To better understand the variety of access options available for the SUS Digital Library collections, the Working Group investigated several consortia and their database access methods. Consortia investigated include: Virtual Library of Virginia (VIVA), Illinois Digital Academic Library (IDAL), California Digital Library (CDL), Georgia Library Learning Online (Galileo), Kentucky Commonwealth Virtual Library (KCVL), OhioLink, and the University of Toronto (which serves multiple Ontario colleges and universities). The working group asked representatives of these consortia a list of questions designed to provide information about the background of the organization, their preference for remote or local mounting of resources, the systems architecture used by the consortia, the possibility of internal linking between records of differing types, interface issues, access and security methods used, and their own evaluation of the services provided. The initial reports on these consortia were used to foster discussion amongst the members of the Working Group, and to develop models for the report as outlined below. The submitted reports of the Working Group members are included as Appendix B.

The basic architectural distinction is whether resources are stored and accessed through a service run by a publisher or vendor ("remote access") or whether they are stored and accessed through a service run by the consortium ("local loading"). We found no pure case of either exclusively remote or exclusively local access, and no two consortia had identical policies towards this. Rather, the consortia occupied positions along a continuum from strongly preferring remote access to strongly preferring local loading. We grouped these into three models, described below.

Model 1: Remote access strongly preferred; no services beyond consortial purchase.

Description: In this case, the main function of the consortium is to get consortial pricing and handle licensing for the member institutions. The consortium may provide a gateway to licensed resources, or may encourage member libraries to give access from their own websites.

Sites: VIVA, IDAL, CDL

Technical architecture: Not applicable; projects in this category provide little infrastructure. They may have a website with links to licensed resources.

Possible Advantages:

• Requires the least amount of central effort (programming, maintenance, etc.)
• Allows participants to take advantage of high-end, possibly database-specific, features offered by vendors.
• Demands of a large customer base may speed enhanced development where desired.
• Data is typically more current on vendors’ own sites.
• Given rapid evolution of these services, it is more efficient to have the vendor responsible for format changes needed over time (e.g. PDF to SGML or XML).
• Resources licensed separately by individual members of the consortium may integrate better with consortium-wide resources in terms of vendor-provided links if resources are not in a "closed" system.

• Cost to access content directly from vendor may be higher than if loaded locally (although local costs of staff and storage, etc. must also be considered)
• May require expensive leased lines to vendor site to guarantee acceptable response time.
• Dependent on vendor for maintenance and delivery of information (vendor may have server problems, not enough ports, Internet delays, etc.)
• Perpetual access to content less certain (archiving issues).
• Access to older content may change over time (some may not remain online; additional access or storage fees may be charged for retrieval of older content)
• Cancellation may mean loss of all content previously paid for.
• More difficult to expedite interlibrary loan services by capturing record and patron information from the local LMS.

Description: In addition to licensing resources, the consortium attempts to provide a unified interface to citation resources via Z39.50, a real or virtual union catalog, and "hooks to holdings" between citations and the local OPACs. Z39.50 is used to access both remote and local resources; if Z access is not available from a remote service, this is added incentive to load the data locally.

Sites: Galileo, KCVL

Technical architecture: Projects in this second category had at a minimum citation servers and Z39.50 gateway software. Both of them are using OCLC’s SiteSearch and WebZ.

Possible Advantages:

• Benefits of common interface to multiple resources (easier for novice/occasional user, less training required, etc.).
• May be able to set up broadcast (simultaneous) search of multiple resources and combine search results.

• May sacrifice valuable search engine, interface, and special features functionality.
• May require significant local maintenance (e.g. SiteSearch setup and maintenance over time is reported to be quite labor-intensive).

Description: The consortium attempts to build a comprehensive local environment; remote vendor services are used only if the data cannot be acquired for local loading. In addition to a common interface and hooks to holdings, the consortium provides value-added services for local data (such as links between citation databases and ejournal collections) and data archiving.

Sites: OhioLink, Toronto

Technical architecture: These implementations have both citation servers and content servers for ejournal collections. Both OhioLink and Toronto use Elsevier’s ScienceServer for journal content; Toronto also runs Ovid’s server for full text medical journals. For citations, OhioLink uses DataWare (a BRS-based search engine) for citations, while Toronto uses SilverPlatter and Ovid.

Toronto plans an interesting new architecture for the future: they will combine both citation files and ejournal collections into a single XML-based database and use web search engine technology for indexing and retrieval.

Possible Advantages:

• Benefits of common interface to multiple resources (easier for novice/occasional user, less training required, etc.) May possibly be the same interface as the library catalog.

• Currently, can build a more complete linking environment: in addition to hooks to holdings, the consortium is able to provide links between citations and the articles they cite because it has control of both citation and ejournal databases.

• The consortium can guarantee archival access because it has physical possession of the data .
• Local control of search engine, interface, and special features development (including timing of introduction of changes).
• Can probably set up broadcast (simultaneous) search of multiple resources within the local collection and combine search results.
• Can generate customized usage analysis through statistical reports.

• Considerable amount of time and money may be spent on equipment and staff needed to program, maintain, and update systems.
• Interface and search engine may not include some desirable features available from some vendors directly (e.g. thesaurus searching, enhanced reference linking, etc.).
• Data likely somewhat less current than if connecting to vendor site directly.
• Depending on vendor production methods, gaps in content may exist (e.g. missing journal issues) that do not if connecting to vendor site directly.
• Links to content outside of local collection may not be possible.
• Links from outside services licensed separately by consortium members to local content may not be possible.

