<nettime> David Gugerli on Data Management as a Signifying Practice

[Geert Lovink <geert@xs4all.nl>]

(Dear nettimers, the videos of most of the presentations at the INC  
conference Society of the Query are now available online at http://networkcultures.org/wpmu/query/videos/ 
. Konrad Becker, who was there to launch the Deep Search book,  
announced a next search event in Vienna, in May 2010, if you wish the  
fourth in a series of events in Europe on this topic. Here at INC  
we've discussed to turn the blog of the Amsterdam event into a more  
permanent location where interested can find, share information on the  
politics, aesthetics and culture of 'search'. If you're interested to  
join this collaborative blog, please write to Marijn at  
networkcultures.org. Below you'll find the text of David Gugerli on  
the theory and history of databases. Enjoy! Geert)

Data Management as a Signifying Practice
David Gugerli, ETH Zurich
November 13, 2009, Amsterdam

Edited by: Baruch Gottlieb

Databases are operationally essential to the search society. Since the  
1960’s, they have been developed, installed, and maintained by  
software engineers in view of a particular future user, and they have  
been applied and adapted by different user communities for the  
production of their own futures. Database systems, which, since their  
inception, offer powerful means for shaping and managing society, have  
since developed into the primary resource for search-centered  
signifying practice. The paper will present insights into the genesis  
of a society which depends on the possibility to search, find, (re-)  
arrange and (re-)interpret of vast amounts of data.

I am aware of the fact that the title of my talk is both very  
ambitious and theoretically subversive. The “Culture of the Search  
Society” undermines the distinction Gilles Deleuze once made between  
the operating principles of the Foucauldian societies of discipline on  
one hand and the operating principles of late capitalist societies on  
the other hand, i.e. societies, which seem to replace earlier  
disciplinary surveillance techniques of inclusion and exclusion with a  
diverse set of juxtaposed rules that rather serve to control “input /  
output relations”, i.e. societies that are tightly linked to the  
notion of management and the allocation of resources. In a somewhat  
paradoxical sense, Deleuze’s control society is a society which is  
characterized by a high degree of flexibility, by distributed, rather  
than hierarchical, networks, by stochastic processes, and by an  
increased level of tolerance with regard to norms. And it is a society  
which is flourishing on a both infrastructural and cultural seedbed of  
search practices or search technologies. It is, I want to argue, not  
so much a control society but rather a search society.

The world as a database: CSI

Let me start with something probably familiar, something which you  
actually might have seen on television. “CSI: Crime Scene  
Investigation”, one of the most popular, Emmy Award-winning, CBS  
television series, trails the investigations of a team of Las Vegas  
forensic scientists as they unveil the circumstances behind mysterious  
and unusual deaths and crimes.

Most episodes conform to the traditional detective story whodunit- 
structure and depict the work of two forensic teams, which are usually  
analyzing two different cases of murder at a time, using, in both  
cases, the most sophisticated technological and scientific means for  
their forensic laboratory work.

One reviewer stated that the series’ techno-scientific orientation has  
an astonishing effect both for the role of the murderer as well for  
the figure of the victim. In fact, both of these figures are,  
dramaturgically speaking, only interesting as carriers of evidence.  
There is no attempt at understanding the social dynamics between  
murderer and victim. The motives of the suspect are almost irrelevant,  
the tragedy of the victim is not really taken into account. There are  
a dead and a living body whose encounter in the past has produced  
forensically relevant evidence. In addition to the crime scene, the  
two bodies involved in the crime are, to put it bluntly, a mere  
repository of traces, a hub of evidential markers, a base of pieces of  
information that can be retrieved, technically stabilized and  
scientifically analyzed in order to – and this is crucial - recombine  
them in such a way that the whole set of aggregated data might verify  
or falsify parts of an ever more differentiated hypothetical narrative  
of the crime under investigation. These scenarios are shown in a  
somewhat blurred view of the investigating agents' imaginations as  
they indicate where they should look for more data at the evidence- 
providing crime scene. One can say without exaggeration that CSI is  
depicting the world as a database; its dramatic development is about  
the excitement of search and query.

Global Software Business 2007

The popularity of the database as a general model for search and query  
processes is not only evident if one is wasting his or her time in  
front of a TV set. Database management systems are also shaping our  
“real” world to a great degree. Most importantly, DBMS are big  
business. A glance at one of many market reports can give us a  
sufficient orientation. Carl W. Olofson wrote in a study released last  
year by IDC: “The RDBMS market is estimated to have grown by more than  
12% from $16.7 billion in 2006 to $18.8 billion in 2007.” Moreover, it  
is noteworthy that the top three vendors control 84% of the global  
market (IBM, Oracle, and Microsoft) and that customers face a high  
client lock-in due to the costs of migrating data and data models.“  
They are, so to speak, highly dependent on their data management  
systems."

A signifying practice

Popular culture, globalized markets, and the path dependencies of  
enterprise are certainly the most promising entry points for a  
historian’s account of this emerging computer-based field of practice,  
especially if we remember the crucial fact that DBMS - like all other  
technologies - are mainly instruments both for their producers and for  
their users to shape society and to manage social change.

Probably the most important feature of a state-of-the-art database is  
its virtually unlimited possibility to fine-tune, to enlarge, to  
aggregate, to export and to recombine datasets or even entire  
databases. Current database management systems have enormous  
flexibility. However, this flexibility has been developed over a  
considerable period of time. They have, so to speak, a history of  
their own. This becomes clear if you have a look at the computer and  
software landscape of the late 1950s and the 1960s, where computers  
were dealt with as exceptionally powerful calculators, whose programs  
were run in a linear, batch-processing mode.

