| DNF seems useful to me. The other deficiencies you mention don't matter much to me, though the Goodreads UI is bad. (Don't care if it looks old, but right-clicking on things often doesn't work because they're stupid JS link, and it's way too focused on being a social feed.) As far as features, the things I value about Goodreads / would like from replacements: * I can see what my friends have read and rated books. (I specifically care about a tiny handful of friends who I know have similar tastes to my own.) Obviously, this will be very hard for you to replicate. Note that I do not care about the social feed. It doesn't matter that my friend is currently reading xyz. All that matters is that when I look at book X, I can see that [friend with very similar taste] read it (don't care when!) and whether they rated it highly or poorly. * Goodreads recommendations are bad. Related to the above point. Goodread's "Reads also enjoy" wavers from "moderately useful" to "fundamentally broken". Lists are okay, but broad lists are dominated by super-popular books that came out post-Goodreads, and there aren't enough specific lists (or not specific in a useful way; so many pointless lists for "books with an X on the cover".) After importing my Goodreads books to Kaguya, and checking the recs on my latest read, you have roughly the same issue that Goodreads has with obscure books: the "similar books" are...other recent books I recently read. (On Goodreads I'll see this when my brother and I read an obscure scifi novel and something 100% unrelated. He and I will dominate the data set for that book.) I recently discover that LibraryThing's similar books system is actually much better than anything on Goodreads for finding similar books. That's the thing to beat. I suspect they're able to do better because they've built up a large dataset from deep tagging similar to what you're planning. A tagging system isn't enough; you'll have to tag things! That requires a lot of users, copying an existing data set, or some clever LLM use. (Maybe not so clever; it's probably been trained on a lot of these books.) * Libby integration I have a Chrome extension (Available Reads) that will hit the Libby API for each book on a page. This lets me quickly see which books on my To Read list are currently available as ebooks or audiobooks at my local library. It's useful enough that I open up Chrome (not my primary browser) just to use that extension from time to time. * Filters and sorting on genre I have a big "want to read" list. When I want to pick my next book to read, I usually am thinking "I want to read scifi" or "I want to read about history" or whatever. In Goodreads, my Want To Read list doesn't even let me filter or sort fiction from non-fiction. What would be cool would be if I could filter "1980s scifi that I haven't read that my brother has read and rated at least four stars". Or "middle-grades scifi that I read 10+ years ago and rated 4+ stars" for when I want a nostalgic read. The "more stats" I think could be interesting, albeit not particularly useful. I do like the "date read / date published" graph on Goodreads stats. (The "date published" axes gets squashed to uselessness if you read "The Odyssey", though.) |
1. Friends reviews on book is actually not that hard at all. We will implement it right after we get the friends and follow system working.
2. This one will take a decent number of users and ratings. We need a lot of data before we can make recommendations using ML.
3. Noted.
4. We will add that in the next few weeks. Filtering TBR seems to be a common request.