[WEB4LIB] Betr.: Building a database for e-journals

Eric Hellman eric at openly.com
Wed Oct 17 16:28:17 EDT 2001


It is an over-simplification to say that OpenURL and person link 
pages are the solution to this problem. They are certainly a part of 
the solution. The number of resources that work with personal link 
pages, in particular, is at present insignificant.

First, a disclaimer: my company, Openly Informatics, 
(http://www.openly.com/) is deeply involved with the full variety of 
players working to address this situation. We develop and sell 
linking components, software, and linking data. I am a member of the 
NISO committee AX, which is working to standardize OpenURL.

THE GOAL

Our goal is one-click access to everything. In other words, when a 
user sees a reference to literature, either in an A+I database or in 
the references in an article, they should be able to click on the 
reference and get to an appropriate resource for that article in as 
few clicks as possible. Maybe asking for one-click access to 
everything is a bit like asking for Peace on Earth, but it's very 
much worth striving for.

THE TASK
In order to make this happen, a lot of groups with different and 
often conflicting interests have to cooperate. (See... I told you it 
was a lot like striving for peace on earth!)

1. Information providers (publishers, aggregators, libraries, etc.) 
need to embed hooks in their content that can be pointed at a user's 
information agent. The format for the hooks will be customizable 
links called OpenURL's, which is what the NISO Committee AX is 
working on. The OpenURL standard will be based on a submission 
authored by Herbert van de Sompel, now of the British Library, and 
Oren Beit-Arie of Ex Libris.

2. Libraries need to deploy and configure OpenURL link-servers. 
Link-servers are software agents that know what electronic content a 
user has access to and how to direct the user to that content. The 
link-server acts as a hub which directs users to the appropriate 
content. Within a few years, link-servers are likely to become an 
integral part of any Integrated Library System worth paying money 
for, but for now link-servers are separate components or services.

3. Information providers (publishers, aggregators, libraries, etc.) 
need to be friendly to incoming "deep" links. If a user has 
bibliographic information for an item, either entered by hand or 
carried on an OpenURL link, the user ought to be able to immediately 
get to resources for that item that the user's library has paid for. 
This means sensible authentication mechanisms and automated access. 
No user wants to drill down through ten search screens or directories 
when they already know what they want.

THE PRESENT
Here's how the current situation looks in each of these areas.

1. OpenURL. Thanks to the marketing push by Ex Libris for SFX, a lot 
of information providers have committed to adding OpenURL support in 
their products. The number that have actually implemented OpenURL is 
much smaller. Notably, many of the large aggregators have 
hooks-to-holdings features in their current products that can be used 
to construct OpenURL links to library link servers. EBSCO, Proquest, 
CSA and Gale and many others  have such capabilities which range from 
sophisticated to trivial.

Libraries can accelerate this trend primarily by asking for OpenURL 
when it comes time to make purchasing decisions. The web site for 
OpenURL is http://library.caltech.edu/openurl/

2. There are two parts here, deployment and configuration.
A. deployment. Link-servers come in a variety of shapes and sizes. 
While SFX is the best known and most full-featured, you can also 
deploy the free link server (based on Jake data) available on our web 
site. OpenURL link-servers have been announced or are being offered 
by Fretwell-Downing and Endeavor. Some universities have turned their 
e-journal databases into link-servers. GODOT, developed at Simon 
Fraser University for COPPUL, is a good example. A Korean Company, 
KINS, has developed a link server for the asian market. Certain 
aggregators may be able to turn their existing linking systems into 
lightweight link-servers. This is likely to be a market with a 
variety of options.

