What can you do with a Library?

“The catalogue of an institution like the British Museum, dealing with a mass of matter already accumulated, and intended to register an ever-accumulating mass of matter for ever and ever, must not aspire at perfection, and can never attain finality”. – Richard Garnett, Keeper of Printed Books, 1890-99.

A principle starting point of this research is that data can be expressive of a cultural practice. The British Library’s operational infrastructure was designed for finding and delivering books and other items by encoding them as easily sorted and queried electronic databases. When readers request books for instance, they are generating “born digital” data, partly independent of the collection items themselves. This requesting data is all stored in a database called ABRS (Automated Book Requesting System), a system which has recorded every transaction since the building opened 20 years ago, and therefore having the potential to become an cultural resource in its own right. Yet this is a very functional and, as far as is possible, streamlined process that values efficiency above all else. And also users’ expectations of the systems are mostly very straightforward. We should not begin by assuming any cultural practice whizzing around the British Library’s data networks. So how can we break out of this straightjacket?

For a database that is highly instrumentalised like ABRS it may be best not to start by trying to expose a rich cultural process going on under the surface that may not currently exist, but to try to build one instead. To do this we also need to attend to the fact that such a practice forms amongst the relations between the data infrastructure, the actual conditions in which it is accessed and the significance that lives on from their pre-digital history. We have to attend to how to situate this born digital data in the wider context from which it takes shape and in which its affordances address themselves to a public. And we start with what we have got. It will be impossible to rewrite the software that governs the operation of a national library. So intervention is one key strategy here – diverting its current functioning, placing foreign bodies inside it, offering alternatives. We cannot do anything which prevents it from functioning, but we might create a situation in which people are motivated to use it in a non-functional way.

The history of the British Library’s digital catalogues seems to reveal a greater intrinsic cultural richness when we delve into their histories, including an incredible range of incompatible collection types and traditions of maintenance. In particular, we can detect a clash of cultures when decisions have been made about how to encode them into a database form. The Manuscripts and the Sound and Moving Image collections rebel against the dominant catalogues of printed matter with their single author, monographic model. Every few years there is a suggestion by a new senior manager to try to amalgamate all materials into one great database, an attempt which is allowed to run its course, regarded as it is by some as a utopian ideal. Will there ever be one database for everything? Is it even desirable? There are debates over the continuing necessity for shelfmarks, which can be seen as a relationship between a machine readable system number or a human readable code for categorising items. Now that the software need only system id’s with which to function, what do we want from the shelfmark?

The catalogues themselves are growing at a ferocious rate yet this is not so noticeable because in the background, not as a result of any the library users activities. The digitisation conversions of the late 1980s and early 1990s seem like and were herculean efforts at modernisation. Yet they were minor compared to the later efforts of migrating from previous databases to new designs, incorporating new sources of metadata and correcting errors from the old, a task which still continues in the background as people daily come across bugs and mismatches. There are constant programs of converting new sources of incompatible metadata, resolving the cataloguing requirements of radically different media types (who is the author of a sound recording of a haddock’s mating call?), dealing with the traditional cataloguing ways of identifying a person with their title, deciding whether a work is a version of a work or a new creation. The most interesting bits are those that don’t work properly because they reveal the databases histories and the efforts of members of staff to give the outward appearance of frictionless efficiency. The cultural practices that must be disciplined.

Like the requesting system, the usage of the catalogue is also recorded and the data analysed. This reveals two things that could be said to be going in opposite directions. The Explore search engine which is the “discovery” system for most of the catalogued collections, a proprietary system branded by BL, has several common features of a modern database interface. Readers can publicly Tag records, they can write Notes and have their own Workspaces where they can organise these references. Yet these features are rarely used. And when they are it is more as if for personal bookmarking purposes, or as though clicking a “like” rather than for collective knowledge enrichment. Is this an example of a Facebook style practice interfering with the practice of metadata classification and responsible cataloging? At the same time it is not like social media, seeming to default back to the practice of the solitary scholar. Here the trail runs cold. Now look back a stage at the activity of catalogue searching, the stage at which the British Library welcomes the reader as a user of a friendly and familiar search engine. The Explore interface is called a “discovery” system, emphasising its serendipitous abilities like a browser. When we examine how it is being used we quickly discover that Advanced Search with its powerful boolean operators and numerous fields is hardly ever chosen. Instead people type keywords into the single Anywhere field of Simple Search and then use the Author, Subject, Date, etc., filters to boil down the thousands of search results to something they can scroll through. Give me everything and I’ll junk what I don’t want – the opposite of what librarians and IT professionals alike regard as “good searching”. Here is what is described in disparaging tones as the Google-isation of the British Library catalogue. It is a strong cultural practice but is it a rich one we can work with or a ubiquitous software behaviour that is now set in stone? We may need to look in our toolbox of avant-garde shock tactics to move past these habits.

