|
Indeed. It is also web search engine specific - some of those websites the metadata downloaders hit play nicely, and others are annoyingly finnicky, like not wanting spaces between an authors initials.
There is an extra complication with the Goodreads plugin and it is something I will tweak when I do the next version. The way it works is it fires a search at Goodreads, and then takes the first search result that comes back. However the Goodreads search engine can be "overly helpful" and give you a completely different book if it did not find an exact match (this only applies to title/author searches, not ISBN ones). So I have some code which does some crude checking between the title/author of the search result it found, and that in Calibre, just to make sure you don't get data for an unrelated book back, which is worse than no data at all imho. It can be in some cases that my validation checking is a little "too" strict, and so some extra characters like "(2010)" at the end of a title might reject the Goodreads result for not matching closely enough.
As itimpi says it is a very fine balance, I have tried to take a conservative approach under the belief that no data is better than wrong data. However you have to "help it a little" at times by cleaning the cruft out of your authors and titles.
|