Are bookmarks made to an edition or to a work?

9 views
Skip to first unread message

Anthony G

unread,
Dec 7, 2010, 8:31:10 AM12/7/10
to Open Bookmarks
I was looking through the wiki when I saw this:

"What is the universal ID for a book? ISBNs are likely to be included,
but they refer to editions, not to works."

The implication seems to be that bookmarks are thought of as being
made in a work, regardless of which edition of the work is being read.
If that's the case, I can foresee problems: works change between
editions, particularly in non-fiction but in fiction too.

Imagine Dave makes a bookmark in an edition of Monkey Training for
Dummies and exports it. At some point, he buys the new edition; the
world of monkey training has moved on. He tries to import his bookmark
from the old edition, but the text he bookmarked is no longer found at
the same position marker. Or it's no longer in the book at all - he'd
bookmarked something about whips but more humane methods now prevail.
What happens then?

Just a thought. I'm afraid I don't have a solution to offer at the
moment.

Anthony

Dave Pawson

unread,
Dec 7, 2010, 8:54:11 AM12/7/10
to openbo...@googlegroups.com
On 7 December 2010 13:31, Anthony G <anthony...@gmail.com> wrote:
> I was looking through the wiki when I saw this:
>
> "What is the universal ID for a book? ISBNs are likely to be included,
> but they refer to editions, not to works."

True. And (generally) to a media specific version.

> Imagine Dave makes a bookmark in an edition of Monkey Training for
> Dummies and exports it. At some point, he buys the new edition; the
> world of monkey training has moved on. He tries to import his bookmark
> from the old edition, but the text he bookmarked is no longer found at
> the same position marker. Or it's no longer in the book at all - he'd
> bookmarked something about whips but more humane methods now prevail.
> What happens then?


Seems right that the bookmark is edition+version specific,
since the page reference (or link target) could change
between the two.

Perhaps the task is to provide a x-ref method of importing
bookmarks (where possible) from edn n to n+1
or media A to media B. Onerous to say the least.

regards


--
Dave Pawson
XSLT XSL-FO FAQ.
Docbook FAQ.
http://www.dpawson.co.uk

James Bridle

unread,
Dec 7, 2010, 11:51:24 AM12/7/10
to openbo...@googlegroups.com
I think the language needs to be amended, but the intention stands.

When I wrote "editions" I was referring to printings, formats and
publications, i.e. the same book as a paperback, as a hardback, as
published separately by different presses.

An *updated* edition of a book is a different book in this context; a
different format or printing is not. Open Library would give an
updated edition a new work number, so there isn't a problem with the
proposed ID in this context.

In addition, the system should be simple enough that Dave *would* be
able to import his bookmarks where they were still relevant into the
new edition of Monkey Training. Many of them would still be findable.

James

Katong

unread,
Dec 10, 2010, 12:09:05 AM12/10/10
to Open Bookmarks
Can't we expect more frequent and incremental updating in ebooks, at a
level below what would trigger a new edition? In fact, shouldn't we
welcome that possibility? I certainly expect self-published authors to
update relatively continuously.

In my (very limited) experience of the process for submitting ebooks
to current vendors, there is no big problem in updating an
amended .epub file, if the metadata stays the same.

In addition to ISBN and Open Library work numbers, should we consider
the emerging ISTN standard as an identifier? Tracking changes to and
location in a text in relation to its ISTN seems more logical than
tracking at an ISBN level, esp when the same text (single ISTN) has
multiple ISBNs.

