Is the Browser the Library?

Tony Barry, Tony.Barry@anu.edu.au, Head of the Centre for Networked Access to Scholarly Information, Australian National University Library.
Paper presented to the Information Online & On Disc 97, 21-23 Jan 1997, Sydney.

Introduction

With the stampede of a range of publications from paper to the network, the role of the professions which support communication between writer and reader in publishing, book selling and libraries will be radically altered and rearranged.

The neat division of format between monographs, serials and manuscripts disappears and the concept of an "edition" blurs as documents become dynamic.

The reduction in the costs of publishing and distribution threatens the viability of traditional publishers and booksellers and direct access to end users by publishers bypasses the role of libraries.

Hypertext integrates document delivery into catalogues while blurring the distinction between a catalogue, a bibliography and a library all of which become finding tools and delivery mechanisms of similar functionality.

The role of indexers and cataloguers is threatened by the development of search engines which, while less effective, can provide indexes which can be compiled at a thousandth of the cost of manual indexing.

The key to the operation of the network as a library will be selection of sources and organisation of the material selected. Whoever does this will be the librarian of the future.

Why are there libraries?

Libraries exist for a number of reasons -

People's need for information exceeds their capacity to pay
The time required to locate quality material from the sea of available publications exceeds the time people have available
The skills required to organise large quantities of information are not readily available to people nor do they have the time to do so.
The skills needed to retrieve information of interest even when organised may be absent.

By changing the above factors network publishing will alter what libraries do and how they do it.

The costs of network publishing

The costs of electronic publishing have been extensively debated and although there is controversy over the amour with which costs will be reduced there is agreement that they will and that extensive self and vanity publishing is possible and clearly is occurring. Charles Bailey[1] has published an extensive bibliography of this literature. Without getting into detailed debate the net effect will be that more, probably far more, literature will be available, and that much of it will be unrefereed and of uncertain quality. Much of it will also be freely available.

The response of the mainstream journal publishers to networking have been to recognise the threat to their livelihood while attempting a variety of methods to provide network delivery for their products. For academic journals their best defence comes from the entrenched reward system in the academic community where publication in established peer review journals is the path to promotion. The publishers are able to sell this prestige and also their editing skills to enhance the material submitted to them and this continues to support the prices they charge for the foreseeable future. It is possible that we may see little change in the traditional literature in the short term but an explosion in the grey literature which will be increasingly useful.

One of the reasons for the existence of libraries is reduced by this cost shift. Other reasons are enhanced.

Filtering and Selection

I have argued in a paper[2] to the 1995 Questnet conference that the central problem on the network will not be information retrieval but filtering, both for the obvious practical reasons but also from the political need to restrict access to objectionable material however defined. That paper is badly dated now and there has been much development mainly centred round the PICS (Platform for Internet Content Selection) protocol. PICS is a protocol that allows groups to rate publications by whatever sets of labels they deem appropriate. These labels can then be accessed remotely to determine whether and item is worth retrieving. In part this could substitute for the decision by referee panels and publishers in deciding on the quality of an item or that of a collection development librarian deciding whether an item is of sufficient quality to add to a library collection. The rating is decoupled from the collection and can be done independently. Will we see libraries putting up PICS servers to indicate whether items are of value?

So far libraries have been active in putting up "subject trees" or pointers to quality material on the web but the most massive of these has been the W3C virtual library which is not supported by a library although parts of it are, as it is a distributed system. The BUBL subject tree is library based , on a cooperative model, and gives a classified approach to useful internet sites. Individual libraries have also assembled subject tries geared to the needs of their clientele such as the Australia National University Library.

From this we see that the role of libraries in providing selective access to information sources may be challenged by other groups wishing to do the same. In particular many professional groups are doing this for their membership as can be found by looking through their web pages linked from the Scholarly Societies Project at the University of Waterloo.

Indexing

One of the few areas which are making money on the internet, is in the provision of indexes. Richardson and I recently reviewed [3] internet indexes. The bulk of these indexes are now giving full text indexed access to many tens of millions of documents. None are run from libraries although some may now be employing librarians to assist in rating material and to provide advice on organisation. The central problem of these indexes is that they are based on free text indexing and therefore their precision performance even with relevance feedback will suffer.

Some attempts are being made to provide mechanisms to provide better subject descriptions through meta data proposals of which the most significant is the Dublin Core specification and consideration is being given to apply this to government information in Australia in the recent IMSC report [4]. As these mechanisms will only work if indexing terms are provided by the publisher the prospects of using controlled vocabulary for the indexing terms is slight.
While there have been experiments to catalogue the content of the internet in a traditional way more interesting approaches are being made in Britain in the ROADS project.

So far the impact of the library profession on indexing the internet has not been great and the dominant indexes are occurring outside the library sector.

What will and electronic library look like?

Network publishing places material effectively direct to the desk of the end users and effectively direct to the shelves of a global library. The creation of a database or set of pages which access material selectively creates what can be regarded as a searchable bibliography but equally it has the functionality of a library.

Library catalogues which consist of a database of document surrogates give way to full text indexes of material selected for groups of clientele access via pages which can equally be regarded as bibliographies or catalogues but which in the end are delivery systems. Once such a database is provided however it can be made available to anybody across the network and make the creating of any similar services unviable unless there are deficiencies in what is provided. The need for local libraries are reduced.

What then becomes the role of the library when there is no longer stock to house and the tools to access the material might also not be required as they are provided by others?

From the end users point of view the "library" will be the network and the tool to access this will be the browser and its successors - be they Network Computers (NCs) or the browser function merged into the operating system. While the end user my acquire information direct from publishers (who may also be the author)
there will still be the key tasks of selecting material on the basis of quality and providing search and discovery facilities for the user. Whether it will be done by librarians or some new professional group is still an open question.

References

1. Bailey, Charles W., Jr. "Network-Based Electronic Publishing of Scholarly Works: A Selective Bibliography." The Public-Access Computer Systems Review 6, no. 1 (1995). (Version 26: 10/25/96)

2. Barry, Antony, NIR in not enough , Questnet '95, Bond University, 6-8, Sep. 1995

3. Barry, Antony and Joanna Richardson, Indexing the Net - A review of indexing tools. Paper presented to AusWeb96, 7-9 July 1996.

4 Architecture For Access To Government Information Report of the IMSC -Technical Group, Canberra, 1996.