Interagency Data Stewardship/Identifiers/Table

'	'	Identification scheme	'	'	'	'	'	'	'	'
Axes	Criterion	DOI	LSID	OID	PURL	ARK	UUID	XRI	Handles	URI/URN/URL
Technical Value	1. Is it recommended by any standards bodies? If so, which ones?	ANSI/NISO; under consideration by ISO	I3C and OMG	Basic ISO and ITU stds	OCLC	No	ISO/IEC and IETF		IETF; US military	IETF
	2. Are there any security issues with it?	no more than any other web address	no more than any other URN scheme	not as part of schema itself			First version yes, fixed since then		No	those of any web address
	3. How scalable is it to very large numbers of objects?	Very	Very	Very	Very, but limited by capabilities of OCLC resolver	Very	2^128 values		Very	very
	4. Is it interoperable with other identification schemes?	are registered URI\'s and is handle based; local ID\'s can be from some other ID scheme (e.g., ISBN number or local accession ID)	other internet approaches	Readily interoperable with DOI\'s; potentially operable with UR*, PURLs, and UUIDs depending on detailed structures of those		TBD			used by other schemes (e.g., DOIs); can be represented as a URN or URL	Other schemes use these as a basis
	5. How compatible is it with the main internet naming schemes?	very	LSID\'s are URN\'s	IP addresses are a form of OID	PURL are URLs that require redirection	very, ARKs are URL\'s and can be used as such	very, Can be used as URN\'s		very (see above)	by definition
	6. Does it require third party maintenance (e.g., a registry and a registry maintainer?)	Yes, both registry and resolution services	Not explicitly - though authorities (and therefore a registry of authorities) are required	yes, registries and registry maintainers	Yes	Yes, Name Assigning Authorities are registered	Yes, to maintain the tools used to create them		Yes, Global Handle Registry	Yes, domain names for URL\'s; IANA registers URN schemes
	7. How dependent is it on a naming authority and how do you know that naming authority will be in business decades or more into the future?	Very but well supported by the publishing industry	central registry required?	A matter of choice	OCLC maintains the resolver	California Digital Library is the naming authority; has multi-institute support; and about 60 employees	No naming authority required		Corporation for National Research Initiatives manages the Global Handle Registry; has a fee based business model	?
	8. What is the expected longevity of the underlying technologies? How long will the scheme remain viable?	as long as the publishing community supports them (are used by lots of journals)	widely used in the life sciences; so likely to be long-lived	ensure scheme should be as long-lived as the federal government and equivalent international bodies	are currently more than a decade old and the WorldCat system is based on it (I.e., a long time)	as long as http is viable and users use this scheme	Simple, easy to implement technology - so should last some time		as long as http is viable and users use it
User Value	1. Will publishers allow it in a citation?	Definitely!	Yes, though extent of journals accepting this not established	not currently accepted by publishers	cite as a URL	cite as a URL	not currently accepted by publishers		not explicitly accepted except as a DOI	URL\'s are citeable
	2. Does the identification scheme have any additional trust value?	no	no	may depend on how the heirarchy is defined	no	Yes, ARKs explicitly reference a maintenance policy	no		no	URN more than URL - but still not really
	3. Does the identifier have meaning? (Should identifiers be transparent or opaque?)	Could but not necessary	Up to the user	Up to the user	Up to the user	Supports but discourages an opaque component of the identifier	Opaque ID		Could but not necessary	up to the user
Archive Value	1. How maintainable is the identification scheme when data migrates from one archive to another (or even from one location in an archive to another)?	Theoretically easy; but requires commitment by archives; could be confusing if the ID\'s are not opaque (e.g., if data with an ORNLDAAC ID were moved somewhere else)	As maintainable as any other URN scheme; requires commitment by archives	requires commitment by archives	very - in fact that\'s why they were invented; but they assume the name of the item doesn\'t change; requires commitment by archives	Easy to migrate all the data from one archive to another; moving just some is more difficult; requires commitment by archives	Unique across all domains, so will not change as object moves		requires commitment by archives	URL\'s not very (redirection); URN\'s more so but requires commitment by archives
	2. Are there any actual charges or costs involved (e.g., DOI’s may cost only a few cents each; but, even a small archive may have dozen’s of millions of files that should have identifiers)?	Yes, though charges per name are minimal	no	Depends on choice of registration authority and related technologies	no	No	No		CRNI charges to assign a prefix to an organization and annual maintenance thereafter	yearly fees for domain names…?
	3. Does the identification scheme handle data that is not on the web? What about physical objects?	no, no - digital resources only	web-centric	Probably no conceptual difficulties	no - only envisioned for on-line resources	No, No - on-line resources only	Yes, yes (but nobody seems to do this?)		no, no - digital resources only	no, no
Potential criteria	Existing usage within ESDC\'s	ORNL							CIESIN