Peter

On Dec 8, 12:51 am, James Bridle <jamesbri...@googlemail.com> wrote:
> I think the language needs to be amended, but the intention stands.
>
> When I wrote "editions" I was referring to printings, formats and
> publications, i.e. the same book as a paperback, as a hardback, as
> published separately by different presses.
>
> An *updated* edition of a book is a different book in this context; a
> different format or printing is not. Open Library would give an
> updated edition a new work number, so there isn't a problem with the
> proposed ID in this context.
>
> In addition, the system should be simple enough that Dave *would* be
> able to import his bookmarks where they were still relevant into the
> new edition of Monkey Training. Many of them would still be findable.
>
> James
>
> On Tue, Dec 7, 2010 at 1:54 PM, Dave Pawson <dave.paw...@gmail.com> wrote:

Dave Pawson

unread,
Dec 10, 2010, 3:16:31 AM12/10/10
to openbo...@googlegroups.com
On 10 December 2010 05:09, Katong <peter.s...@gmail.com> wrote:
> Can't we expect more frequent and incremental updating in ebooks, at a
> level below what would trigger a new edition? In fact, shouldn't we
> welcome that possibility? I certainly expect self-published authors to
> update relatively continuously.

+1. When someone feeds a really silly errata back to the author,
the first they they (should) want to do is update it for other readers.


> In addition to ISBN and Open Library work numbers, should we consider
> the emerging ISTN standard as an identifier? Tracking changes to and
> location in a text in relation to its ISTN seems more logical than
> tracking at an ISBN level, esp when the same text (single ISTN) has
> multiple ISBNs.

Perhaps someone more in the know than I can confirm it, but
AFAIK it is a long process and costs money to get a new ISBN?

Perhaps its day is done and something new is needed.
I'd suggest one form of URI
http://www.w3.org/Addressing/URL/uri-spec.html

Stéphane Mourey

unread,
Dec 10, 2010, 3:38:06 AM12/10/10
to openbo...@googlegroups.com
Hi,

I've 2 very partial answers to this question:
  1. if you use the Lulu website (http://lulu.com), they can provide you an ISBN instantly and freely, but, as they handle the editor part of the number, you cannot use it with an other editor. I don't know if it have a cost to get an editor number, but I suppose that when you got it, you can use it to produce book numbers freely.
  2. In France, it's also free to get an ISBN, but you have to send one copy of the book to the "Bibliotheque Nationnale de France" (the nationnal library in France). As I did not use this way to get an ISBN, I do not know if the number is provided after or before they received the book, and I don't if you can proceed online. But it seems to take some times to complete the whole process.
Best regards,

Stephane Mourey

2010/12/10 Dave Pawson <dave....@gmail.com>

James Bridle

unread,
Dec 10, 2010, 5:08:30 AM12/10/10
to openbo...@googlegroups.com
Hi Peter -

Can you point us towards some more information about the ISTN?

James

Stéphane Mourey

unread,
Dec 10, 2010, 5:19:44 AM12/10/10
to openbo...@googlegroups.com
Hi,

Maybe he was writting about ISTC (http://www.istc-international.org) ?

Cheers,

Stephane Moruey

2010/12/10 James Bridle <james...@googlemail.com>

Richard Nash - Cursor

unread,
Dec 10, 2010, 6:21:18 PM12/10/10
to Open Bookmarks
Yup, that is it, Stephane.I have heard that adoption has been very
very slow. You could email Laura Dawson who sits on the Book Industry
Study Group's committee that deals with identifiers for her sense as
to where adoption is at.

On Dec 10, 5:19 am, Stéphane Mourey <stephane.mou...@gmail.com> wrote:
> Hi,
>
> Maybe he was writting about ISTC (http://www.istc-international.org) ?
>
> Cheers,
>
> Stephane Moruey
>
> 2010/12/10 James Bridle <jamesbri...@googlemail.com>
>
>
>
>
>
>
>
> > Hi Peter -
>
> > Can you point us towards some more information about the ISTN?
>
> > James
>

Katong

unread,
Dec 10, 2010, 7:47:24 PM12/10/10
to Open Bookmarks
Sorry for the sloppiness! Yes, ISTC - International Standard Text Code
(http://www.istc-international.org/). Would be great to hear the
latest from BISG, my understanding is that the registration companies
(Bowker, Neilsen) were going to deploy in their systems before pushing
publisher adoption. It would be good to know the latest.

