> >
> > To cite from ANSI/NISO Z39-84: "The syntax for URIs is much more
> > restrictive than the syntax for the DOI." in general you will have
> > lot's of fun to encode a DOI so that it can be send to the DOI
> > resolver. In case you're a bit lazy with that and push naked DOI's
> > into some XML and turn on datatype=URI checking. Grrrr.....DOI is a
> > non-URI encoding scheme. Maskerading a DOI as URL makes it a URL -
> > it's no longer a DOI. From it's functionality DOI is a URN Namespace
> > - but not syntactically. DOI is not a currently recommended encoding
> > scheme for dc:identifier or dc:relation.
>
> So, my guidelines should EITHER recommend only encoding DOIs as URLs using
> the http://dx.doi.org/10.1000/182 form OR recommend escaping the DOI in
> appropriate ways before encoding it as a URI using the 'doi' URI scheme.
>
> After looking at Z39.84 I would suggest changing my wording to
>
> --- cut ---
> The Digital Object Identifier [DOI] is a persistent identifier of
> intellectual property entities. DOIs are most commonly encoded as URLs
> using the http://dx.doi.org/ prefix.
> For example:
>
> http://dx.doi.org/10.1000/182
No! This is not what i think: The way DOI people seem to prefer to cite DOI's
is "doi:10.1000/182"
DOI is an application of the "handle" system and as such UTF-8 is the authoritive
encoding. (I think doi's are those "handle"s which begin with 10. - but there are certainly more
knowledgeable people)
The URL you cite is a composite of the actual doi - value with the name of a DOI resolver.
You may use http://hdl.handle.net/ as alternative DOI resolver address.
Encoding: Character escaping might be necessary for DOI's depending on the specific
environment. (XML/SGML is rather curious about < and >) .
Taken these subleties the consequence is, that DOI's form a subClass of "handles" not of URI -
As they stand!
The simplest solution is along the lines ISBN, ISSN have choosen
urn:doi:10.1234/whatever, where whatever is the (weakest functional) URN escape of the actual doi.
The best what we can do as DCMI is to ask the owners of the DOI system, whether they consider to
register for a urn Namespace.
>
> Or are you suggesting some other recommendation?
see above
rs
>
> [Z39-84] http://www.niso.org/standards/standard_detail.cfm?std_id=480
>
> Andy
> --
> Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
> http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933
> Resource Discovery Network http://www.rdn.ac.uk/
>
>
>
>
|