Lesser used Dublin Core Metadata and Attributes

Nabodita · 06-14-2017, 01:20 AM

Hello

I've been doing a little research for an internal training workshop; I found that several Dublin Core properties are not described in detail in the OPF specification... nor could I find any specific examples of usage. The specification also provides for two optional attributes: xml:lang and xsi:type and I'm having trouble finding usage examples of the latter.

I've constructed the following metadata entries based on my present understanding of the Dublin Core documentation. Any suggestions, corrections (especially corrections!) or links to further reading material would be sincerely appreciated!

The dc:date property
Possible usage:

Code:

<dc:date opf:event="publication" xsi:type="dcterms:W3CDTC">2017-06-14</dc:date>

The dc:type element
I haven't seen this property used in any ebook, yet it seems fairly straightforward..
Possible usage:

Code:

<dc:type xsi:type="dcterms:DCMIType">text</dc:type>

The dc:format element
I haven't seen this used anywhere either..
Possible usage:

Code:

<dc:format xsi:type="dcterms:IMT">application/epub+zip</dc:format>

The dc:language element
This, of course is a mandatory element; in this case, my question is that RFC3066 has been made obsolete, first by RFC4646 then RFC5646... so is this code correct:

Code:

<dc:language xsi:type="dcterms:RFC3066">en</dc:language>

or should it be:

Code:

<dc:language xsi:type="dcterms:RFC5646">en</dc:language>

The dc:source element
I'm presuming that the ISBN of the print book from which the ebook is created might be listed here. In that case, how does the xml:lang attribute refine the value of this property?

The dc:identifier element
This, of course, is a mandatory element but the OPF specification clearly states that the identifier may have an xsi:type attribute but I can't find any other information about this.

Last of all, I've sort of given up on the dc:relation and dc:coverage elements... unless someone could give me an example of appropriate usage, please?

Any suggestions and thoughts sincerely appreciated!

Cheers
Nabodita

DaleDe · 06-14-2017, 12:28 PM

There may be some help in our wiki Dublin Core

Dale

Nabodita · 06-17-2017, 01:54 PM

Thanks DaleDe, somehow missed seeing the wiki.

I do have another question, though...

It's recommended that the value for the dc:type property be selected from the DCMI Type Vocabulary. Now, the DCMI Type vocabulary has entries like collection, event, dataset, image, text etc. (as read here). Yet the Wiki says that this property is used to describe the document type such as novel, poem, working paper etc.

What am I missing?

DaleDe · 06-19-2017, 11:55 AM

Quote:

Originally Posted by Nabodita

Thanks DaleDe, somehow missed seeing the wiki.

I do have another question, though...

It's recommended that the value for the dc:type property be selected from the DCMI Type Vocabulary. Now, the DCMI Type vocabulary has entries like collection, event, dataset, image, text etc. (as read here). Yet the Wiki says that this property is used to describe the document type such as novel, poem, working paper etc.

What am I missing?

The DC is not oriented toward ePub books. The type in the wiki or oriented toward books but in general they are the same sorts of things.

Dale

KevinH · 06-19-2017, 01:50 PM

FWIW, only a subset of the Dublin Core has been selected for use inside epubs. That set varies from epub "2" to epub "3.0" to what is the latest epub "3.1" spec.

You really have to read the detailed epub specs to find out the exact subset, attributes allowed, extensions, etc.

For example: Here is the latest epub "3.1" spec for metadata:

http://www.idpf.org/epub/31/spec/epu...c-pkg-metadata

Here is the epub 3.0 spec:

http://www.idpf.org/epub/301/spec/ep...-metadata-elem

And the same for the epub2 spec:

http://www.idpf.org/epub/20/spec/OPF...htm#Section2.2

Whereas the full DC spec is quite extensive and can be much more involved.

Nabodita · 06-21-2017, 01:55 AM

As KevinH pointed out, you really have to go through the OPF specs in detail to know what is supported and what isn't... and believe you me, I've spent so much time going over the OPF spec for epub 2 (which is what I'm still playing around with) that I'm seeing it even when I close my eyes!

So the properties I mentioned are part of the DC subset recommended by the IDPF... and hence the confusion.

I barely know what I'm doing. How am I supposed to teach 3 new boys how to make ebooks?