You could see ISTC as an attempt to solve for the difference between
the work (text) and the product (particular edition). ISBN tracks
products, not works.

Eric Hellman has a recent post making this point -- highly relevant to
Open Bookmark. See http://go-to-hellman.blogspot.com/2010/12/biblio-social-objects-copia-mendeley.html.
His view is that social annotations mostly likely want to refer to the
work, not the particular product.

He mentions other attempts to solve the work/product identification
problem including xISBN (from WorldCat) and Library Thing.

He also mentions http://www.openannotation.org/ , which is a Mellon-
funded equivalent of openbookmark, focused on academic books. I
confess I haven't dug into it yet, but will do so. The scholarly
market has a handle system for identifying works (Digital Object
Identifier see http://www.doi.org/) across changing or multiple URLs,
and a consortium driving reference linking (CrossRef see http://www.crossref.org/)
between DOIs. Both are very well developed now, though they focus on
journal articles over books.

Peter

On Dec 11, 7:21 am, Richard Nash - Cursor <rn...@thinkcursor.com>
wrote:

Katong

unread,
Dec 10, 2010, 7:59:59 PM12/10/10
to Open Bookmarks
Book Industry Study group and others have looked at the updating
problem vs ISBNs, and have this to say:

"Digital technologies allow for content – whether delivered online,
downloaded or printed using
POD – to be in a state of continuous revision. This has implications
for identification and version
control. While the ISBN standard has stipulated that “a separate ISBN
shall be assigned if there
have been significant changes to any part or parts of a publication”,
publishers should follow
the basic principle that identifiers will be assigned only where there
is a clear and compelling
need to do so. In most cases, dynamic changes in digital content will
not meet this criterion.
The same principle applies to the updating of files within a POD
context. For most changes it
will not be necessary or desirable to assign a new ISBN."

See www.bisg.org/docs/DigitalIdentifiers_07Jan08.pdf

I believe there are well established systems for identifying updates
and changes in technical documents (not to mention open source
software development), perhaps these could be useful for our
discussion (assuming we are still trying to link to sentence level).
Numbering systems that look at edition.release.update, that sort of
thing.


On Dec 10, 4:16 pm, Dave Pawson <dave.paw...@gmail.com> wrote:
> On 10 December 2010 05:09, Katong <peter.schopp...@gmail.com> wrote:
>
> > Can't we expect more frequent and incremental updating in ebooks, at a
> > level below what would trigger a new edition? In fact, shouldn't we
> > welcome that possibility? I certainly expect self-published authors to
> > update relatively continuously.
>
> +1. When someone feeds a really silly errata back to the author,
> the first they they (should) want to do is update it for other readers.
>
> > In addition to ISBN and Open Library work numbers, should we consider
> > the emerging ISTN standard as an identifier? Tracking changes to and
> > location in a text in relation to its ISTN seems more logical than
> > tracking at an ISBN level, esp when the same text (single ISTN) has
> > multiple ISBNs.
>
> Perhaps someone more in the know than I can confirm it, but
> AFAIK it is a long process and costs money to get a new ISBN?
>
> Perhaps its day is done and something new is needed.
> I'd suggest one form of URIhttp://www.w3.org/Addressing/URL/uri-spec.html

Dave Pawson

unread,
Dec 11, 2010, 1:59:51 AM12/11/10
to openbo...@googlegroups.com
On 11 December 2010 00:59, Katong <peter.s...@gmail.com> wrote:
> Book Industry Study group and others have looked at the updating
> problem vs ISBNs, and have this to say:
>
> "Digital technologies allow for content – whether delivered online,
> downloaded or printed using
> POD – to be in a state of continuous revision. This has implications
> for identification and version
> control. While the ISBN standard has stipulated that “a separate ISBN
> shall be assigned if there
> have been significant changes to any part or parts of a publication”,
> publishers should follow
> the basic principle that identifiers will be assigned only where there
> is a clear and compelling
> need to do so. In most cases, dynamic changes in digital content will
> not meet this criterion.
> The same principle applies to the updating of files within a POD
> context. For most changes it
> will not be necessary or desirable to assign a new ISBN."
>
> See www.bisg.org/docs/DigitalIdentifiers_07Jan08.pdf

That makes it sound like burying their heads in the sand?


