|
[Smile on] Sorry that you can't find a Apple solution... [Smile off]
To be honest, why do you think, this is an calibre problem? You are on the wrong side of the chain. You need to go to the source and this is obviously not calibre. calibre is only pulling what it becomes from metadata source providers and nothing else. You can take a look into the protocol and see what exactly providers are sending back. It often is crap and quite a wonder if you get the correct information. Use their API's and test it yourself if you don't believe it. An other point is that users tend to provide too much information for a query. E.g. if you have a ISBN available, provide only this for a query and nothing else. A ISBN is mostly unique for a publication and every data more will give you more not accurate results.
I did and do this quite often and all I can say is that you will be faster if you do it manual and check the data afterwards with other sources you may trust (in case you want accurate metadata).
Your problem in the end is, that there is no source available that have for all books of the world all kind of metadata available. Also publishers and authors (equal if they are self publishers or not) do not provide accurate metadata at all. It is a pain in the neck that they are not able to produce a correct set of DC metadata these days. It seems to be quite new for them. The description is public available since end of the nineties last century...
I'm doing this kind of research since 10 or 15 years and can trust you this is a never ending mess and quite impossible to find a source you can trust to 100%. It begins with publishers and authors who don't want to do this job accurate. Best resources I use are mostly national library systems around the world. But the quality there also is depending on how old the metadata are as metadata systems and definitions change and grow during the past decades. This is why The Dublin Core Metadata Initiative (DCMI) came in place long time ago (driven by libraries and their systems).
And because you mention Amazon - this is for my metadata searches one of the worst source of metadata. They not only change quite often their API (I feel like they don't want that anyone else use their API's), the "quality" ... is most time I take a look to it also a mess.
By the way, there is a nice plugin from DaltonST that is quite good when you need more than Amazon quality. It is called Library codes. It queries DC metadata from Library of Congress, VIAF, ISNI and Worldcat and save you a lot of manual work if data are available. An other available plugin is Citronalco's DNB_de (metadata from Deutsche Nationalbibliothek API, you need to ask for access and register at DNB before you can use it).
As a last point, feel free to make your own plugin for collecting correct metadata. calibre have a wonderful and well documented interface - you only need to use it. You can also try to improve / modify existing plugins. Guess, Kovid will be quite happy to implement your modifications once they are available.
|