If one of them asks me why the IDPF has defined properties which no one uses or knows about, what am I supposed to say?

Oh, at least I've managed to confirm that I've got the syntax for the dc:date property correct.

Turtle91 · 06-21-2017, 05:43 PM

Speaking of 3.1 spec....

I looked at the comments and there was a good thread about incorporating series name and volume # into "collection" and "volume" (or something) and that it had been "agreed upon". Then the draft came out and it was nowhere to be found... Then the thread was summarily closed!

Does anyone have any more intel on if/how the spec is going to officially support Series and #??

KevinH · 06-21-2017, 06:19 PM

I don't think it will. I think the spec is now final but I could be wrong.

So you could use a meta tag with property of "collection" or "volume" or some other similar terms from the dc or dcterms namespaces.

The 3.1 spec does allow for multiple dc:title metadata elements. So you could add the collection title as a secondary title with or without a volume or series number.

Turtle91 · 06-21-2017, 06:41 PM

That's too bad

The problem is that without a "standard" way of doing it, there won't be any devices/apps that support the display/use of that metadata in sorting or any thing else.

I guess I'm stuck with using the Calibre metadata which is at least supported by Marvin.

Thanks for your thoughts, though.

kovidgoyal · 06-22-2017, 01:30 AM

I have to say I find the EPUB specs absolutely hilarious. They started with an overly complicated and still utterly inadequate spec in EPUB 2.0. They made it even more complicated and still utterly inadequate in EPUB 3.0 and then they rolled back the metadata spec to something somewhat, but not quite similar to what it was in EPUB 2.0 for EPUB 3.1.

And just to put icing on the cake EPUB 2.0 metadata is incomatible with EPUB 3.0 metadata which is incompatible with EPUB 3.1 metadata and EPUB 3.1 metadata while superficially similar to EPUB 2.0 metadata is incompatible with it as well.

Does anyone take these people seriously?

KevinH · 06-23-2017, 10:40 AM

Unfortunately, you are more than right. They have been horrible shepards of the epub standard, implementing many backwardly incompatible changes for no real benefits of any kind. They seem to not understand about the user's perspective at all and have dismissed the importance of metadata with very little thought at all.

The only good thing that has come from the epub3.1 spec is deprecating silly non-standard things like the switch statement, refinements, erc, and trying to move the spec closer to what html browsers already implement.

If they had thought things through from the beginning, they would have done that, and everone would have implemented a webkit-like widget and general support for features would have been much broader. And any changes would have been backwards compatible.

The IDPF really is out of touch with how people use, read, and catalog their ebooks. It is a real shame.

Toxaris · 06-23-2017, 01:30 PM

That is because they listen to companies like Apple instead of actual users. It is not that the issues were not mentioned in the reactions on the drafts of the specs...

Hitch · 06-24-2017, 04:28 PM

Quote:

Originally Posted by Toxaris

That is because they listen to companies like Apple instead of actual users. It is not that the issues were not mentioned in the reactions on the drafts of the specs...

This.^

Hitch

kovidgoyal · 06-25-2017, 12:00 AM

Quote:

Originally Posted by Toxaris

That is because they listen to companies like Apple instead of actual users. It is not that the issues were not mentioned in the reactions on the drafts of the specs...

The ironic thing is that Apple does not follow the EPUB specs either. At this point, I really dont know for whom these specs are written. They aren't written for users, in that, they dont try to address the most frequent problems actual book readers have with the EPUB format. Big companies such as Apple and Adobe regularly create EPUB viewing programs that dont conform to the specs. And that is not even to mention Amazon, which is the biggest and ignores them entirely. So what is the point of the specs? Who benefits from their creation? Why do people spend significant effort on creating them?

It seems to me that the entire EPUB spec effort has pretty comprehensively failed.

Hitch · 06-27-2017, 07:08 PM

Quote:

Originally Posted by kovidgoyal

The ironic thing is <snip>

It seems to me that the entire EPUB spec effort has pretty comprehensively failed.

Yes, Kovid:

You are surely right.

