something, they simply won't. It doesn't mean that if you recommend Smirnov's Beethoven 9th, then no-one will ever listen to Blenkinsop's. Spotify means that they will listen to Smirnov's because you told them to... and then they'll be able to listen to lots of other versions and then make up their own mind. But a pointer in a vaguely right direction is an indispensible first step, I think.
Confession time: I had a lot of Beethoven Symphonies. None of them made particularly coherent sense and I generally didn't listen to any of them. Then I read your multi-multi-part guide to a bazillion cycles, made my choice from that and now have 4 I listen to quite a bit. Lose the wall of choice; focus down to a few recommendattions... then people can learn to find their own way forward. It's worked for me with you and Beethoven; it\s working for me with Tchaikovsky... I figure it can work for other people, too.
Then theres the problem of tastes changing over time. Years ago I rarely enjoyed many string quartets apart from a few warhorses but now they are core to my listening. If id made a list of core repertoire in 2002 it would look very different today. Dont get me wrong, im not being critical, its just that its so difficult to make recommendations to anyone. When I do my Beethoven reviews I know there are some people sat at home thinking "this guy is talking utter *****, conductor X's Beethoven is fantastic." It doesnt stop me but even I change my mind over time.
That's fine too. Plastic information is a plus. The trouble with those Penguin Rosettes is they couldn't be updated in real time. We're not constrained in that way today.
And I don't think changing tastes is an issue. Again, look at the Tchaikovsky thread I keep banging on about. 6 people mentioned symphonies, ballets and a couple of opera. Only then did one lone voice pipe up and mention 'Orchestral suites, anyone?", and then lots of people piled in and said 'Oh, of course, forgot about them". If
you are not a String Quartet person today, then someone else will be. That's the whole point of a collaborative effort in a public forum, I think.
Short version: I wouldn't over-think this. Come up with (for example) a 'core Beethoven' corpus, throw it out there and see what happens. The worst that can happen, surely, is that 43 pages later, no-one agrees on anything. But chances are, if it was phrased as 'this is how I would start with Beethoven', I think people would constructively add to it, but wouldn't -I'd hope- try to argue subtle nuance at interminable length.