Re: [Discuss-DOI] Killing 2 birds

[Eric Hellman <eric@hellman.net>]

Maybe it wasn't such a good idea.

My suggestion was that the prefix should reflect the primary type declared
in the DOI metadata. Since each DOI should be declared with only one
primary type, there should only be one prefix which is valid with each DOI.

You could of course associate wrong prefixes with a semantic such as "find
DOI's of type x associated with the doi that follows" but it's not clear to
me that that would be workable.

And no, I wasn't suggesting that uri: is better than urn: I'm not up on urn's.

Eric

At 10:51 AM -0500 3/24/99, Tony Hammond wrote:
>Eric:
>
>     uri:doi:w:10.1000/131
>
>Interesting suggestion - if we really do have to be up front with the Work v
>Manifestation distinction. I can see that this could be used for citation
>purposes, but in actual usage wouldn't
>
>     uri:doi:w:10.1000/131
>     uri:doi:p:10.1000/131
>     uri:doi:d:10.1000/131
>
>all get resolved by the Handle system to the same URI (typically a URL)?
>And BTW
>is there any notion that we would extend this scheme as
>
>     uri:doi:w:10.1000/131
>     uri:doi:p:10.1000/131
>     uri:doi:d:html:10.1000/131
>     uri:doi:d:pdf:10.1000/131
>     uri:doi:d:tex:10.1000/131
>     uri:doi:d:etc:etc:10.1000/131
>
>Still don't really understand if there is any difference between 'p' & 'd'
>- so
>if we need to distinguish between the abstract and the real, wouldn't we
>rather
>use 'm' for manifestation?
>
>As for print/web syntax suitability, AP have answered that by opting for
>DOIs of
>the form
>
>     uri:doi:10.1006/geno.1998.4536
>
>short enough (and legible enough) to be printed on our journal articles and to
>be reproduced in citations, especially with article by article publications
>where we may not have page numbers. And they don't require hex encoding, or
>entifying.
>
>Tony
>
>ps/
>Are you suggesting that we have to use "uri" rather than "urn" because the DOI
>is not yet registered as a URN namespace ID with IANA?
>
>
>
>
>
>From: Eric Hellman <eric@hellman.net> AT ~internet on 23/03/99 14:37 EST
>
>To:   <discuss-doi@doi.org> AT ~internet@CCMAIL
>cc:    (bcc: Tony Hammond/AP/LDN/HARCOURT)
>
>Subject:  [Discuss-DOI] Killing 2 birds
>
>
>Content-Type: text/plain; charset=US-ASCII
>Content-Transfer-Encoding: 7bit
>Content-Description: cc:Mail note part
>
>
>Here's a modest suggestion on the DOI syntax.
>
>First of all, note that the DOI syntax, as last proposed, is not compatible
>with being a URI without encoding. Note also that it is inappropriate for
>non-digital representations, (such as print) because non-printing and
>ambiguous characters are allowed. There is a need for a clearly defined
>extented syntax for representing the DOI in URI's and in print.
>
>Second, note that once metadata frameworks get decided, it will be easy and
>fast (assuming the metadata is done competently) for machines to look up
>metadata such as primary type, given the "pure" DOI string. Thus, for
>machine use, the proposed syntax addition ( "(W)" and the like) is
>redundant.
>
>Nonetheless, the primary type is useful to humans.
>
>A DOI-URI syntax prefix can address this issue without touching the "pure
>DOI", which frequently will require encoding anyway.
>
>something like:
>
>uri:doi:w:10.1000/131
>
>would be the URI-DOI, suitable for printing and use on the web
>
>while
>
>"10.1000/131"
>HEX(31302E3030302F313331)
>Binary(001100010011000000101110001100000011000000110000001011110011000100110011
>0
>0110001)
>
>are representations of the "pure" doi, suitable for machine use.
>

Eric Hellman
Openly Informatics, Inc.
http://www.openly.com/           Tools for 21st Century Scholarly Publishing

------------------------------------------------------
Discuss-DOI maillist  -  Discuss-DOI@doi.org
http://www.doi.org/mailman/listinfo/discuss-doi