Hitch

06-14-2017, 01:20 AM	#1
Nabodita Connoisseur Posts: 97 Karma: 44418 Join Date: Jul 2013 Location: Mostly in my own head! Device: Kindle Fire, iPad	Lesser used Dublin Core Metadata and Attributes Hello I've been doing a little research for an internal training workshop; I found that several Dublin Core properties are not described in detail in the OPF specification... nor could I find any specific examples of usage. The specification also provides for two optional attributes: xml:lang and xsi:type and I'm having trouble finding usage examples of the latter. I've constructed the following metadata entries based on my present understanding of the Dublin Core documentation. Any suggestions, corrections (especially corrections!) or links to further reading material would be sincerely appreciated! The dc:date property Possible usage: Code: <dc:date opf:event="publication" xsi:type="dcterms:W3CDTC">2017-06-14</dc:date> The dc:type element I haven't seen this property used in any ebook, yet it seems fairly straightforward.. Possible usage: Code: <dc:type xsi:type="dcterms:DCMIType">text</dc:type> The dc:format element I haven't seen this used anywhere either.. Possible usage: Code: <dc:format xsi:type="dcterms:IMT">application/epub+zip</dc:format> The dc:language element This, of course is a mandatory element; in this case, my question is that RFC3066 has been made obsolete, first by RFC4646 then RFC5646... so is this code correct: Code: <dc:language xsi:type="dcterms:RFC3066">en</dc:language> or should it be: Code: <dc:language xsi:type="dcterms:RFC5646">en</dc:language> The dc:source element I'm presuming that the ISBN of the print book from which the ebook is created might be listed here. In that case, how does the xml:lang attribute refine the value of this property? The dc:identifier element This, of course, is a mandatory element but the OPF specification clearly states that the identifier may have an xsi:type attribute but I can't find any other information about this. Last of all, I've sort of given up on the dc:relation and dc:coverage elements... unless someone could give me an example of appropriate usage, please? Any suggestions and thoughts sincerely appreciated! Cheers Nabodita

06-21-2017, 01:55 AM	#6
Nabodita Connoisseur Posts: 97 Karma: 44418 Join Date: Jul 2013 Location: Mostly in my own head! Device: Kindle Fire, iPad	As KevinH pointed out, you really have to go through the OPF specs in detail to know what is supported and what isn't... and believe you me, I've spent so much time going over the OPF spec for epub 2 (which is what I'm still playing around with) that I'm seeing it even when I close my eyes! So the properties I mentioned are part of the DC subset recommended by the IDPF... and hence the confusion. I barely know what I'm doing. How am I supposed to teach 3 new boys how to make ebooks? If one of them asks me why the IDPF has defined properties which no one uses or knows about, what am I supposed to say? Oh, at least I've managed to confirm that I've got the syntax for the dc:date property correct. Last edited by Nabodita; 06-21-2017 at 02:00 AM. Reason: Update

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Greetings from Dublin Ireland	Niall2701	Introduce Yourself	11	01-16-2011 05:17 PM
Encoding Dublin Core Metadata into XHTML	rogue_ronin	Other formats	38	10-11-2010 08:23 PM
Hello from Dublin, Ireland	jaqian	Introduce Yourself	8	01-12-2010 10:51 AM
Hello from Dublin Ireland..	Peter Williams	Introduce Yourself	4	11-24-2009 01:24 PM
Hello from Dublin	piper	Introduce Yourself	7	09-25-2009 02:32 PM

06-14-2017, 12:28 PM	#2
DaleDe Grand Sorcerer Posts: 11,470 Karma: 13095790 Join Date: Aug 2007 Location: Grass Valley, CA Device: EB 1150, EZ Reader, Literati, iPad 2 & Air 2, iPhone 7	There may be some help in our wiki Dublin Core Dale

06-17-2017, 01:54 PM	#3
Nabodita Connoisseur Posts: 97 Karma: 44418 Join Date: Jul 2013 Location: Mostly in my own head! Device: Kindle Fire, iPad	Thanks DaleDe, somehow missed seeing the wiki. I do have another question, though... It's recommended that the value for the dc:type property be selected from the DCMI Type Vocabulary. Now, the DCMI Type vocabulary has entries like collection, event, dataset, image, text etc. (as read here). Yet the Wiki says that this property is used to describe the document type such as novel, poem, working paper etc. What am I missing?