This project references techniques of data visualisation, but that cannot be limited to spectating the data as it flows around a library system. The purpose is not to create a mirror, even if it reflects arterial paths not normally seen. It must reveal something that creates a new context in which to think or act, a new rhythm that might resonate out and infect behaviour. That is the theme of the research that seeks new modes of cultural practice. It should ideally lead to as ongoing a process of exploring the digital and physical scene as possible, it should tend towards the prescriptive rather than the descriptive. The “born digital” data that is the basis of the British Library’s operations is still currently perceived in terms of a functional context. We could almost speak of a mismatch when compared to the way the physical space is evolving. To what extent is the British Library already a cultural institution based on data or is that what we have to make it? We may have to create our own born digital data, to be transformative we may need to invent new data relations.

Whether a new cultural practice is grasped or is picked up depends on public engagement. We cannot explain to users the whole functioning of the library on its technical level, nor would they wish to know. But perhaps we can create an image of it which they can build on. A joke, a provocation or a riddle. Perhaps it will turn out to be a lie but by that time it will not matter. This is also a point about the purpose of artists data visualisation, rather than revealing some hidden principle or turning a morass of data points into a recognisable pattern, it can be used to invert this process, by taking an assumed structure and showing where the gaps are, where there is space to move, the undefined data field, the misrecognised functions.

That software is not a tool but that it can participate in a poetics or is capable of providing a space for imaginative thinking is what makes it amenable to artistic practice. And so art is the last and most challenging goal here. Hard enough to make something that provides a new feature for someone to play with, an experiment that works or doesn’t work but tests the limits of a system or set of behaviours. This may be design or it may be an arts project or research. But it is an artwork that has the degree of autonomy necessary to form a new reference point, an idea with the distance from which it can permeate both feelings and actions. Unlike a good design but like any big institution such as the British Library itself, the best art will never work efficiently, its components aren’t compatible and it always demands another adjustment. Not a part of the operational infrastructure itself but a foil against which the infrastructure always finds itself lost, displacing the library always somewhere else.

Image gallery

Lewis Carroll Tittle Manuscript Courtesy of Alan Danskin Collection Metadata Manager at the British Library at Boston Spa.

This is how new books and other items were first recorded when they were acquired by the British Library up until the 1990s. After being printed, these original slips would be taken to the Title Room where they would be stored in boxes in alphabetical author order. These still exist in  storage at Boston Spa.


Printed Tittle Courtesy of Alan Danskin Collection Metadata Manager at the British Library at Boston Spa.

Once the manuscript title slip had been checked and found its correct place in the catalogue, it would be sent away to be printed. These printed title slips would then be copied four times, three of them for the catalogues and the “fourth copy” being taken to the Title Room to be stored in boxes in shelfmark order. Because the shelfmarks at that time were organised by subject, this “fourth copy” effectively constituted a subject index, although used primarily by staff and only occassionally being made accessible to Readers. These also still exist.

UKMARC machine readable catalogue record Courtesy of Alan Danskin Collection Metadata Manager at the British Library at Boston Spa.

This was the first version of the digital records produced by keyboard converting the printed catalogue during 1987 to 1991. On line 024 you can see a long number “21578907” which was a code devised to refer back to the volume, age and entry number in the old printed catalogue just in case it needed to be checked.

MARC21 machine readable catalogue record Courtesy of Alan Danskin Collection Metadata Manager at the British Library at Boston Spa.

This is the latest version of the MARC format for machine readable library records. The migration from the previous version, including more checking, deduplication and merging of additional metadata created the current catalogue database called ALEPH, and took place from 2004. It turned out to be an even bigger job than digitising the printed catalogues in the first place.


Title Slips

This is one of the old boxes of printed title slips. These are part of the so-called “fourth copy” which were stored in shelfmark order, the same order as the books on the shelves. They functioned as a check to make sure none of the books were missing as well as being used as a subject class catalogue.

Title Room archive at Boston Spa

This are where the old paper Title Slips are still stored, the manuscript set and the printed set. There are approximately 20 million in total. The hand written slips were the first record of the item made when it entered the library and functioned as a base line check for all subsequent information about that item.





“General Catalogue III (“GKIII”), interleaved or “laid down” set”

Up until the digitisation of the catalogues starting in the 1980s, the British Library used a hybrid system to keep up with the influx of new acquisitions. Each set of printed catalogues were cut up and interleaved with extra pages to make “laid down” sets. The title slips of new items could then be glued into the new pages in the right order but lightly enough to be moved around to make more space if necessary. In this way it combined some of the flexibility of an index card system with a fixed printed catalogue. This process of “maintenance” was a continual task that required enormous amounts of bibliographic knowledge, linguistic expertise and precise, concentrated error checking.

By Richard Wright Commissioned artist for the IoCT project

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s