Re: Persistent Documents and Locations

>Patrik Falstrom's concern about reuse of URNs is, in my estimation,
>overblown.  It may well be that ISBNs are imperfect or are imperfectly
>managed.  Real Life.  Deal with it.  We should be so lucky as to end up
>with a system that serves us as well as ISBNs serve the book world.

>Persistence of names will not be assured by technology, anymore than
>the design of ISBNs assures their perfect application.

Stu's point is well-taken.  Persistence will vary according to methodology 
applied to deriving and maintaining persistent, unique names for 
system identifiers.  ISBN works for the document domain (publishing)
for which it was designed.  But it is specific to that domain and as such,
might not be the best candidate for a URN.  OTH, a URN should
not prevent a community of interest from using established identifiers.
The certain result of that would be rejection of the URN solution.
 
Is it not likely that several naming schemes will be in effect which a local 
service must map?  So while the ISBN approach is viable on some 
systems and types of documents, message-ids will be used on
others.  Shouldn't the URN be independent of these but mappable.
That is very likely what will be done in the SGML applications.  Other
notation processors can choose their own techniques including direct
application of the URN or URL, but we can't make that decision for them as
it affects human interaction and system maintenance.

Someone correct me here, please:  is the only goal for the URN that it be a
unique, persistent name for the global system (i,e, WWW documents)?


Len Bullard

Received on Tuesday, 29 August 1995 11:29:14 UTC