The Librarian and the Webmaster
Paula Garrett, Fisher Library, University of Sydney, NSW 2006, Sydney, Australia. Phone +61 2 351 3257 Fax: +61 2 351 2890
David J. Ritchie, Fermi National Accelerator Laboratory, Batavia, Illinois 60510 USA. Phone 1 708 840 3940 Fax: 1 708 840 2783
Keywords: World Wide Web, High Energy Physics, Collaboration
In early 1992, Ruth Pordes and Jonathan Streets of the Fermilab Computing Division's [HREF4] Online Systems Department [HREF5] (OLS) were considering the problem of providing information about online data acquisition systems to high energy physics experimenters. Seeing the WWW presentation to Artificial Intelligence in High Energy Physics (IHEP'92 ) at La Londe, France in February 1992, Streets recommended WWW as being "the best thing around," and OLS decided to adopt it.
A server to provide online data acquisition system documents [HREF6] was set up by Tim Berners-Lee [HREF7], CERN , and Jonathan Streets, Fermilab, on a visit by Berners-Lee to Fermilab in July 1992. Tim Berners-Lee went on to visit NCSA [HREF8], the National Center for Supercomputing Applications in Urbana, Illinois beginning the collaboration whose result was NCSA Mosaic for Mac [HREF9], PC [HREF10], and X (Unix) [HREF11].
That first server ran on a VAX/VMS system and used DCL (Digital Command Language) in its implementation. In response to browser requests, it extracted full text documents from a previously existing Fermilab-written document database. This 1992 technology would be considered fairly primitive by today's standards. Using a revised implementation, the server continues in operation today.
July 1992 through January 1994 saw the creation of an estimated ten to fifteen servers by various groups associated with Fermilab. Some, such as E781 [HREF12], were created by experimenters to help organize the dispersal of information within their far-flung collaborations; others, such as the Sloan Digital Sky Survey [HREF13] were created by Fermilab support groups to help with the development of software and hardware intended for a particular project.
In February 1994, it became clear that WWW was gaining a real presence at the laboratory and that an actual Fermilab Home Page [HREF14] was needed to bring together the servers created by the "early adopters" as well as set the stage for further use of the web technology. Initially, a process involving consultation with a broad set of relevant groups was imagined. However, the impending announcement of evidence for the top quark [HREF15] made it imperative to have the home page up quickly. Instead, a small team drawn from the Fermilab Directorate and Fermilab Computing Division was charged to accomplish the task. Beginning with a concept authored by Judy Jackson, Fermilab Directorate, and Liz Quigg, Fermilab Computing Division, an implementation was put in place by the date of the announcement on April 28, 1994. The Laboratory continues to add additional links to its home page and to update it with significant announcements-the most notable one being the announcement of the discovery of the top quark [HREF16] posted on March 2, 1995.
Just prior to the April 1994 date, the Fermilab Library, knowing the intense interest of its readers in the scientific papers concerning the evidence of the CDF [HREF17] and D0 [HREF18] collaboration, set up a Fermilab Library web server [HREF19] in order to provide references to the top quark papers. The Library continues to extend and enhance its server with reference offerings as well as information about the Library.
In the months since, the Laboratory and Library servers have evolved in distinctly different ways. An effective collaboration between the librarian and the webmaster has occurred. Some information has been listed on both servers. In other cases, the choice has been made not to list the item on one or the other. A certain intentional complementarity and dissonance of approach has evolved. Looking back, we have come to realize that this process has been valuable and stimulating to the evolution of both servers and may hold a key to the way other organizations may wish to orchestrate their web offerings.
In the remainder of this paper, we describe the architecture of the Laboratory and Library servers, discuss the goals, the implications of those goals, the drivers of structure and content, and the effect of these factors on the architecture. Comparisons are made. The implications for Libraries within organizations are described. The application to large Libraries which begin to have multiple servers within their own organization (e.g., one in reference, one in cataloging, one in circulation) is made. The broader implications are discussed.
A year before the setting up of the Fermilab Home Page, however, the web was simply a device for technical communications between an experiment's collaborators with little, if any, visibility to the external community. The growing public presence of the web meant that creators of web servers had to accommodate themselves to the Directorate's beginning to play a role and the Directorate had to educate itself about the importance of the web as a significant vehicle for presentation of the laboratory to the public.
Result: An individual within the Directorate who already had certain responsibilities for media relations was designated as Editor of the top level page and those pages designed for the general public. An individual from the Computing Division was designated Editor of the technical pages-these being Fermilab At Work [HREF26], and its related links.
A relatively flexible procedure involving e-mail consultation between the two Editors and the webmaster was set up to coordinate proposals for changing the home page. See the United States Department of Energy Televideo Service [HREF27] home page for a much more structured approach in this area.
Experiment web pages truly describe "work in progress"-unpublished, unrefereed material inappropriate to be cited outside the collaboration. Support groups similarly use the web for provisional matters.
The concept of a "work in progress" area became clear. What made this concept particularly useful was that "work in progress" was a sensible disclaimer generally understood by many-not only individuals involved in research but also those associated with funding agencies supporting that research.