The Working Group discussed the relative importance of some of the cited advantages of these models.

Common interface: A common interface to multiple resources can be provided by loading resources into a single system (model 3), or by providing Z39.50 access to multiple local and/or remote systems (model 2). Although the advantages of a common interface in providing a consistent look and feel are not negligible, they may be less important than in the past, as web-accessible services tend to be easier to learn and use than older telnet systems, and most share a basic level of commonality through the browser interface (point and click, pull-down menus, etc.). Another benefit is that the common interface may allow concurrent search of multiple resources. A disadvantage is that it may prevent users from taking advantage of vendor-specific interface features. A possible approach would be to offer both features as options: one option to enter a broadcast search through a common Z39.59 interface, and another option to choose the typically more powerful, vendor-specific interface.

Authentication Options and Affect on Access: As longer term SUS plans are developed, issues involved with access by users connecting from known IP's and remotely by dial-up need to be studied to insure access to as much system functionality as possible. For example, use of a local webscript to a remote service may interfere with some desirable linking. On the other hand, too much dependence on a proxy server may slow response time.

Interface functionality: Since commercial online services are competitive, vendors can be expected to develop new and differentiating features which usually are available first (or exclusively) via the vendor’s own remote interface (model 1). For example, Elsevier’s Science Direct (the remote service) provides reference linking between citations in journal articles and the articles cited, which is not yet available in Elsevier’s Science Server (the onsite product). In general, these features were felt to be important for the user community.

Hook to holdings: "Hooks to holdings" are links between a citation and a display of library holdings of the cited object. These are usually a citation to a journal article and a display of holdings of the journal title in print, but the holdings may also show individual (non-consortial) institutional subscriptions to electronic resources. Hooks to holdings are theoretically possible in all models but rarely implemented in model 1. The feature is important, but it was noted that the trend in consortium licensing to allow all members electronic access to all titles subscribed to in print by any member tends to diminish over time the importance of local holdings, and hence of hooks to holdings.

Citation linking: Linking between citation files (A&I databases) and the electronic full text of the articles cited is currently possible within the closed systems offered by some consortiums (model 3) or cooperating vendors (model 1). For example, ISI’s Web of Science can link to Springer-Verlag and Elsevier journal collections. Linking from the references section of journal articles to the articles cited in the references is currently possible inside some vendors’ systems (e.g. Elsevier and Ovid). Citation linking of both types was felt to be very important for the research community. It was noted that several efforts are working on models for open citation linking so the advantage to closed systems may diminish over time.

No model was found to be so compelling as to exclude consideration of the others. Also, no conclusions can be drawn without comparing relative costs, a major undertaking which the Working Group did not do. However there was some consensus that models 1 and 3 are most promising. Model 1 offers two important benefits, currency of data and vendor-specific functionality, while requiring very little overhead in central staff. Model 3 is attractive because of its potential for implementing hooks to holdings and citation linking, although these advantages may diminish over time as more open models for linking are implemented. It was also noted that the benefit of Model 3 increases with the number of resources in the system – the fewer consortially available resources, the less the power of the citation linking and other services provided. Model 2 appeared the least attractive as it required significant work to support while offering the fewest or weakest advantages.

We also recognize that in addition to resources arranged for through the SUS consortium, each institution may arrange for additional resources independently, with a subset of SUS Libraries, or through other consortia. A successful model will have to incorporate these resources with central SUS resources in a way that effectively meets the needs of our users.

The Work Group also discussed the possible disposition of the roughly 20 citation databases currently offered via WebLUIS. If the FLCA NOTIS implementation is replaced with a Unix-based LMS, the citation databases will also need to be moved off of the mainframe within the same timeframe.

There are at least three alternative solutions:

1) move the databases to the new LMS;

2) move the databases onto a Unix platform independent of the LMS;

3) switch to remote access to vendor systems.

Option (1) may or may not be possible or desirable depending on the LMS selected. Since it would be useful to have a solution independent of the LMS selection process, this is the least desirable of the options.

Option (3) is only possible for the subset of databases that are commercially available. Locally-developed databases such as the FORMIS Ant Bibliography will require a local server solution.

For commercially available databases, the choice between options (2) and (3) should be based on a rigorous comparison of both functionality and costs. Unix-based packages which could support option (2) include OCLC’s SiteSearch, DataWare (the BRS-based search engine used by OhioLink), SilverPlatter and possibly others. SiteSearch is already licensed for SUS use, and has been analyzed by FCLA and found suitable in general for citation databases.

The Working Group recommends that:

• FCLA implement a pilot project to move a citation database to SiteSearch so that real experience can be obtained.
• This or some other Working Group be continued with the charge of further examining the issues related to the various models. This should include a comprehensive study of the relative costs of local access through SiteSearch and remote vendor access for the set of databases currently offered through WebLUIS.

