That's really not a big issue with the large library rules. I'm more concerned about the fact that a library based off a L6Q21 summae let's someone study 21 times at quality 21. 441 experience. Plus 24 more if you start below 6. That's enough to take someone from zero to thirty.
I'm not familiar with these rules, but how in the world is a level 6 summa pushing a summa above level 6? I already don't like it.
Large libraries tend to consist of a mix of texts, mostly tractatii.
The idea behind the rules is that when you're studying at or below the level of the library, you're studying from various summa. Once you pass the level, you're into the tractatus collection - at which point you're now limited in the number of times you can study from it.
One thing I immediately spot: a tractatus struggles to get a quality above 14, so getting 21xp per season even when you're past what the library summa collection is strikes me as odd.
The problems the previous posters mention all stem from the fact the Transforming Mythic Europe's ideas for simplifies stats for large libraries in fact only make sense for large libraries.
The example with the L6Q21 summae is easily seen as mathematically broken. If you look at just the one book. Agreed, it does seem excessive that you can study for 21 seasons for Q21.
The rules are for when you have so many books that individual stats and tracking which books have already been read are tedious and all but impossible. IMHO it would make sense to use 'summed up stats' if you have 10+ books on the same Art. Also, these new rules also create some abstractions and changes as to how study normally works. IMHO having a large number of books allows you to cross-reference and gain some synergu from all the different books, rather that just reading them one at a time for consequetive seasons.
Specifically, these (optional) rules are not for when you have less than 10-15 books on a given subject. That's not a large library, why are you using a large library approximation?
I think I'd do things a little differently.
Not really thought-out, but I'd probably have something along these lines:
Level and Quality are considered as per the Abilities XP table.
Tractatii add their Quality to Library Quality, twice that to Library level.
Summaes add their quality to Library Quality XP, and twice their level to Library level XP. Realia work as summaes.
You might want to lower quality XP, but you get the idea. This represents that you can study from multiple sources, and that even a Q10 summa may have insights useful to a master of an art.
As an option, if your art level is > Library level, yu can still study from half quality, scrapping bits here and there. If it is twice that, yu've read everything.
That seems pretty reasonable. Although if you wanted to expand the rules to abilities I would divide tracti quality by 5 for level experience.
The problem is that in order to get the library score you are still tracking individual books, and at that point why vary from just tracking the individual books?
In theory the maximum level for any subject in a library should be the highest summae level plus the experience from all tracti on the subject. So if you are buying levels for a library you need to either raise the highest summae or be buying tracti. As for a generalized quality of a library, if the library is large enough to use large library rules I would think there should be an element of chance to that as you might not find the best option amongst all those books- perhaps quality would be the library level in the subject plus an organization score minus your level plus a stress die, where botches count as negatives on the total score and organization also reduces botch dice.
Durenmar's library rules are good large-library rules. No upper cap that I'm aware of, but plenty of reasons why you can't just study from it constantly.
If a covenant library is large enough to be 'large', why not?
I've played around with the numbers a lot since my OP. We are talking several messy pages of scrap paper filled with hand written tallies. Not anything I'll bother transcribing but I'll share some of my conclusions.
First thing is you probably can't get a really high Level library with these rules. Raising a library from Level 20 to Level 24 takes 90 summae. YSMV but I'm not sure I expect there to be over 90 fairly high value summa for any given art in the Game.
I agree there is IMOHO a problem with Low level (extremely) high Quality summa like the aformentioned L6Q21. I think there are a couple of easy house rules/gentlemen agreements that keep these texts from breaking verisimilitude.
Fortunately the problem caused by locking out books isn't as big a deal as but it is still annoying. In a couple of tests, sets of books that best case scenario could raise both Level and Quality by +4 each only ended up being about half as effective when added in a worst case scenerio order. (Either L+2 Q+2 or L+0 Q+4 depending on your definition of worst) So in a real game situation where story guides are putting some thought into what texts to seed into the story it should become mostly a non-issue.
And you are talking a lot of books. Unless you start with a collection with relatively low scores, the library's Level and Quality won't rise very high above the core text without adding dozens and dozens of books.
