> That means that we'll need identifiers for works, expressions,
> manifestations, etc. and I know this is a huge problem. Ideas welcome.
Ideally, as you're basing much of RDA around RDF, I would expect to
see URIs (that you can actually request) as the identifiers. But then
you still have to decide what to use to make the URI of course.
I agree with Karen that ISBNs are unsuitable, not only for the reasons
given, but also because the manifestation is only part of the problem.
You need identifiers,as Karen says, for the FRBR entities as well as
anything else you decide to make a first-class citizen of the data
(such as authors).
The reason I think ISBN appeals to many people is that it is well-
known and widely available. Meaning that existing data can be
transformed and analysed without having to work out what identifier
you should have for the thing. That matching can be complex in many
cases and requiring access to some central service that assigns
identifiers is undesirable IMHO.
There is another technique that's possible though, natural keys.
Deriving an identifier from data elements that are applicable and
widely available. This has many of the upsides of the ISBN without the
downsides - it has different downsides ;-)
I published a paper on this earlier in the year at the Linked Open
Data Workshop in Beijing. http://events.linkeddata.org/ldow2008/#program
rob
On 22 Nov 2008, at 23:29, Karen Coyle wrote:
> On Tue, Nov 18, 2008 at 4:47 AM, Alistair Miles
> <[log in to unmask]> wrote:
>
>>
>> URIs denote. I.e. URIs are names for things. If I said
>>
>> <urn:isbn:0723888560> rda:identifierForTheManifestation <urn:isbn:0723888560
>> > .
>>
>> that would be equivalent to saying the book is its own identifier.
>>
>> I could, however, say
>>
>> <urn:isbn:0723888560> rda:identifierForTheManifestation "urn:isbn:0723888560
>> "^^xsd:anyURI .
>>
>> which would be a perfectly reasonable alternative to
>>
>> <urn:isbn:0723888560> rda:identifierForTheManifestation "ISBN:
>> 0723888560" .
>>
>
> Alistair,
>
> Unfortunately I don't think we can use the ISBN as the identifier for
> the manifestation. It has its own meaning, and won't be one-to-one
> with manifestations that the library holds. For example, books before
> 1968 didn't get ISBNs; some books today don't get them (government
> publications as one example); and publishers goof and give the same
> ISBN to more than one work. So we need to keep the ISBN as the number
> assigned by the publisher (when it does so).
>
> That means that we'll need identifiers for works, expressions,
> manifestations, etc. and I know this is a huge problem. Ideas welcome.
>
> kc
>
> p.s. This is a very slow convertation, but my mind just doesn't make
> it into the fast lane much these days. I'll try to take a close look
> at all of your scenarios over the next few days.
>
> --
> -- ---
> Karen Coyle / Digital Library Consultant
> [log in to unmask] http://www.kcoyle.net
> ph.: 510-540-7596 skype: kcoylenet
> mo.: 510-435-8234
> ------------------------------------
Rob Styles
tel: +44 (0)870 400 5000
fax: +44 (0)870 400 5001
mobile: +44 (0)7971 475 257
msn: [log in to unmask]
irc: irc.freenode.net/mmmmmrob,isnick
web: http://www.talis.com/
blog: http://www.dynamicorange.com/blog/
blog: http://blogs.talis.com/panlibus/
blog: http://blogs.talis.com/nodalities/
blog: http://blogs.talis.com/n2/
|