APPENDIX B: CASE STUDIES OF CONSORTIAL PROJECTS
Contents

1. California Digital Library
2. Galileo
3. Illinois Digital Academic Library
4. Kentucky Commonwealth Virtual Library
5. OhioLink
6. University of Toronto
7. VIVA: The Virtual Library of Virginia

  1. CALIFORNIA DIGITAL LIBRARY

Submitted to Monica Metz-Wiseman by Cecily Johns, Deputy University Librarian,

Davidson Library, University of California, Santa Barbara

-- what is the intended user population?

The intended user population is the faculty, students and staff of the University of California. Some data bases, like the Online Archive of California and the Melvyl catalog are available to the public at large.

-- does the project have a bias towards local mounting of resources, remote access to resources, or neither

There are a group of about 12 or so legacy databases that are mounted centrally on a server in Oakland, CA but the remaining databases are available remotely, e.g., OCLC databases

-- can they provide links between ejournal articles and citation databases

and if so how?

One of the conditions of license agreements with database publishers, like Web of Science or INSPEC, is that links will be established between database citations and the full text electronic journals.

-- what are the options for authentication?

The issues of authentication have not been resolved. Right now the campuses ftp their patron files to Oakland and these are used for authenticating users. It is a primitive method and does not work very well. Some campuses have set up their own authentication systems but there is no UC wide authentication.

-- how well does this architecture accommodate non-commercial resources

(e.g. locally digitized collections)?

I can't really answer the question about how well the architecture accommodates non-commercial resources. The Online Archive of California which is currently a database of finding lists must be scanned if not already in digital form or formatted at Berkeley before they are mounted and made available.

-- why did they design it the way they did?

CDL maintains 9 databases in-house and this involves substantial staff time. There are a number of databases that we access through bibliographic utilities, i.e., RLG and OCLC. The remaining CDL databases are available on the web (and of course FirstSearch

databases are now web-based). There are another group of databases available from Stanford and UCLA servers. The answer really is determined by the limit to the ability of operational staff at CDL in Oakland to maintain in-house databases, the opportunity for cooperative ventures such as access databases that are maintained on another campus server, and finally the migration of these databases to the web.

-- do they feel they are successful in meeting their goals?

I believe that CDL has been very successful in meeting its goals. The major initiative was to develop a Science, Technology and Industry collection and that is well underway. It began with a survey of faculty for top priority publishers of ejournals. At the same time we are developing a database of special collections finding aids and will be

adding full text content this year (Online Archive of California). Collection specialists in the social sciences, arts and humanities are also working on developing collections -- we added the Chadwyck-Healey products, LION and PCI, last year. Surveys are used to determine the highest priorities. The major limitation is money for content.

-- what are their future plans?

Future plans are to continue to develop in the areas described above and to add government publications and social sciences data.
 
 

2. GALILEO

Rich Bennett

Background:

--what is the intended user population?

Their vision is for a statewide library for the entire state of Georgia. Their first goal is "To ensure universal access to a core level of materials and information services for every student and faculty member in the University System of Georgia-regardless of geographic location, size of institution, or mode of instructional delivery: traditional residential, off-campus, or distance learning." Long term it is more than that.

-- who are the participants?

GALILEO has public databases which are available to anyone, and licensed databases which are available to currently enrolled students, faculty, and staff at participating GALILEO institutions based on contractual agreements with database vendors. Currently participating institutions include the University System of Georgia, the Atlanta/Macon Private Academic Libraries (AMPALs) and Georgia Private Academic Libraries (GPALs) consortia, the Department of Education and K-12 schools, public libraries, and the Department of Technical and Adult Education's technical institutes. List of participants is at http://www.peachnet.edu/galileo/commune.html#usg . They do not all have access to the licensed databases.

Local vs. remote

-- Are they storing some commercially licensed content locally, providing access to content directly from vendor servers, or both?

Some is still stored locally. When GALILEO began, the Web was new and there were great concerns about response time, so databases already stored locally were continued in this way.

-- Do they have a bias towards local mounting of resources, remote access to

resources, or neither?

Definitely moving toward the remote access model for commercially-available resources.

Architecture:

-- what is the overall technical architecture and software components

SiteSearch is used as the citation server. Did not ask what they are using for serving local commercial citation content. My understanding is that they do not have any locally-stored E-Journal content.

-- What is relationship (if any) between library catalogs in consortium institutions and digital library resources?

Users must all enter GALILEO through the front door of GALILEO. Links to individual databases within GALILEO via library catalogs are not possible.

-- Are resources that are separately arranged for by a member of the consortium accommodated so that an individual institution's resources can all be incorporated into a single place? Or is this not currently possible? SiteSearch does allow for customized GALILEO menus by institution, so this is possible.

Linking:

-- can they provide links between ejournal articles and citation databases and if so how?

No

-- do they provide direct links to journal articles and other digital resources for course reserve? If so, is this limited to materials stored locally? Or to articles on commercial vendor sites as well

No

-- Can they link citation database records to detailed local library

holdings information?

Not really. They do have a union serials list that is updated twice a year, and there is some mechanism for checking about general title holdings by institution. I wasn't clear how it works. But it does not approximate FCLA's hook-to-holdings feature

Search and user interface:

--Do they have a common user interface for all resources provided?

No, but a majority have been available using SiteSearch. However, they seem to be planning to use the new FirstSearch interface for delivering FirstSearch databases since it has improved dramatically and is superior to their current SiteSearch. NOTE: They have not fully evaluated a recent upgrade to SiteSearch, but upgrading their local SiteSearch would be a major commitment. Apparently SiteSearch is not easy to set up and maintain.