Result: The Fermilab At Work [HREF28] section assumed a prominent role in the web offering.
Fermilab is very much a "laboratory without walls" and has operated in this fashion since its beginning in 1967. Those who have contemplated the "library without walls" concept will understand the unique challenges such a structure brings. A member of the Fermilab research community may be anywhere-on the laboratory site, at the college just down the street or at a university in Australia.
It would have been impractical and administratively very costly to restrict access to the "work in progress" collection of pages at an overall level. Indeed, it would have had a chilling effect on collaborative work. Restriction by browser IP address would not have worked because of varied individual researcher affinities. Usernames and passwords would not have worked because of the dynamic way individuals move in and out of collaborations.
Collaborations that wish to isolate the more sensitive areas, such as minutes of collaboration meetings, particularly controversial results, etc., do make use of access restrictions. It is easy for specific collaborations to employ these techniques because they, with their knowledge of the collaboration, are able to manage access much more easily.
Result: As with any public information channel, it was important to have appropriate structures set up to allow the laboratory speak clearly and with one voice. The designation of the editor, based out of the Directorate, addressed that matter. By design, the "public voice" of the laboratory could be presented in the top level page and the pages designed for the general public while work in progress could be described in the Fermilab At Work [HREF31] section.
Result: The matter was deferred. No area on the home page was formally designated as a publication area.
There are clearly documents within the Fermilab At Work [HREF32] section that are nearing publication. However, they are regarded as either (a) work in progress, (b) in support of that work, or (c) material presented in the spirit of inviting "community refereeing," additional collaborative work or co-development. A particularly good example of the latter is the Fermilab Software Tools Program [HREF33].
To address the deferral, a collaboration between the Fermilab Publications Office, Library, and Computing Division was set up to work through the issues. A number of items (varied input formats, the need for patent and copyright review as well as peer review) make this a complex task. Once accomplished, a Publications Fileserver will be put in place. Authors will be able to submit their papers and readers will be able to obtain publications-all over the web.
Within the Fermilab At Work [HREF34] page and related pages, the approach of providing different access points was followed-borrowing here from library organizational techniques.
Many people simply want to access what is new. Others know the laboratory organization chart and wish to find their material via that route. Others still know the name of the experiment or large scale project. Others simply know the general category of information. Finally, others want a very focused guide to external web resources with some "jumping-off" points to broader "meta-page" guides.
All these access methods are provided but without elaborate page structure. Duplication under each access method, if it serves the reader, is allowed. Since screen real estate is at a premium and short load time is very desirable (this is, after all, a list for those at work.), simplicity is the rule: the structure is that of a simple list indented as appropriate.
This "local webmaster" is expected to be a contact with whom the Fermilab webmaster could discuss technical issues (e.g., server down, change of server node, etc.). The local webmaster is also expected to make sure that the server content is consistent with various proper computing usage policies. A distinction is made between collaboration/support-group "institutional" pages and home pages for individuals. The "local webmaster" is held responsible for the content of the "institutional" pages. The individuals are held to conformance with proper computing usage policies in the content of their home pages. When setting up the individual's home page, the local webmaster is expected to remind individuals about these policies which they have seen before at the time they received their computer account.
An element of "line management responsibility" was put in place by requiring the local webmaster's supervisor (or official spokesperson in the case of collaborations) sign a form designating the particular individual as local webmaster. In addition, a supervisor is strongly encouraged to obtain the blessing of their Division or Section head in order to make sure there is a clear understanding at a high level about what is being done.
With respect to the collaboration/support group pages and servers, there has been a modest amount of comment about the sign-off bureaucracy. This has been weathered. The delegation of responsibility for content to an accountable, recognizable line management chain has given management (and certainly the webmaster) a certain comfort level in linking independent servers to the Fermilab Home Page.
At least in one circumstance, the arrangement worked well when a particular content was questioned. Thus, the procedure appears to work and, most importantly, the collaborative use of the web and web servers at Fermilab appears to be healthy, on-going, and adopted in greater and greater numbers.
When a library's major collection changes formats, focus on access must shift. The typical Fermilab Library reader is highly literate in electronic media and in turn has high expectations in regard to information access and delivery. At the same time preprints were migrating to such places as the xxx e-Print archive [HREF36] at Los Alamos National Laboratory [HREF37] and to any of a number of other preprint bulletin boards [HREF38] as well as onto WWW servers of specific research individuals and groups, more and more information sources were also becoming available via the web. Through its own single web server [HREF39], the Library was able to direct its readers not only to WWW resources, but also to gopher, ftp and telnet sites. WWW has become a common denominator in this high energy physics setting as a way of making active references to information sources.
With its own server, the Library is free to make links, change links, break links, and change the design of the hierarchy of resources-all in conjunction with the ever-changing reader needs. As the maintainer of the Library's web server, the head librarian incorporates new web resources discovered by means of listservs, library literature, library and laboratory colleagues. Most of these are placed under the "Hot Links" [HREF40] section of the Library's home page.
