Hmm. Now I'm imagining an alternate reality where Holmes really dies at Reichenbach and every story from "The Empty House" onward is just Watson fabricating fictional Holmes cases to keep the series going, because The Strand offered him too much money to turn down. Maybe he based them on real cases from Holmes's notes and wrote himself into them.
Watson explicitly states this in the ending of Dibdin's
The Last Sherlock Holmes story. (This is not a novel for the purists. It breaks
all the toys.)
That's one thing that bugged me a bit when I recently rewatched the Jeremy Brett series, and that sometimes comes up in other Holmes adaptations, where Watson writing up the cases for publication in The Strand is mentioned or shown within the stories. Given that a lot of Holmes's cases involved highly confidential and sensitive matters that could have ruined the reputations of a lot of important people, you'd think that Watson would've fictionalized the names and personal details when he wrote up the stories. So seeing it dramatized as if we're seeing the real cases that Watson then adapted, but with the characters having the same names that they had in the published stories, seems incongruous.
I think back to our discussion last year in the TrekLit forum about "the King of Bohemia" in "A Scandal in Bohemia"--there's a long-standing belief in literary fandom that the name and character is a substitution to obscure the
real person involved in the scandal. ("The King of Bohemia" is commonly believed to be the Prince of Wales.) So I view adaptations this way...
Actual Events -> Watson's version of events with details & names obscured -> Adaptation using Watson's version of events
In that sense, while I get where you're coming from, the adaptations are already two steps removed from the actual events Sherlock Holmes was involved in. Henry Baskerville may not have been named Henry Baskerville in the real events of
Hound... or maybe he was and he told Watson it was fine to be named in Watson's account.
I'm not a great fan of pastiches that have Holmes and Watson interacting with real, historical people because I just don't think Watson would have written the stories that way. David Gerrold talked at the Farpoint convention in February about how much he enjoyed writing a story that brought together Holmes with, IIRC (I didn't take notes), Oscar Wilde, and while yes, it's an interesting intellectual puzzle to solve (Gerrold talked about how he had a very short window of time, like 18 hours, in which this story could fit into Wilde's life), it's also not Watsonian in the slightest.
I made a slight exception for Holmes versus the Ripper pastiches, because if you're going to put Holmes into those events, you
have to use the historical people and the canonical murders. It's why I don't care for Ellery Queen's
A Study in Terror; all of the Ripper details are fictional.