--Is user access always via a web browser? Is there alternative access available (e.g. a Windows client)?

Web only.

--Can they search multiple databases with a single search? Describe.

No

--Can they search only citation records - or full text as well?

There is nothing stored locally that includes full-text searches. Not sure if any of the remote services have full text searches.

Access and security:

-- Do they have a way of supporting access to resources that are licensed only by some consortium members? If they are stored locally? If accessible directly from vendor?

It is possible for an institution to include separately arranged databases on their SiteSearch menu. GALILEO databases are only considered to be those that are available to all.

-- can they limit the number of simultaneous users into any particular commercial database or collection?

I forgot to ask about any locally stored data. Remote access limits are controlled by the vendor.

-- do they provide local scripting as part of the gateway to commercial databases?

I didn't ask specifically. I don't think so

-- what are the options for authentication?

Access to licensed databases is available campus-wide, at satellite centers, and at off-campus locations via IP addresses registered with GALILEO. Remote access to those GALILEO licensed databases which provide it is available to currently enrolled students, faculty, and staff via password. The password is institution-specific, changes every quarter, and is e-mailed automatically to the library director or designee at each institution prior to the change.

Currently enrolled students, faculty, and staff should not supply the password to any person not meeting the qualifications for access. Passwords cannot be e-mailed to faculty or students by GALILEO; they must obtain the current password from their library or media center. Each library has a local policy regarding the verification of affiliation and the distribution of passwords.

Evaluation:

-- how well does this architecture accommodate non-commercial resources (e.g. locally digitized collections)?

I didn't ask. There are about 20 publicly-available databases which I believe are not centrally managed. So a variety of platforms are probably being used.

-- why did they design it the way they did?

When the concept for GALILEO was being discussed, SiteSearch was really the best approach available.

-- do they feel they are successful in meeting their goals?

Yes.

-- What pros and cons do they see with what they're doing? What would they

do differently?

Maintaining SiteSearch interfaces is very time-consuming. They are not sure whether there is a better approach available right now, but SiteSearch was/is a lot of work the way they have implemented it (with customized local interfaces) and has required a lot of self-education.

-- what are their future plans?

Definitely moving toward a model of remote access to commercially-supplied resources. Not sure what they will do with the recent upgrade to SiteSearch - will involve a major commitment. Will probably use new FirstSearch interface for FS databases - not SiteSearch interface. Needing to educate users to multiple interfaces is not considered as

important a consideration as was the case originally. There's no way to serve everything using one interface anyway.

-- Other Info

GALILEO Home Page -- http://www.galileo.peachnet.edu/

GALILEO Database Listing --

http://neptune3.galib.uga.edu/database_listing.html

Lots more about GALILEO -- http://www.peachnet.edu/galileo/about.html

3. Illinois Digital Academic Library (IDAL)

Carol Turner

-- General information

The Illinois Board of Higher Education has approved an $870,000 grant to launch the Illinois Digital Academic Library (IDAL). IDAL will provide access to a core collection of digital resources on-line, primarily journals, with full text and images.

The bulk of the funding will be used to acquire digital collections. The first year acquisitions program will focus primarily on acquiring digital materials appropriate to a primarily undergraduate audience. First year funding has been divided into two categories:

* $500,000 to fund the acquisition of a license(s) to full text periodical collections for seven months with the intention of continuing service in succeeding years

* $300,000 to fund the acquisition of digital resources that have a one time cost to the program. These may be general interest digital monographs or topical collections of digital texts and images.

Project Timeline

* 8/31/99 - RFP distributed

http://ilcso.aiss.uiuc.edu/Web/Services/IDAL/IDAL_RFP.pdf

* 10/15/99 - Proposals due

* 11/30/99 - Begin contract negotiations with recommended vendor

* 1/13/00 - Action on contract by University of Illinois Board of Trustees

* 1/31/00 - IDAL implementation begins

The IDAL initiative will be managed by the Illinois Library Computer Systems Organization (http://www.ilcso.uiuc.edu), with administrative support from the University of Illinois Office for Planning and Budgeting. The program will be open to colleges and universities beyond the 45 ILCSO member institutions, and these additional institutions will be involved in program governance.

Responses to the questions below are taken from the RFP document

Background:

-- what is the intended user population?

Students, faculty, and staff of participating Illinois institutions of higher education. A strong emphasis is placed on support for distance learners.

-- who are the participants?

Participating academic libraries must be affiliated with post secondary institutions accredited by the Illinois Board of Higher Education and must be members of the Illinois Library and Information Network (ILLINET).

Architecture:

-- what is the overall technical architecture and software components

"The proposer shall describe the hardware and software platform for the service, and indicate what measures will be taken to ensure that the increased workload from IDAL users does not impact overall system performance."

-- What is relationship (if any) between library catalogs in consortium institutions and digital library resources?

"The proposer shall describe the capability for indicating local library holdings information in a multi-system environment."

Linking:

-- can they provide links between ejournal articles and citation databases and if so how?

"The proposer shall describe the capability for providing linkages between the citation or abstract and the full text that is available in electronic collections to which the library has access."

-- do they provide direct links to journal articles and other digital resources for course reserve? If so, is this limited to materials stored locally? Or to articles on commercial vendor sites as well?

