Key Words

5 posts / 0 new
Last post

I use Adobe InDesign CC to create ePubs. In the File info palette, as well as adding Document Title, Author and Description metadata. You can also add Keywords. When I export to ePub and i open the file the Keywords appear in the content.opf file in the dc:subject, but I'm told the Keywords are not implemented in ePub. Does anyone have any information on this matter?

You mean, there is a Dublin Core element subject for each key word? (not just one for all?)
With some viewers the audience has access to this, but most viewers have in general bugs or gaps in interpreting the meta data from Dublin Core, often the authors are not recognised or not all authors or in general not the function of a creator or contributor. Some hide rights and license conditions or they remove them, if a format conversion takes place.
This means it depends on the bugs and gaps of each viewer, what is accessible to the reader and what not.
There is no common implementation strategy for different viewers - but most are somehow incomplete - but this has no dependency on EPUB directly. Unfortunately for some meta data the notation method of EPUB 3 differs from EPUB 2, therefore old viewers will not necessarily do it right for EPUB 3 books, even if they do it for EPUB 2, but for Dublin Core subject the same notation can be used.
Therefore I provide all meta data additionally within an XHTML document in the spine to work around such annoying bugs and gaps for human reading and access in case the viewers fail.

Yes - A dc subject element for each key word(s) - here's a bit of mine:
<dc:title>The Hastings Road and the &quot;happy springs of Tunbridge&quot; </dc:title>
<dc:creator>Charles G. Harper</dc:creator>
<dc:subject>Kent history</dc:subject>
<dc:subject>Tunbridge Wells</dc:subject>
<dc:subject>local history</dc:subject>
<dc:subject>stage coach</dc:subject>
<dc:subject>social life and customs</dc:subject>

This should be extracted by some viewers. Often they claim, that the creator is the author, what is not always true, but in general there should appear something, if the viewer does it right.
Try for example EPUBreader for Mozilla/Gecko viewers like firefox or lucifox or lucidor and you should find, that they extract this. I think, calibre will extract subjects as well, but has some major other gaps for meta data, but your simple list might work with calibre.

Currently for example you do not mention the role of the creator, but if C. G. Harper did all the work on this EPUB archive and no one else contributed, this is ok.

What a reading system chooses to expose to the reader doesn't quite fall into "bugs and gaps", even if what is used (and how) can be frustrating (e.g., using the last dc:creator as the author, which actually is a bug).

There are few requirements on reading system presentation or use of metadata, as when you start getting to that level of reading system conformance you start to impede the ability of developers to create the interfaces they want. The metadata is primarily to facilitate development of bookshelfs, and those are completely outside the scope of the EPUB specifications.

That said, from the content creator perspective, I'd never leave out any metadata. As OH notes, there are some reading systems that will expose the information. It's better to make your publication as rich as you can and let the reading systems sort out what they can do with the information than strip it down to what seems to be supported.

You also never know how development will progress. If you include the information, and the reading system interface changes to accommodate it in the future, then all your content is immediately compatible instead of having to be re-released.

Secondary menu