06-19-2017, 01:50 PM	#5
KevinH Sigil Developer Posts: 8,732 Karma: 5703586 Join Date: Nov 2009 Device: many	FWIW, only a subset of the Dublin Core has been selected for use inside epubs. That set varies from epub "2" to epub "3.0" to what is the latest epub "3.1" spec. You really have to read the detailed epub specs to find out the exact subset, attributes allowed, extensions, etc. For example: Here is the latest epub "3.1" spec for metadata: http://www.idpf.org/epub/31/spec/epu...c-pkg-metadata Here is the epub 3.0 spec: http://www.idpf.org/epub/301/spec/ep...-metadata-elem And the same for the epub2 spec: http://www.idpf.org/epub/20/spec/OPF...htm#Section2.2 Whereas the full DC spec is quite extensive and can be much more involved.

06-21-2017, 05:43 PM	#7
Turtle91 A Hairy Wizard Posts: 3,346 Karma: 20171571 Join Date: Dec 2012 Location: Charleston, SC today Device: iPhone 15/11/X/6/iPad 1,2,Air & Air Pro/Surface Pro/Kindle PW & Fire	Speaking of 3.1 spec.... I looked at the comments and there was a good thread about incorporating series name and volume # into "collection" and "volume" (or something) and that it had been "agreed upon". Then the draft came out and it was nowhere to be found... Then the thread was summarily closed! Does anyone have any more intel on if/how the spec is going to officially support Series and #??

06-21-2017, 06:19 PM	#8
KevinH Sigil Developer Posts: 8,732 Karma: 5703586 Join Date: Nov 2009 Device: many	I don't think it will. I think the spec is now final but I could be wrong. So you could use a meta tag with property of "collection" or "volume" or some other similar terms from the dc or dcterms namespaces. The 3.1 spec does allow for multiple dc:title metadata elements. So you could add the collection title as a secondary title with or without a volume or series number.

06-21-2017, 06:41 PM	#9
Turtle91 A Hairy Wizard Posts: 3,346 Karma: 20171571 Join Date: Dec 2012 Location: Charleston, SC today Device: iPhone 15/11/X/6/iPad 1,2,Air & Air Pro/Surface Pro/Kindle PW & Fire	That's too bad The problem is that without a "standard" way of doing it, there won't be any devices/apps that support the display/use of that metadata in sorting or any thing else. I guess I'm stuck with using the Calibre metadata which is at least supported by Marvin. Thanks for your thoughts, though.

06-22-2017, 01:30 AM	#10
kovidgoyal creator of calibre Posts: 45,307 Karma: 27111242 Join Date: Oct 2006 Location: Mumbai, India Device: Various	I have to say I find the EPUB specs absolutely hilarious. They started with an overly complicated and still utterly inadequate spec in EPUB 2.0. They made it even more complicated and still utterly inadequate in EPUB 3.0 and then they rolled back the metadata spec to something somewhat, but not quite similar to what it was in EPUB 2.0 for EPUB 3.1. And just to put icing on the cake EPUB 2.0 metadata is incomatible with EPUB 3.0 metadata which is incompatible with EPUB 3.1 metadata and EPUB 3.1 metadata while superficially similar to EPUB 2.0 metadata is incompatible with it as well. Does anyone take these people seriously?

06-23-2017, 10:40 AM	#11
KevinH Sigil Developer Posts: 8,732 Karma: 5703586 Join Date: Nov 2009 Device: many	Unfortunately, you are more than right. They have been horrible shepards of the epub standard, implementing many backwardly incompatible changes for no real benefits of any kind. They seem to not understand about the user's perspective at all and have dismissed the importance of metadata with very little thought at all. The only good thing that has come from the epub3.1 spec is deprecating silly non-standard things like the switch statement, refinements, erc, and trying to move the spec closer to what html browsers already implement. If they had thought things through from the beginning, they would have done that, and everone would have implemented a webkit-like widget and general support for features would have been much broader. And any changes would have been backwards compatible. The IDPF really is out of touch with how people use, read, and catalog their ebooks. It is a real shame.

06-23-2017, 01:30 PM	#12
Toxaris Wizard Posts: 4,520 Karma: 121692313 Join Date: Oct 2009 Location: Heemskerk, NL Device: PRS-T1, Kobo Touch, Kobo Aura	That is because they listen to companies like Apple instead of actual users. It is not that the issues were not mentioned in the reactions on the drafts of the specs...

Advert

Advert