"The proposer shall describe the ability to create and use an online course reserve system."

-- Can they link citation database records to detailed local library holdings information?

"The proposer shall describe the capability for providing linkages to the catalog entry and library holding for the item."

Search and user interface:

--Do they have a common user interface for all resources provided?

Proposers are asked to provide information on each if multiple interfaces or search engines are supported.

--Is user access always via a web browser? Is there alternative access available (e.g. a Windows client)?

"Access to these databases shall include but shall not necessarily be limited to a web-based interface from end user locations to one or more proposer locations."

--Can they search multiple databases with a single search? Describe.

Asked to describe system's functionality in ... "ability to search multiple databases simultaneously, including automatic de-duping of records retrieved from a multi-file search."

--Can they search only citation records - or full text as well?

Asked to describe system's functionality in ... "ability to have searches retrieve full text documents only, as a system-wide default (i.e., so that any user's search retrieves only full text documents, by default, with searching for citations and abstracts available as an option)

Access and security:

-- what are the options for authentication?

"The proposer shall detail the handling of access control including but not limited to IP authorization. IDAL institutions use a variety of integrated library systems. Remote access by end-users will be a key component of IDAL. These end-users frequently will use Internet Service Providers not affiliated with the user's home institution, limiting the effectiveness of IP authorization. The proposer should propose a means to facilitate such access in a straightforward way that does not impede the process of connecting to the product(s) and/or service(s) by the end-user. The proposer may utilize a third-party software supplier to provide authentication services, provided that the costs of such services are included in the total maximum award amount of $500,000 for 7 months of FY2000 access....The proposer shall specify whether IDAL institutions may pass valid users to its service using various standard authentication mechanisms (e.g., Lightweight Directory Access Protocol (LDAP), Kerberos, Bluestem, etc.).

4. The Kentucky Commonwealth Virtual Library

Walter Rowe

http://webz.uky.edu/

Introduction

The Kentucky Commonwealth Virtual Library is a work in progress. Discussed are the various decisions leading to the construction of the final site to be presented to the public on November 1, 1999.

Background

The KCVL databases are for all Kentucky citizens and Kentucky institutions' students and faculty. Anyone who walks into a Kentucky library may use the databases while in the library and for remote users will receive ID/Password. Participation is intended to include all Kentucky libraries. Each library has input as to planning, research design issues, and construction. Representatives in multiple committees and working groups include representatives from Kentucky libraries. Included in the KCVL is a site management group that includes a Director, an Assistant Director, a Secretary, and two Virtual Reference Librarians.

Local vs. Remote:

Commercially licensed content will be stored locally if possible but even if not access to all databases will be from a single internal search engine.

Architecture:

KCVL is presently using the OCLC SiteSearch WebZ interface with its Z39.50 protocol to accommodate all searches in Kentucky libraries on-line catalogs and all databases. Kentucky digitization sites will use the DynaWeb platform integrated through the KCVL in order to accommodate the "same feel/same touch" expectations of the site. The KCVL plans to incorporate consortium members resources through the promise of cooperative statewide licensing agreements.

Linking

Many Kentucky libraries will have their own Website. KCVL will be a link placed on each library's local page. WebZ will supply the e-journal articles and citation databases a "hook to holdings" allowing for detailed local library holding information. This site is a Kentucky state- wide project; Course Reserves for the universities within the state will be handled by the individual institution. Though the capability is available, there is no indication that Course Reserves will ever be added.

Search and user interface

The reason for creating the site was to build a common user interface for Kentucky citizens, faculty and students; this interface is the KCVL site. All access will be via the Web, there is no alternative access planned via a Windows client or any other alternative access.

KCVL will use SiteSearch WebZ with the Z39.59 protocol to allow searches to encompass all available databases. Available full-text versions will be presented to the user as found via the database.

Access and Security

Access will include resources shared through the consortium with the expectation for statewide licensing agreements. All databases will be searched through SiteSearch, which in some cases can effect the specific accessibility to a site. This decision is based the "same feel/same look" philosophy of those constructing the Kentucky Commonwealth Virtual Library.

Local scripting is a gateway to the site by users connecting from their personal computers. Those using the site at participating libraries will not need an ID and password. The process of authentication is done via a connection at the site where one can request access.

Evaluation

Final evaluation of the site is difficult until the site is made available on November 1, 1999. The site had initially planned to migrate to Endeavor and its Voyager integrated information system but delays and software problems made them reevaluate.

Design and architecture of this site is based upon a broad-based user group. This factor led to the "same touch/ same feel" philosophy but a well-versed user may become frustrated.

Architecture for the non-commercial resources and locally digitized collections seem to

be well thought out though DynaWeb is not a familiar product. The Architecture for the commercial sites seem to be a bit flawed. The decision to use exclusively SiteSearch limits the patrons capability to accomplish an all-inclusive search within many databases. A link to a databases own search engine would allow for better more specific search results.

Future plans for the site includes a reevaluation of Voyager in the spring. They feel if it works as presented it could allow for more utilities such as Inter-Library Loan and more efficient reports.?

5. OhioLink

Priscilla Caplan

Background:

-- what is the intended user population?

Students, faculty, staff of Ohio colleges & universities

-- who are the participants?

17 public universities, 23 community/technical colleges, 35 private colleges and the State Library of Ohio

-- what resources are offered?