" dynamic changes in digital content will not meet this criterion."

>


> I believe there are well established systems for identifying updates
> and changes in technical documents (not to mention open source
> software development), perhaps these could be useful for our
> discussion (assuming we are still trying to link to sentence level).
> Numbering systems that look at edition.release.update, that sort of
> thing.

Except that it's any book rather than just technical books surely?
I see this as a far more general problem.

stml

unread,
Dec 14, 2010, 6:55:07 AM12/14/10
to Open Bookmarks
I have to say, that having dealt with them on many occasions, I'm
pretty reluctant to rely on anything involving Bowker/Nielsen and ISBN/
ISTC, as these closed systems are antithetical to an open approach.

In particular, there's the question of identifying a work when saving
the annotation. Most works will include the ISBN, but few (none?) will
include the ISTC, and there's no free and open way of finding or
obtaining it. However, when based on an open identifier like
OpenLibrary, it's possible to query for and obtain this for free.



On Dec 11, 6:59 am, Dave Pawson <dave.paw...@gmail.com> wrote:

Dave Pawson

unread,
Dec 14, 2010, 6:59:58 AM12/14/10
to openbo...@googlegroups.com
On 14 December 2010 11:55, stml <james...@googlemail.com> wrote:
> I have to say, that having dealt with them on many occasions, I'm
> pretty reluctant to rely on anything involving Bowker/Nielsen and ISBN/
> ISTC, as these closed systems are antithetical to an open approach.
>
> In particular, there's the question of identifying a work when saving
> the annotation. Most works will include the ISBN, but few (none?) will
> include the ISTC, and there's no free and open way of finding or
> obtaining it. However, when based on an open identifier like
> OpenLibrary, it's possible to query for and obtain this for free.

Anyone know the legal position about using the ISBN, then
adding revision data to that?

Without a better option, a formalized extension (ISBN - format - edition - date)
etc might be viable?

Richard Eoin Nash

unread,
Dec 14, 2010, 7:47:20 AM12/14/10
to openbo...@googlegroups.com, Open Bookmarks
Oh yeah, institutionally it's like interacting with slightly toxic molasses...

_____________
http://ThinkCursor.com
http://RNash.com @R_Nash

Dave Pawson

unread,
Dec 14, 2010, 8:16:55 AM12/14/10
to openbo...@googlegroups.com
On 14 December 2010 12:47, Richard Eoin Nash <rn...@thinkcursor.com> wrote:
> Oh yeah, institutionally it's like interacting with slightly toxic molasses...

Heck of an image that Eoin :-)

I'll try and remember that one.

stml

unread,
Dec 14, 2010, 9:26:21 AM12/14/10
to Open Bookmarks

> Anyone know the legal position about using the ISBN, then
> adding revision data to that?
>
> Without a better option, a formalized extension (ISBN - format - edition - date)
> etc might be viable?

We can *use* the ISBN, it's just not much use for what we're doing.
I'm not sure how the extension helps - we'll still have to go
elsewhere to aggregate works.

But I don't want to reinvent the ISBN or any other system. I want to
choose a better one. Arguments *against* using the OpenLibrary ID?
They have promised support, have backing from the Internet Archive
for the long term, and are committed to the same principles of
openness.

James

Dave Pawson

unread,
Dec 14, 2010, 9:53:27 AM12/14/10
to openbo...@googlegroups.com
On 14 December 2010 14:26, stml <james...@googlemail.com> wrote:
>
>> Anyone know the legal position about using the ISBN, then
>> adding revision data to that?
>>
>> Without a better option, a formalized extension (ISBN - format - edition - date)
>> etc might be viable?
>
> We can *use* the ISBN, it's just not much use for what we're doing.
> I'm not sure how the extension helps - we'll still have to go
> elsewhere to aggregate works.

