Quote:
Originally Posted by kovidgoyal
To be precise, IIRC google sometimes has tags and amazon almost never.
|
Kovid,
I have noticed the same drop in tags retrieval success over the last three months roughly. I ran a metadata download JUST for metadata and not covers last week. Took a VERY long time because I added the Fantastic Fiction beta and turned back on the full set of what I had used at various times. That includes Amazon, B&N, Big Book, Goodreads, Google and Open Library. The conversion rate of adding new tags (previously the searched for books had no tags) was abysmal. Something approaching 1 per cent. I wonder if it's a case of too many cooks in the kitchen. But I DID pare back to Amazon and FF (not knowing that it wasn't working) before running the last search. Those settings were no more successful than my 1 percenter, albeit returning the results in something like a twentieth of the time.
As a matter of strategy, moving forward, do you advise stopping depending on downloaded tags? If it's going to fail, might as well go manual and spend the time taken, if not the lesser effort.
Thanks as always, GM
ONE LAST IDEA: If tags ARE retrievable, say from a specific source I'm not using, it would be great if one could do a metadata DL of Just Tags, Just Comments, Just Meta Data, Just Covers or some intelligent combo of the four. I often have comments without tags. I've fooled around trying to deduct tags FROM comments. The results are not horrible, certainly better than 1 per cent. But I wonder if searching for Just Tags might not be something fairly useful.