About 76 databases, including full text and citation indexes. The subset of resources which is called the "Electronic journal collection" (EJC) consists at this time of all the Academic, Elsevier, Kluwer, Springer-Verlag, Wiley and project MUSE journals. They only include in EJC publishers who allow them to load ALL of their titles.

Local vs. Remote:

-- Are they storing some commercially licensed content locally, providing access to content directly from vendor servers, or both?

Both. In addition to the library union catalog, they store about half of the citation databases and the EJC locally.

-- Do they have a bias towards local mounting of resources, remote access to resources, or neither?

Strong bias for local mounting. They only go remote if the publisher won't allow local loading.

Architecture:

-- what is the overall technical architecture and software components

There is a union catalog of the libraries of participating institutions stored in Innovative Intervaces (III). There is a server for citation databases that uses DataWare (based on BRS) as the search engine. The EJC uses Elsevier's ScienceServer software as the content server. ScienceServer also has some search capability, so a user can initiate a search for a journal article in either a citation database or in the EJC.

-- What is relationship (if any) between library catalogs in consortium institutions and digital library resources?

The participating libraries all have their own catalogs at their own institutions, and also contribute copies of their catalog records to the union catalog maintained by OhioLink. OhioLink sends to each participating library a set of catalog records for all the journal titles contained in the EJC, with appropriate URLs, which the libraries can load into their own catalogs. The Union Catalog maintained at OhioLink also has URLs in catalog records for journal titles. The URLs point to the entry page for each journal (list of issues).

-- Are resources that are separately arranged for by a member of the consortium accommodated so that an individual institution's resources can all be incorporated into a single place? Or is this not currently possible?

OhioLink is not involved with any data except consortial purchases. For the non-EJC databases, all OhioLink members can access all databases. For the EJC, an OhioLink member can decide whether they want to get it or not. There is no level of selection within EJC (i.e., "I want these titles but not those") and ScienceServer currently has no access management that would support this, although this function is an enhancement expected in the next version.

Linking:

-- can they provide links between ejournal articles and citation databases and if so how?

They provide links between citation databases in the DataWare server and the EJC in ScienceServer. The DataWare server has a manually maintained list of journals included in the ScienceServer. If a user is searching the DataWare server, and a retrieved citation is for a journal found in the list so known to be in ScienceServer, the DataWare server sends a request to the ScienceServer; if the ScienceServer finds the article, it returns a link to the full text, which is displayed to the user with the citation.

The request sent to ScienceServer consists of the ISSN, volume, issue and page of the article. Since this can be generated from the citation data, there is no need for the DataWare server to store URLs, DOIs or other article identifiers in the citations themselves. This also obviates the need to update the citation databases when issues are added to the EJC.

There are also links into the ScienceServer from ISI's Web of Science, a remote database. These links are added by a rather complicated exchange of files between ISI and OhioLink which I won't try to explain unless I have to.

-- do they provide direct links to journal articles and other digital resources for course reserve? If so, is this limited to materials stored locally? Or to articles on commercial vendor sites as well

Yes. It is up to the individual reserve room, but schools can do this with the EJC and for external resources when the vendor allows this kind of access (for example, UMI ProQuest's SiteBuilder). For EJC articles, the ScienceServer provides static URLs which can be bookmarked and copied into reserves records.

-- Can they link citation database records to detailed local library holdings information?

Yes. When citations are retrieved from the DataWare server, there is a "check availability" button that returns all locations holding the journal in print. Individual locations can be selected to return detailed holdings information.

Search and user interface:

--Do they have a common user interface for all resources provided?

There is a common interface to all citation databases in the citation server, and to all the journal collections in the EJC. The citation server and the EJC have different interfaces. External resources also have different interfaces, obviously.

--Is user access always via a web browser? Is there alternative access available (e.g. a Windows client)?

I didn't ask if there was a non-browser client. I don't think so.

--Can they search multiple databases with a single search? Describe.

I didn't ask. From playing around, it does not look like it is possible to search multiple databases in the DataWare citation server with a single search. It is possible to search all the different journal collections in the EJC with a single search, but these are actually a single physical database in ScienceServer.

--Can they search only citation records - or full text as well?

In the DataWare citation server there are (obviously) only citation records to search. In the EJC, searches are also on bibliographic data provided along with the journal articles. Most of the articles are PDF files, so there is no full text to search. If there was full text, however, the ScienceServer could search it. However, I was told the full text search wasn't very good, at least not in an earlier release of the system.

Access and security:

--Do they have a way of supporting access to resources that are licensed only by some consortium members? If they are stored locally? If accessible directly from vendor?

No.

-- can they limit the number of simultaneous users into any particular commercial database or collection?

No.

-- what are the options for authentication?

The primary form of authentication is by IP validation. If the IP address is not recognized, there is a rather complicated authentication process which requires the user to identify his home campus. Authentication was written by OhioLink and was done this way because, at the time, the participants did not want to run their own proxy servers.

Evaluation:

-- how well does this architecture accommodate non-commercial resources (e.g. locally digitized collections)?

It doesn't. Locally digitized collections, as well as commercially available collections of non-textual data (e.g. visual images from AMICO) are served outside of this system by a relatively new component of OhioLink called the "Digital Media Center" (DMC). The DMC has its own image server using Informix as the relational database. They are developing the capability of supporting various forms of metadata suitable for images such as the VRA Core Categories.