There are a couple of cool things about these rules though. If the library rules are "part of the setting" it means that almost all books have some value. Summa and even Tractus that no one would ever read on there own can still be valuable additions to a library. Particularly high L low Q summa make great (possibly abusive) core books for a Library, and will still usually be able to add experience to the Level of most established collections. So authorship is no longer just a game for high com good teachers. Almost anyone with an art score 50+ should consider writing a book IMOHO.
It also make's sense to keep certain books out of collections using these rules. Great works and other books with scores that are almost impossible improve using the library rules should probably never be added to collections. But this makes sense to me in setting and becomes a feature not a flaw. You have your treasured books on display or locked away for special occasions and your "stacks" for general reference and study.
A lot will depend on how much magi (especially young magi) are encouraged to write (and at what quality). Considering that every magus with level 20 in an art can write 4 tracti on it, if a library were to simply charge for its use in tracti it shouldn't take too long for it to improve it's capacity.
Much of that falls under YSMV. But a bunch "Vain" Tractati with a score of 6 or less can still usefull in raising a libraries quality to 10. IMOHO a positive is fairly common in the order so more magi then not should be able to produce Tractati in the 7-9 range. That would make it "realistic" to achieve libraries with Qualities pushing 14.
No more, no less, than with the RAW large library rules 
If you just want to abstract things a lot, you don't have to: Don't bother asking yourself if you already have a copy of book x by author y in your library, just buy/exchange that q10 lvl 10 summa, and let's assume you're not dumb enough to have it already.
If I'm tracking individual books, I want titles (and perhaps author names) for all of them in order to differentiate between them. Plus I need to keep track of who read what already. However Metacreator can make this easy for me, and I use this tool all the same. But it isn't required, you could keep track of books read with pen and paper.
But If I use the large Library Rules I merely need to track the Summae plus the number of Tractatus - numbers only! Also, I need only to keep track of how many seasons I've studied with a score above the Library's level rating. Again, numbers only not titles.
Well, you still need to track books (by name and author) added to the library (these can be really flavorful) so you don't have duplicates increasing the score, unless you abstract the numbers completely. You just don't have to track them for each and every character.
Strictly speaking, no, you don't need to track books in the library by name an Author. You can go "I brought a list of books in our library, and go shopping in (friendly covenant)'s library. How many tomes can I find that we don't have already?" and make it completely abstract. Yes, this diminishes the impact of each tome, but we're already talking about libraries so large that we're using the Large Library Approximation ("rules").
And yes, book titles and authors can be extremely flavourful, but that's not relly the issue at hand, eh?
Well, to be fair, I did say unless you abstract it completely, after that, so you are taking my statement out of context.
So, yes, you can completely abstract everything. But what is the issue at hand? It's a discussion about the large library rules. You can have a large library with each tome listed by name and author. Having that information doesn't impact the library.
I like the idea of flavorful books. I dislike having to track them individually per magus. The rules provided do need a bit of polishing, but I haven't had time to look at them in depth.
The thing is that at a certain point more books is more hindrance than help. If you already have another book with a higher level and quality, why would you buy an inferior book to add to your library instead of hiring someone to make a copy of the existing text so more people can use it? Accumulating new tracti is always helpful, but the prices in Covenants don't really reflect that.
That's not the way the large library rules work. They assume that taken and used as a collection a large quantity of books can in some ways be greater then the sum of parts. 23 summae none with a level limit higher then 10 could give the library a Level 12. And as long as a lone readers level is below the library's they get a +3 study bonus over and above it's established quality. And any one can continue to study from the library after reaching the library's level for a total number of times equal to it's quality.
There are down sides. Added tracti increase the quality, usually to levels higher then any one text but ultimately a post Level Limit library will usually provide less xp then reading each Tractus seperatly should provide.
Using the rules seems to imply you gain speed at a loss of total potential.
The problem is that it kind of flies in the face of the rules for books themselves, and strikes me as a Macguffin rather than fitting the setting. At what point doe a few level 10 books suddenly become an effective level 12 book? Or to push it a bit further if I have an exceptional level 25 book how many level 10 books do I need to have sitting around to read my way up to level 27 or 28 in that art?