B. Configuration. This is the hard part at present. To configure a 
link server, you have to know titles you have access to, and how to 
access them. You have to build an e-journal database. Unfortunately, 
title lists at many aggregators are not as well maintained or as 
complete as they ought to be, making this a lot more work than it 
should be. A number of solutions have popped up to alleviate the 
aggravation being experienced by many e-journal librarians. Serials 
Solutions has helped a lot of people, and their data has been used in 
link servers, in a prototype with us and in at least one SFX 
installation. TDNet offers a bundled service, but they where not 
doing OpenURL, at least when I talked to them last. I don't know 
anything about JournalWebSite. Jake was a great solution a year ago, 
but  data maintenance has stalled. Ultimately this problem has to be 
solved with closer involvement by the information providers and the 
implementation of standards, and that's the tack we're taking.

Libraries should start thinking about link servers now. The time and 
effort  put into getting e-journal databases into shape will not be 
wasted.

3. Linking. This part is rapidly being solved in two ways.

A. CrossRef. I tell people that CrossRef is a miracle, because it 
represents unprecedented cooperation and agreement between all the 
major science publishers. CrossRef is a consortium of over 70 
publishers that have agreed to contribute article metadata and links 
to a common database, which currently has over 3.5 million items. 
Libraries can become affiliate members of CrossRef for $500/year. For 
more info on how to use CrossRef, see http://www.openly.com/crossref/

B. Link friendly sites. More and more, both aggregators and 
publishers are recognizing that their products become more valuable 
when linking is easy and stable. We can now direct-link to 7700 
e-journals with the JournalSeek database. Even most of the 
aggregators who used to be hostile to direct linking are now trying 
to retrofit their services to allow it.  The California Digital 
Library in particular has been active in prodding publishers to make 
linking easier, by putting it as a requirement in their procurement 
process; this is something that could be emulated to beneficial 
effect.

THE FUTURE

Although today the talk is mostly about link-servers for libraries 
and consortia, I think that eventually everyone will have a link 
server  (or personal link page) operating as a plug-in to their web 
browser, with preferences customized transparently to each 
individual. Authentication and rights management will be built in, 
and the resulting experience will make browsing the professional 
content as simple and easy as the free-content web is today. It's a 
sad reality today that paying for high quality content results in a 
poor user experience because of all the primitive and clunky ways 
that access control is implemented. Oldtimers will tell their 
unbelieving younger colleagues war stories about how they once had to 
go and munge title lists by hand while they sat at terminals in 
*buildings* called libraries.

Eric


>
>
>The solution to your problem is OpenUrl,  a person link pages (plp), 
>that act as OpenUrl resolver and the mechanism to let your browser 
>know to which personal link page to link. A simple demonstration can 
>be found at http://www.kb.nl/persons/theo/  .
>When you look into these pages you see what mechanism is used to let 
>the browser know where to link to. It would be great if everybody 
>uses the same mechanism. In that case everyone can always link to 
>his own content provider independend of where you found your 
>metadata.
>
>With regards,
>Theo van Veen
>
>  >>> Marlène Delhaye <marlene.delhaye at bu2.univ-mrs.fr> 17-10-01 08:24 >>>
>Good morning,
>
>I have to build a large database to manage and provide access to ca.4500
>e-journals for universities members of a consortium. The main difficulty
>for me is that people on different locations shouldn't have access to
>the same content (as universities don't subscribe to the same journals
>packages).
>I'd like to have information about the technical solutions you've
>choosen in your libraries, and about the time, money and skills you've
>needed to realize this project.
>
>Thanks for your help and experiences,
>
>Best regards,
>
>Marlène Delhaye
>Documentation électronique
>-----------------------------------
>Service Commun de Documentation
>Université de la Méditerranée
>27 Boulevard Jean Moulin
>F-13385 Marseille cedex 05
>t : 04 91 32 45 38
>f : 04 91 25 60 22
>e : marlene.delhaye at bu2.univ-mrs.fr
>http://bu2.timone.univ-mrs.fr/
>-----------------------------------

-- 
Eric Hellman
Openly Informatics, Inc.
http://www.openly.com/1cate/      1 Click Access To Everything
http://my.linkbaton.com/                Links that Learn
http://addaflag.org/                        Raise the Flag on your Website


More information about the Web4lib mailing list