-- why did they design it the way they did?

They wanted ownership of data, links between citations and articles, and a single user interface.

-- do they feel they are successful in meeting their goals?

Yes.

-- What pros and cons do they see with what they're doing? What would they do

differently?

The pros are, they got what they wanted. They did not volunteer any major disadvantages. I inferred they would do use local proxy servers for authentication if they were designing that today.

-- what are their future plans?

For the DataWare citation server and the EJC, the future plans are just more of the same -- they want to add more data from more publishers, and take advantage of enhancements in the ScienceServer as they come out. The DMC is relatively new and is run by a different group in OhioLink, I didn't speak with them.

6. University of Toronto

Athena Hoeppner

-- who are the participants?

19 Ontario Universities, community colleges, various consortiums.

The U of T has three campuses: the main campus downtown and two Suburban campuses. One of our the guiding principles is that access to library services should be independent of location.

As the largest library in Canada we have a role (but no formal mandate) as a resource supporting other Canadian libraries. We have extended this into the area of electronic resources. When we develop/implement our systems, we always have the potential use by other institutions in mind.

Local vs. Remote:

-- Are they storing some commercially licensed content locally, providing access to content directly from vendor servers, or both?

We have all of the Elsevier journals loaded locally; the rest are accessed via the internet. We are just starting to load Academic Press on our local system and will be adding Springer, Wiley & Kluwer after that. Of the journals only about 8-9,000 are complete, full text; the rest are abstracts only. The Indexes & Abstracts include the usual things you would expect -- most are loaded locally. The rest are available via the internet.

-- Do they have a bias towards local mounting of resources, remote access to resources, or neither?

We anticipate loading locally the journals from as many of the primary publishers as we can manage. Solves archive issue. Linking can be done independently. Take advantage of intranet links.

Architecture:

-- what is the overall technical architecture and software components

Our locally mounted electronic journals are loaded on an IBM RS6000 using the ScienceServer system. We also have all of the Ovid journals available for local loading on our local Ovid system.

We operate 7 Sun servers for our SilverPlatter & Ovid databases and 5 NT servers with WinFrame to provide access to those databases which are Not available from SilverPlatter. We also provide remote access to all the databases from Cambridge Scientific Abstract and to other odd title.

One and half staff for tech support. Venders do programming.

-- What is relationship (if any) between library catalogs in consortium institutions and digital library resources?

E-journals cataloged. Concerned about updating. Future plans to integrate catalog and citation databases.

-- Are resources that are separately arranged for by a member of the consortium accommodated so that an individual institution's resources can all be incorporated into a single place? Or is this not currently possible?

Local loading at U of Toronto, IP checking.

Linking:

-- can they provide links between ejournal articles and citation databases and if so how?

Citation-to-article linking is a key strategic direction for us. We are working with all of our major primary and secondary publishers to make this happen. Our aim is to never give the user a dead-end. All citations should lead to at least one of the following: 1) an electronic article, 2) information about paper holdings & locations, 3) a document request

form (i.e., ILL or document delivery).

-- do they provide direct links to journal articles and other digital resources for course reserve? If so, is this limited to materials stored locally? Or to articles on commercial vendor sites as well

Like the idea, working towards it. Considering SiteBuilder w/ Proquest. Not limited to local.

-- Can they link citation database records to detailed local library holdings information?

No.

Search and user interface:

--Do they have a common user interface for all resources provided?

As much as possible.

Whenever we develop/implement information systems, we try to do it from a "user-centric" perspective which says "Information providers should modify their systems to match the information seeking behaviour of users". I contrast this with the "vendor-centric" which says "Information seekers should modify their behaviour to use the systems on which information resides."

We are working toward the development of a new information system (code named AltaCat) in which we will combine our library catalogue records as well as our index & abstract records into one massive data store. The records will (probably) be XML-based and we will use one of the web search engines (probably AltaVista). We anticipate building a lot of "smarts" into the front-end to make use of the rich metadata built into these records; we also plan to parse the user's query strings and deliver expert guidance with hand-crafted pages and databases – much like Yahoo and others do. We will then build in or create links from these sources to the local and remote journals we have licensed.

--Is user access always via a web browser? Is there alternative access available (e.g. a Windows client)?

Whenever possible, electronic services will rely upon web deliver; we avoid proprietary clients like the plague.
 
 

--Can they search multiple databases with a single search? Describe.

Yes. Can search across Elsevier, Kluwer, and Academic Press. Silverplatter databases as well.

--Can they search only citation records - or full text as well?

Uncertain.

Access and security:

-- Do they have a way of supporting access to resources that are licensed only by some consortium members? If they are stored locally? If accessible directly from vendor?

Terminals on campus are OK. Has a U-Tor dial ISP accounts for faculty staff

and students. Get CD to set up connection. Have 20 hrs free then charges

$5 per 20 hour. Enter id number for authentication. Also have proxy.

-- can they limit the number of simultaneous users into any particular commercial database or collection?

As necessary for remote. Not for local.

-- do they provide local scripting as part of the gateway to commercial

databases?

Yes

-- what are the options for authentication?

IP

Evaluation:

-- how well does this architecture accommodate non-commercial resources (e.g. locally digitized collections)?

