David Stern
		Director of Science Libraries and Information Services
		Kline Science Library
		219 Prospect Street
		P.O. Box 208111
		New Haven, CT  06520-8111

		phone:  203/ 432-3447
		fax:  203/ 432-3441

		email:  dstern@minerva.cis.yale.edu

  ************************************************************


                                  ASIS Midyear Meeting
                  Electronic Publishing: Applications and Implications
                                           Minneapolis
                                         May 25, 1995 

   ***********************************************************

Price Models for Document Delivery

Pricing in relation to USER DEMAND 
	based largely upon ideas at the Los Alamos Eprint Workshop

The models proposed by Karen Hunter discuss the logistics and 
economics of the present paper-based distribution  paradigm in 
which revenue comes from property rights.

I would like to emphasize 3 points:

1. most of this simple distribution will be done much cheaper soon 
(Ginsparg model, later) and will result in no significant revenue 
(as long as authors don't demand revenue).

2. charging for data packets creates some major problems; examples are
	pay by the number of comments linked to item (changes over time),
	how do you pay for redesined/manipulated and reposted data,
	how do you charge for part of a document (graph),
	complex departmental charge back mechanisms for convenience factor.

3. users want enhanced service/access
	still want VALIDATION
	now want FILTERS (less/better data - not more data)
		better search/browse options, 
		manipulation of data with client software,
		integrated media,
		links between documents,
		commenting capabilities,
		data integrity (distinct from ownership concerns).

These three points show that future revenue will come from 
   ADDED VALUE, not property rights to primary material.

Delivery mechanisms (be they for primary data or filters/interfaces) 
have the same basic logistics and concerns:

options:	

     A. stand alone   
	archiving  (duplication of effort at many universities)
            customization (links to local holdings, etc)

     B. consortia    
	economies of scale, leverage for R&D demands
	standard interfaces without as much local customization

The final solution will be a mix of many options:  
	niche solutions in relation to local population needs
	multiple mechanisms - local subject clusters, regional 
	     consortia, local systems (even CD-ROMs)

therefore, multiple cost options:
	subscriptions for access options (e.g. reserve systems, packages) - not data
	Pay Per View (transactional) for access to packages of data - not data elements 			(within consortia for discounts)

implementation concerns:
	copyright (Fair Use = unburdened educational exemption) plus cost levels for
		commercial use  ,
	data integrity,
	reciprocal (walk-in) arrangements among academic organizations
		(denied by password control, but IP address problems for sabbaticals),
	platform independence (browser/viewer software must work for both library kiosk 			mode and office/lab configurations),
	hardware and software support,
	predictable budgets,
	long term access to previously published information/services,
	user behavior studies -> revised and refined models.

Briefly comment on the proposed models:
	Local storage - support costs are high, but customization is possible.

	Transactional - unknown annual budgets and unbearable charge-back mechanisms
		(too much overhead for tracking, and won't scale-up as modeled),
		very difficult to customize.

	Lease options -  why multiple subscriptions or higher costs for simultaneous users?
		Never done with paper, and no relation to any serious increased costs to the 		system.

	Consortia and Host models -  ADDED VALUE and archiving must be provided.


Future models:

1.  ISI Electronic Library Project
	The eventual revenue source will be the added value of the transparent link between 	the filter and the delivery mechanism.


2.  "Virtual Journal" Model


electronic manuscript creation
                        |
                        |
                       \/

PrePrint Distribution  (FREE)  

- remote files with pointers  (local servers at universities and labs)
- centrally loaded files (Ginsparg http://xxx.lanl.gov/)
                                    (American Mathematical Society)


	|
	|       NOMINATION
	|           with rationale
	|
       \/

OR

	|
	|       ACCESS COUNTS
	|
	|
       \/


Editorial Review Board

- society approved =  VALIDATION and FILTERING
- standard format (SGML)
- added value editorial data  (image descriptions, concept hierarchies)

	|
	|
	|
	|
	|
       \/

VIRTUAL JOURNALS
Enhanced Abstracts
Review Articles (annual updates, comprehensive)

- integrated media  (simulations, animation, movies, equations)
- hyperlinks betwen related documents, parts of documents
- enhanced search capabilities


 *****************************************************************

EXPLANATION OF ABOVE:

PrePrint  -->  ePrint Distribution stage will produce a clearinghouse and index for 			local and remote documents.  Archiving and integrity are easier for centrally 		handled material (AMS).  Archiving of this non-peer reviewed material is 			the concern of the discipline-specific society/community.  The High Energy 			Physics community is served by the Ginsparg project, which has recently 			received a 1 million dollar NSF grant.  This advanced technology 				community has eagerly adopted the free distribution system, and new 			enhancements are being planned.  

Nomination and Retrieval Statistics will bring specific documents to the attention of 		an Editorial Review Board ... which will provide the user desired   				VALIDATION (tenure concerns) and FILTERING options.  The added 			content from the editorial process will also be a marketable item.

	The selective Virtual Journals that result will provide what users are demanding:
		integrated, dynamic, transparent, high quality material.  
		A variety of current highlights, review articles and SDI options will allow 			users to once again browse their specific and related fields as they strongly
		emphasized they would greatly enjoy.

Publisher niche:

	There are a number of places for the publishers to enter into this model.

  They can provide the R&D funds to collaborate in the development of 
	a) the revenue-generating text and image processing methods,
	b) the supporting database structures and storage mechanisms, and
	c) the interfaces and gateways for seamless user access to information
		(intelligent thesauri creation and concept hierarchies).