ONIX code in metadata

[eggheadbooks1]

Hello:

I am trying to embed my BISAC categories and keywords into my ePub's metadata. The ONIX for Books, section P.12.6, gives the example of delivering both BISAC and keywords using reference names:

	
    <Subject>	
        <MainSubject/>	
        <SubjectSchemeIdentifier>10</SubjectSchemeIdentifier>	
        <SubjectSchemeVersion>2012</SubjectSchemeVersion>	
        <SubjectCode>GAR006000</SubjectCode>	
    </Subject>	
    <Subject>	
        <SubjectSchemeIdentifier>10</SubjectSchemeIdentifier>	
        <SubjectSchemeVersion>2012</SubjectSchemeVersion>	
        <SubjectCode>GAR028000</SubjectCode>	
    </Subject>	
    <Subject>	
        <SubjectSchemeIdentifier>20</SubjectSchemeIdentifier>	
        <SubjectHeadingText>garden design; urban; patio; container</SubjectHeadingText>	
    </Subject>

With "10" referring to BISAC, "2012" referring to the BISAC version, "GAR006000" the BISAC code; and "20" refers to "keywords."

Seems straightforward, but when I try to validate the ePub it fails because none of these elements are recognized. So next I tried

<dc:subject>

and that was recognized but none of the others. I also tried putting "dc:" in front of the remaining elements but it made no difference (see SS attached).

Does anyone have any idea how to embed my BISAC categories and keywords into the metadata in a way IDPF will accept?

[JSWolf]

They aren't standard and no ePub reader will do anything with them.

[eggheadbooks1]

The code isn't for the ePub reader; it is intended to be read by the retailer. It gives the user more control over information disseminated with the file. So, for example, if you do not have a way to input keywords -- which you cannot do if you use KWL or you hire an aggregator to distribute your ePub -- you can put them in the metadata. Turns out only ePub3 supports this.

[wannabee]

When you upload an ePub to itunes, google play, amazon, sony reader store, B&N and Kobo there are fields for the meta data in either the spreadsheets or the store upload portal to enter a couple of BISAC codes.

[eggheadbooks1]

Yes, I am aware of that. But KWL, for example, does not have a field to input keywords but the programmers are looking at ways to use embedded keywords to assist searching -- so the keywords have to be in the metadata. Then there is the problem that, while you can input BISAC codes when uploading your file to, for example, Amazon, some BISAC codes are not used by Amazon yet factor into search results. So once again, it's better to have it in the metadata. And finally, as noted in my earlier post, if you use an aggregator you may not have any means to input keywords and BISAC codes. Thus, the ability of the author to input such integral information into the metadata would be a huge step forward.

[wannabee]

I'd be pretty surprised if an aggregator uploaded to one if the shops I mentioned via a method that excludes a bisac.

[Hitch]

Originally Posted by wannabee

I'd be pretty surprised if an aggregator uploaded to one if the shops I mentioned via a method that excludes a bisac.

I've worked with pretty much all of the aggregators, and they all require either ONIX metadata input or they have simple Excel spreadsheets in which they require the publisher (for those uploading via FTP) to input the BISAC data. This is all transferred to the various retailers, again, via FTP, with each book. Their intake software, whether it's Apple, Amazon or Nook, etc., reads the fields in the provided document (XLS or XML) and populates the correct data to the correct location for each book prior to/during publication to their platform. They absolutely do not read the book metadata--that's provided for display on the individual reading devices, at least, it is at this time.

(This data--spreadsheet or XML correlation after the FTP is also why all entities with real publisher accounts at these retailers, as opposed to, say, a KDP account, all use ISBN's; so that the data correlates correctly. When using an FTP upload, the book is named ISBN.epub (or mobi), the cover is ISBN.jpg and the metadata is all ISBN-driven. FWIW.)

I haven't seen any aggregator options that exclude Keywords; I'm surprised to see that one would do so. Of course, keywords can be easily included within the metadata, but, as discussed, the retailers at this time aren't reading the metadata from within the ePUBs. It's something of a chicken-egg scenario. Or a dog chasing its tail; you can put the keywords in there...but nobody's reading them.

As there's no ePUB standard for it, I'd be surprised to find any aggregator that would pass an ePUB with (no offense) "kluged" metadata into it, and as they aren't reading the metadata, until KWL tells you that they are, you are kinda beating your head against a rock. Amazon's use of "categories" for searching is created and maintained by them--just like not everyone gets the "leave a review" link at the end of their book from Amazon, not everyone gets the extra categories, either. Amazon's bots and algorithms determine what additional categories (or BISACs) are assigned to any given book.

Hitch

[st_albert]

Our experience is precisely as Hitch described.

AFAIK, no aggregator or distributor digs into the epub metadata for anything.

That said, if I really really wanted to include BISAC codes, or their human-readable equivalents, in the metadata, I would simply put it within the <dc:subject> tags like any other keywords.

We do use the <dc:subject> tags, although I am unaware of any software that makes use of them at the moment. Maybe Calibre? Kind of like the Voyager project; we send it out there, not knowing who is listening....

Albert

[Hitch]

Originally Posted by st_albert

Our experience is precisely as Hitch described.

AFAIK, no aggregator or distributor digs into the epub metadata for anything.

That said, if I really really wanted to include BISAC codes, or their human-readable equivalents, in the metadata, I would simply put it within the <dc:subject> tags like any other keywords.

We do use the <dc:subject> tags, although I am unaware of any software that makes use of them at the moment. Maybe Calibre? Kind of like the Voyager project; we send it out there, not knowing who is listening....

Albert

Hey, Albert:

Ditto, we use the subject tags, also, and the description tag; but like you we don't kluge the BISAC codes themselves in there. No point to it, as humans generally don't search that way. We use them for the theoretical Deep Web and because we think that it's likely that libraries my send out searchbots someday through the DB's looking for keywords, rather than BISACs which they can pick up from the actual metadata/ONIX input.

Best,
Hitch