Links to the full-text of these two preprints were the first ones the librarian added to the Library server's "Hot Links." Also handling most of the reference queries, the librarian increasingly turned to WWW first in tracking down government documents such as Presidential press releases and science policy statements requested by the Laboratory's Director and other readers. Good URL "finds" generated from this work were often times incorporated into the Library's "Hot Links" [HREF41] section.
The frequency of the changes is an element of the chosen architecture for the Library's web server: reader-driven change.
The focus in this paper has been Reference Services but changes are occuring in other service areas as well. It is entirely possible to imagine multiple web servers within a large library-some devoted to circulation (e.g., to support reader access to his/her own circulation records), technical services (e.g., to provide information on recently cataloged materials), and reference (e.g., to provide access to "Hot Topics" pages maintained by individual reference specialists for their particular specialty fields). Indeed, a number of libraries already are making innovative uses of web technology along the lines described.
See for example, Innovative Internet Applications in Libraries [HREF46] maintained by Ken Middleton of Todd Library, Middle Tennessee State University and also Fisher Library at the University of Sydney [HREF47].
When a library begins to have multiple servers with multiple collections of pages maintained independently by varied individuals across the institution, the web organization issues begin to resemble those for a large laboratory, such as Fermilab. It may be that the solutions described here will also apply.
The ease of gathering and authoring via the web empowers librarians with a greater independence and flexibility. As a result, they will be able to change their library's public interface at a pace more in step with reader needs instead of being so highly dependent on vendors, computer professionals, or others.
The web's "Uniform Resource Locator" (URL) concept is key to making this ability to switch quickly possible as is, of course, the underlying presence of that uniformly accessible, totally interconnected, network of networks, which is the Internet. Together, the two provide a seamless interface to these varied information resources, bringing on the next phase of library automation by creating what one might call the information marketplace.
Of course, the information marketplace has existed for a long time-the pivotal aspect of this new information marketplace is that the point of purchase has changed. Formerly, the decision to go with this or that commerical database vendor was likely made in the library administrator's office once a year at contract time. Now, the decision is made by the individual reader as he or she chooses this or that network information resource on possibly a minute to minute basis.
Currently, in the "spirit of the internet," there are many "vendors" supplying information for free. Given the ease of authoring and making documents available on the web, this may continue for some time. Still, the demand for authoritative documents or those with other special characteristics accessible in the same way as the free material is giving rise to vendors who provide access to information resources on the network for a price.
Vendors naturally work to have their resources made essential and placed at the center. Special browsers, special servers, or other software that use the underlying web and network transmission rules are all possible. Regardless of what vendors devise for providing access to their information over and above the basic Mosaic browser kinds of capabilities, libraries must require conformance to standards and mandate interoperability (e.g., require that information servers from various vendors work with a wide selection of information browsers).
Not only will this force competition into the information marketplace, encouraging high quality and low price, but it will also allow the library to stay "light on its feet" and very adaptable to the highly diverse, growing numbers of information resources on the Internet.
By so doing, libraries will insure that they, in the next phase of automation (which in fact is now occuring), can continue to be reader-centered rather than system-centered and thereby be of maximum benefit to their reader communities.
T. Berners-Lee, R. Cailliau, A. Luotonen, H. Frystyk Nielsen, A. Secret, "The World-Wide Web," Communications of the ACM 37, (1994) 76.
For information more generally about the conference, see the URL: http://www.scu.edu.au/ ausweb95/
This document was prepared with MicroSoft Word and converted to HTML for the network-resident version using rtftohtml. For further information, see ftp://ftp.cray.com/src/ WWWstuff/RTF/rtftohtml_overview.html.
1 Head Librarian, Fermi National Accelerator Laboratory, January 1987 - March 1995.
2 Webmaster, Fermi National Accelerator Laboratory.
3 T. Berners-Lee, Presentation given at Fermilab, July 1992 (unpublished).
4 J. Streets, private communication.
5 Penelope Constanta-Fanourakis, Guide to Using the Document Database - DOCDB, PN 336, March 29, 1988, (unpublished).
6 The top quark is the last to be discovered quark in the Standard Model theory linking all sub-atomic constituents. Researchers have searched for it in physics experiments for some eighteen years. In April 1994, evidence for it was announced by the Fermilab CDF and D0 Collaborations. In March 1995, CDF and D0 announced its discovery.
7 For example, in March 1993, one experiment's private unpublished meeting minutes requested collaborators to store material in the experiment's VAX Notes electronic "conference," a vendor-specific, platform-specific, system somewhat like a bulletin board. By June 1993, the experiment's minutes had announced V1.0 of the experiment's "Documentation System," a WWW Server and implied the demise of VAX Notes. The sense in the announcement was that merely of setting up another channel of collaboration communication-of little relevance to the laboratory Directorate and the laboratory's public posture.
8 P. Garrett and D. Ritichie, "Fermilab Library Projects," FERMILAB-TM-1667 (1990) and P. Garrett and D. Ritchie, "Fermilab Library Directions," FERMILAB-TM-1668 (1990).
Submitted to AusWeb95 The First Australian WorldWideWeb Conference firstname.lastname@example.org