Music, of any genre, is never about reproducing the notes and dynamics accurately. MIDI does that, and the day we can create historically informed MIDI we will have it perfected.
Rather, music is about the spontaneous communication and individuality that occurs WHILE reproducing those notes and dynamics. Genres differ in which dimensions of freedom the musician can use, but not in that fundamental objective. In classical music it's what makes it worthwhile to hear different interpreters of the same piece. It's why it's worth the 5 year waiting list to see Wagner in Bayreuth, when arguably the greatest rendition of The Ring in 100 years (Solti, Chicago, fight me) is available relatively cheaply everywhere.
If that's the case...
1) Encourage your classical friends to break out of the mindset of "cooking the recipe the master chef made exactly." Audiences complain that classical music is sterile and doesn't speak to them, precisely because this mindset is so common among all but the most elite performers. Also, notice that NO elite performers approach their musicianship that way.
2) Notation and theory must have a slightly different purpose than the one you articulated. I believe they're simply there as tools of communication between musicians, to describe recurring patterns we hear. Otherwise we'd be forever saying things like "Beethoven 3, is the one that starts with that thing where it sounds like it's going to end but really doesn't." Calling it a "deceptive cadence" or even a "surprise VI" is just way easier and lets us communicate (and operate) on a higher level of abstraction. It's the same way "dependency injector class" tells you about a given chunk of code, and lets us reason about its place in the larger structure. A great engineer isn't defined by their ability to reproduce a textbook dependency injection class, but rather by their ability to adapt the concept of dependency injection well, in the right places and times.