Print

Print


Yes, I would love for there to be a global solution. 

We are not squatting on those domains (although there are some we ARE squatting on).

From: Karen Coyle <[log in to unmask]>
Reply-To: <[log in to unmask]>
Date: Saturday, August 3, 2013 10:35 AM
To: <[log in to unmask]>
Subject: Re: [BIBFRAME] Modeling Question

Laura, the problem with a Bowker-specific URI is that with an ISBN in hand you would have to know that it's a US ISBN in order to turn it into a URI. It needs to be a pattern that can take any ISBN. Bowker has the isbn.org domain, while Italy uses isbn.it, France uses isbnfrance.com, etc. It really has to be something global.

(I just did a quick search and isbn.com and isbn.info are registered but available for $$.)

kc

On 8/3/13 7:26 AM, LAURA DAWSON wrote:
[log in to unmask]" type="cite">
We've talked about this at Bowker. Bookwire, our book-listing site, has URL aliases such as: http://www.bookwire.com/978xxxxxxxxxx. But for SEO purposes, we have to include the title and/or author in the URL.

From: Karen Coyle <[log in to unmask]>
Reply-To: <[log in to unmask]>
Date: Saturday, August 3, 2013 10:21 AM
To: <[log in to unmask]>
Subject: Re: [BIBFRAME] Modeling Question


On 8/2/13 4:55 PM, Wallis,Richard wrote:
[log in to unmask]" type="cite">

Looks to me more like an attempt to produce a PURL service to point at publisher web page URLs.  They are only created upon request, and they could point at anything.  - even less like reliable resource identifier than the ISBN!

Richard, that is the nature of DOI, not just DOI for ISBNs. It can point to anything, and at times that anything is a publisher's home page. Still, DOIs are considered important identifiers.

I think this leaves us with a dilemma, which is:
1. Unless a publisher includes the ISBN-A on the package or in metadata, you have no idea if there is one, but
2. Unless the ISBN international agency develops a canonical URI pattern, we're stuck with the URN form - urn:isbn:....

Maybe we should be pounding on the doors of http://www.isbn-international.org/ to get a simple, global ISBN URI.

kc


[log in to unmask]" type="cite">

Once you know the pattern, yes you could calculate what the ISBN-A URL would be (if there is one created for the book with that ISBN) , but that is not a translation to a linked data URI.

In the same way, knowing the oclcnum of a bib record lets you calculate what the WorldCat Linked Data URI for the resource that record describes, because of the URI pattern used in WorldCat data.  However, they are still separate and different things.

Treating it otherwise would be the equivalent of using the text string which is the title of a book to calculate what the dbpedia URI would be for the work with that title and then saying that the URI was equivalent to the title.

~Richard


From: Karen Coyle <[log in to unmask]>
Reply-To: Karen Coyle <[log in to unmask]>
Date: Friday, 2 August 2013 21:42
To: "[log in to unmask]" <[log in to unmask]>
Subject: Re: [BIBFRAME] Modeling Question

Richard, did you see my note about DOI and the ISBN? They claim to be the official ISBN as URI:

http://www.doi.org/factsheets/ISBN-A.html

I believe that there is no problem translating an ISBN string into the DOI URI.

kc

On 8/2/13 1:23 PM, Wallis,Richard wrote:
[log in to unmask]" type="cite">
On 2 Aug 2013, at 20:18, J. McRee Elrod <[log in to unmask]> wrote:

Richard Wallis posted:

[In the linked data world] there is a significant difference between
the >numbers (OCLC number, LCCN, ISBN, etc.) associated with a
resource and the >URI that identifies it.

Of these numbers, only ISBN is associated with the resource.  The OCN,
LCCN, and other national bibliographic agency nubers, are associated
with the description.


That is true, the numbers have been associated with the records (descriptions). However the URI is a [linked data] identifier for the resource. 

Note the '[In the linked data world]' in the text above you referenced.  Linked data uses http URIs as identifiers for resources, so that they can be linked and those links followed.  Obviously there is need to record numbers and other identifying strings (which are not http URIs) that have been used to identify the resource in other domains, as properties in the RDF description.

ISBN is a bit of a special case, it is an identifier for the resource, it is a string, it is not a http URI that can be used as a linked data identifier.  So in RDF it is captured as a string property. 

~Richard.

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet

-- 
Karen Coyle
[log in to unmask] http://kcoyle.net
ph: 1-510-540-7596
m: 1-510-435-8234
skype: kcoylenet