Re: [urn] I-D Action: draft-ietf-urnbis-rfc3406bis-urn-ns-reg-01.txt

[Juha Hakala <[e-mail]>]

Hello Martin,

I agree with you in many respects, but there are some differences of 
opinion as well. See below...

Martin J. Dürst wrote:
> Hello Juha,

> This means that there is a strong question as to how, and to what 
> degree, the classical library-based name-location distinction is still 
> relevant and useful. With that I don't want to say that it's useless, I 
> just want to say that however many years of "practical experience" you 
> can claim, that experience may or may not transfer to a world with 
> essentially different "physical" rules.

 From a library point of view, for instance the following features of 
networked resources seem to be in favour of separating identification 
from location:

1. There may be multiple copies of a single resource in different 
locations. However a resource should have one and only one identifier.

2. Over time, there may be different resources available from a single 
location.

3. A single resource may be moved from one location to the next.

4. A digital resource may disappear (leaving only e.g. a printed 
surrogate).

5. A single file can incorporate multiple resources that should be 
identified separately; a single resource can span over multiple files.

There has also been discussions about persistence (or the lack of it) of 
domain names, and the need of identifying abstract entities such as 
works (which may or may not have manifestations in the Web).

> Also, at least as far as content is digital text, search engines add 
> another dimension to this picture. If things move, search engines just 
> catch up. Especially for specialized content, we are definitely not yet 
> there yet, but for more general content, the fact that something gets a 
> different URI isn't the end of the world these days.

Search engines do catch up if the resource is not in a non-accessible 
silo. There seems to be an alarming trend that instead of making 
resources available in the (semantic) Web, some importand players are 
preferring to hide their data from the open Web.

For resources in silos, it remains to be seen if persistent identifiers 
will be able to provide access.

Moreover, any library has a significant amount of non-textual resources 
which for the time being are not well catered for by search engines. 
There are interesting (pilot) services for searching music or still 
images, but this is not yet mature technology.

> So up to the middle ages, we mostly had the situation that one would 
> hear by chance that a certain library had a certain work, and would 
> travel to that library and read the work inside that library (though 
> indeed not on the shelf) if one had that much leisure time.

It is a little bit ironic that for the Web content, most national 
libraries  that harvest the Web must rely on similar practice. According 
to the existing legal deposit & copyright acts the harvested documents 
cannot be made freely available; the users must come to the national 
library or other legal deposit libraries in the country, and use the 
material on dedicated workstations, which are a modern equivalent of the 
chain that attached the mediaeval books to the shelf so as to protect 
them against thieves.
> 
> In the "library age", one would go to the nearest public library, ask 
> e.g. "I want to read Tom Sawyer", got back a question whether one meant 
> Tom Sawyer by Mark Twain, and then got a "name" (or number) to write on 
> the order slip, or got directed to the relevant bookshelf (location). 
> And one would take the book home, and the library would record the 
> location of the copy in question as "with lender Foo".

As an aside, while there is a well functioning business model for 
scientific periodicals (which, once purchased can be freely used) there 
is as of yet not a fixed model for e-books. There could be a limited 
number of electronic items (allowing for instance two users to read an 
electronic item at the same time), or each lending operation may have a 
fixed cost, so that one really popular book might take the entire 
acquisitions budget in a short time.
> 
> In the digital age, we can type something into a search box, get the 
> content in no time, and never have to "give it back". At the same time, 
> several copies will have been created. Lots of 
> names/locations/identifiers (from metadata down to track numbers on a 
> hard disk) may be involved, but the overall thing may not easily fit an 
> old-style library name/location dichotomy anymore.

The process you describe above applies freely available materials and 
e.g. commercial scientific periodicals. But for other kind of resources 
such as e-books things may become more complex, legally and otherwise. 
As regards the content, an e-book in EPUB 3 format becomes a ZIP 
container that can and will hold many things (each one of which must be 
identified separately) in addition to the text of the book, and even the 
text should adapt itself to different viewers. For the libraries' 
traditional cataloguing principles these resources will pose interesting 
challenges.
> 
> 
> Regards,    Martin.
> 

-- 

  Juha Hakala
  Senior advisor, standardisation and IT

  The National Library of Finland
  P.O.Box 15 (Unioninkatu 36, room 503), FIN-00014 Helsinki University
  Email [e-mail], tel [phone]