Still in infancy. System will support projects.

-- do they feel they are successful in meeting their goals?

Yes - goals set high. Proud of where we are.

-- What pros and cons do they see with what they're doing? What would they do

differently?

Pros: speed of access, linking options,

-- what are their future plans?

See above.

7. VIVA: The Virtual Library of Virginia

Valerie Edwards

Background :

1. What is the intended user population?

• VIVA is a consortium of academic libraries across the state of Virginia. The primary user population is the faculty, staff, and students of these institutions.
• In 1998, the combined FTE of VIVA members was approximately 266,000.

  2. Who are the participants?

• 39 state-assisted colleges (6 doctoral, 9 4-year colleges and universities, 24 community colleges)
• 32 private colleges and universities participate as members when possible
• The Library of Virginia is an Associate member.
• Associate members must be non-profit research/educational institutions chartered in the Commonwealth of Virginia. There is an annual fee of $1000 per institution for Associate membership. Privileges of an Associate Member include an invitation to buy into contracts for electronic resources as they are negotiated, and they may participate in training workshops for those resources for which they have negotiated contracts.
Local Vs. Remote:

1. Are they storing some commercially licensed content locally, providing access to content directly from vendor servers, or both?

• VIVA has several Chadwyck-Healy full text resources loaded locally at the University of Virginia. These were loaded by UVA before VIVA was founded, and the contracts were extended to other consortial members.
• All other products are accessed directly from the vendors.
  
  

2. Do they have a bias towards local mounting of resources, remote access to resources, or neither?

• The bias is towards remote access. VIVA has a very small staff (2-3 persons), and relies mainly on volunteers from the member libraries. To locally load resources would cause VIVA to either 1) hire additional staff and purchase additional equipment, or 2) depend on individual libraries to take on the responsibility of local loading.
• VIVA's director, Kathy Perry, states that as long as Internet response time is good – and it keeps improving every year – they will continue to avoid local loads unless there is a significant price break for local mounting.

Architecture :

1. What is the overall technical architecture and software components (e.g. OhioLink uses DataWare as a citation server and ScienceServer as a content server with hotlinks from one to the other)?

• N/A
  

2. What is relationship (if any) between library catalogs in consortium institutions and digital library resources?

• No relationship. Each library maintains their own catalog, which has no relation to VIVA or database procurement.
  
  

3. Are resources that are separately arranged for by a member of the consortium accommodated so that an individual institution's resources can all be incorporated into a single place? Or is this not currently possible?

• VIVA does not incorporate individual institution's resources into their homepages. Furthermore, they request that each institution produce a webpage with links to their own electronic holdings (whether or not the sources come from VIVA) rather than refer patrons to the VIVA homepage.

Linking :

1. Can they provide links between ejournal articles and citation databases and if so how?

• They encourage individual institutions to contact vendors directly for such services.
  
  

2. Do they provide direct links to journal articles and other digital resources for course reserve? If so, is this limited to materials stored locally? Or to articles on commercial vendor sites as well?

• Each institution is responsible for their own reserves (no centralized catalog)
  
  

3. Can they link citation database records to detailed local library holdings information?

• N/A
  
  

Search & User Interface :

1. Do they have a common user interface for all resources provided?

• No.
  
  

2. Is user access always via a web browser? Is there alternative access available (e.g. a Windows client)?

• Some databases also have telnet access, but most telnet access is being replaced with web access. VIVA only subscribes to online sources for its members (never CDs).
  
  

3. Can they search multiple databases with a single search? Describe.

• No.
  
  

4. Can they search only citation records - or full text as well?

• N/A
  
  

Access & Security :

1. Do they have a way of supporting access to resources that are licensed only by some consortium members? If they are stored locally? If accessible directly from vendor?

• Most resources are bought for the state-supported schools from a state fund, with private schools participating and paying from their own funds. The web site gives access to those products bought from state funds.

2. Can they limit the number of simultaneous users into any particular commercial database or collection?

• VIVA tries to negotiate licenses with unlimited access, or to "pool" the number of simultaneous users for maximum usage. VIVA depends on the database provider to keep track of number of simultaneous users.
  
  

3. Do they provide local scripting as part of the gateway to commercial databases?

• Only if a vendor requires it. Universities can do their own scripting if they wish.
  
  

4. What are the options for authentication?

• IP; password; some scripted passwords.
  
  

Evaluation :

1. How well does this architecture accommodate non-commercial resources (e.g. locally digitized collections)?

• Digital collections are the responsibility of individual institutions. A special collections committee assists the process by coordinating efforts, offering workshops, and helping choose which items to include. VIVA merely provides links to the individual projects on their webpage.

2. Do they feel they are successful in meeting their goals?

• Yes. They are hiring an assessment consultant to help evaluate the project and decide on future roles.
  
  

3. What pros and cons do they see with what they're doing? What would they do differently?

• PROS -- good resources; very collegial; good support; tracking of cost savings; including the community colleges in the consortia brought community college resources to a level they would have been unable to achieve on their own (better for education across the state)
• CONS -- archival issues; statistics control
• DIFFERENT? -- larger staff so less burden would be on the institutions
  
  

4. What are their future plans?

• Hire assessment consultant to evaluate the project.
• Patron-initiated ILL
• More emphasis on full text
• Analyzing cost savings for local institutions