So for existing works (having an ISBN) such a schema would work?
What's the definition of an aggregation?
A new work or a collection of existing works?

>
> But I don't want to reinvent the ISBN or any other system. I want to
> choose a better one. Arguments *against* using the OpenLibrary ID?
> They have promised support,  have backing from the Internet Archive
> for the long term, and are committed to the same principles of
> openness.

For new works, use the OpenLibrary ID?

Retains links to existing works, lets new ones flourish?

Katong

unread,
Dec 16, 2010, 10:41:34 AM12/16/10
to Open Bookmarks
James, thanks for bringing us along. Let me try and give an argument
*against* enforcing an OpenLibrary ID in the standard.

if we really want openness and simplicity, can we really enforce a
single work ID? Even one as open and friendly as an OpenLibrary Work
ID?

Just allow the ereader software to create an annotation file that
refers the 1) work and 2) pointer/highlight (or fragment) to the best
of its ability - ISBN, handle, Google Books URI, Goodreads number,
whatever, for the work ID, and then unique block ID, text string, etc,
all of the above for the pointer/fragment identifier.

What ever the eReader can access directly when the reader makes her
annotation.

(We have to recognize that reading systems may or may not be internet-
connected at the time of annotation. They can only grab what they have
access to and what is coded in the ePub file, html5 data cache, URL,
etc.)

Then *later on* different online services can try and map and
reconcile the work/position information that's carried in the bookmark
xml notation. OpenLibrary could take the ISBN and returning the work
id. (I think ISBN to WORK ID is always many to one.) Different
services would have an incentive to help the mapping of the bookmark
xml entry to different systems, including handle/DOI for academic
works that have a DOI. Including sorting out the necessarily imperfect
position data.

At some point, in someone's service, the bookmark entry becomes good
RDF, incorporating a URi for the target, a URI for the "note" or body
of the annotation ( be it tweet or blogpost or text file) and a URI
for the annotation itself (ie the record of the relationship between
note and noted).

But we needn't demand that of the initial bookmark xml entry.

Peter

Fredrik Bridell

unread,
Jan 25, 2011, 7:06:56 AM1/25/11
to Open Bookmarks
Hi, do you mind if I bump an old thread like this?

One thing that's on my mind is how to deal with things like
collections of short stories or essays. I'd love to be keep the
bookmarks with a particular essay rather than with a particular point
inside a book where it's reprinted.

I think, generally speaking, bookmarks should go with the content, not
with the medium. ("Sounds right, but what does it MEAN?"...) Agreed?

I think maybe a good example is how anybody can just say "Exodus 3:14"
and that's a good enough reference - across editions, translations,
etc. (I just made that reference up, btw, no hidden meaning intended).

It would be great to find a solution that can solve this, and also
handle different versions (hardcover, paperback, epub, pdf... and
editions (revisions) as well as translations. It all has to do with
identifying parts of one particular text as somehow similar (equal to,
related to, derived from) another part of another text. Maybe what you
need is yet another format, something like an extra index that defines
parts in your text and matches them to ... well, either other texts,
or to "nothing", to just a set of unique identifiers. In an ideal
world, this index would be something that was already embedded in your
electronic book, but alas, it's not. (Is it?)

Any thoughts on this?

/Fredrik


stml

unread,
Jan 26, 2011, 6:19:14 AM1/26/11
to Open Bookmarks
Bookmarks are tied - inasmuch as their "tied" at all - to works, not
to editions.

Keeping bookmarks linked to works rather than editions (/reprints) or
formats is precisely what we're trying to do.

We know we can't rely on indexes embedded in books already. But we can
do very clever things with similarity, and that's what we're working
on.

On Jan 25, 12:06 pm, Fredrik Bridell <fredrik.brid...@gmail.com>
wrote:
Reply all
Reply to author
Forward
0 new messages