At that time, computers were the obscure, powerful, tireless and  
efficient assembly-line workers that could sort, count and even filter  
all sorts of almost endless lists of data. Complex data, however,  
could not be treated by such a computer in an efficient way – its  
analysis was a very cumbersome task. No wonder, people started to  
dream of another kind of data retrieval and search than the one that  
was available.

The main problem

Take for instance, the very fundamental limits faced in the 1960s by  
people trying to realize the dream of a Management Information System  
which could provide an endless pool of information ready to be parsed  
by decision makers. It would be too simple to blame the  
insufficiencies of the hardware of the for these limitations. Of  
course, storage and memory were expensive, and processing power was  
nothing compared to even a present-day handheld device. However, the  
more important limitations I am referring to were of a conceptual  
nature.

At that time, data was usually stored in a hierarchical, tree- 
structured array. In order to retrieve a specific item, the programmer  
had to know exactly its logical position, sometimes even its physical  
address on a disk or on a magnetic tape. A manager would not be able  
to actually use a computer other than through the review of  
standardized output on endless reams of paper. The programmer on the  
other hand was a highly specialized navigator who could find a  
particular path to a particular item within a specific set of  
variables. Given these circumstances, even for a specialist it became  
a very troublesome if not an impossible task to rearrange data which  
had previously been stored within a particular expectation of how the  
information will be used.

This is why a database of the 1960s could only answer questions that  
were already foreseen when the database was constructed: The structure  
of the database defined what kind of questions it could answer.  
Nevertheless, the tempting idea of an all embracing pool of  
information which allowed for the recombination of data and the  
aggregation even of whole databases, a database which was open for  
future problems, questions, and interpretations, had already occurred  
to some people by the early 1960s.

The dream of a pool of data from which relevant pieces of information  
could be selected in any desired combination became theoretically  
feasible in the 1970s. This new form of database was gradually  
implemented in the 1980s and has had a huge practical and economic  
consequences throughout the last one and a half decades. The  
fundamental question to be solved was how to separate storage from  
retrieval. This problem was dealt with for the first time in a seminal  
paper presented by Ted Codd in 1970, which can be seen as a pivotal  
moment of transition from hierarchically structured databases to a  
relational model of databases.

Ted Codd and “Large Shared Data Banks”

In terms of intellectual achievement, Ted Codd’s contribution to the  
development of Large Shared Data Banks, as he called them was highly  
original. He was absolutely convinced that “future users of large data  
banks must be protected from having to know how the data is organized  
in the machine (the internal representation). (..). Activities of  
users at terminals and most application programs should remain  
unaffected when the internal representation of data is changed and  
even when some aspects of the external representation are  
changed.” (Codd 1970, 379)

Two principal steps led Codd to his relational model. Firstly, he  
decided to organize all data in tables consisting of records in rows  
(these are also called tuples) and attributes in columns.  
Additionally, each record has its own key which uniquely identifies  
the record. This key eventually allows for the combination of records  
in two different tables.

Secondly, Codd and others developed a search and query language which  
allowed for a mathematically consistent retrieval of data stored in  
related tables. Whoever knows how to use a SQ-Language, is able to  
interpret any possible combination of records in a database, without  
being forced to study the physical address or the logical position of  
an entry.

Recombination of Data

Thus, recombination and relation became the most powerful mode of  
query and interpretation of data. Since the interpretation of data  
suddenly became independent of its form of storage, database managers  
did not have to anticipate future searches which might be done on  
their database. Therefore, an existing database could be confronted  
with new unexpected questions.

Tables certainly have been the most prevalent means of representation  
used by statisticians and political scientists ever since the late  
18th century. The statistical bureaus of the nation-state have  
developed this technology to near-perfection during the 19th and early  
20th centuries.

The database-driven views and reports of the late 20th century,  
however, produced a new category of tables. These are characterized by  
a configuration which is, in principle, dynamic in its composition,  
produced in real time and able to be reconfigured at practically no  
marginal costs. This leads to a remarkable discursive shift from a  
narrative based on a standardized table to a narrative based on a  
flexible tabular view.

Interpretative Flexibility and Critical Theory

The reason why I think that this discursive turning-point is  
historically so exciting is to be found in a temporal coincidence with  
a development, which is was completely disconnected with what happened  
in the realm of relational databases. This coincidentaly simultaneous  
development allows for a very similar emancipation of the signifying  
practice to that of the report-generating user.

In this parallel development I will describe presently, the user is  
usually called a reader. In the second half of the 1960s and the early  
1970s, French critical theorists elaborated a concept of the literary  
work which takes into account the interpretative flexibility with  
which a reader may approach a text. In the theoretical reasoning of  
Barthes, Derrida, and Foucault, a literary work like a novel or a poem  
are seen as a machine which enables its readers to produce  
interpretations with a remarkable degree of flexibility.

For Roland Barthes, for instance, an ideal text “is a galaxy of  
signifiers, not a structure of signifieds; it has no beginning; it is  
reversible; we gain access to it by several entrances, none of which  
can be authoritatively declared to be the main one; the codes it  
mobilizes extend as far as the eye can reach, they are  
indeterminable.“ As Barthes argues in S/Z, published in 1970, i.e. the  
same year Codd published his paper on the relational database model,  
the interpretation of a text cannot be determined by its author. There  
is no direct connection between the author’s intentionally structured  
text and the views or interpretations a reader produces while querying  
the same galaxy of signifiers.

Patterns of the Search Society

These are, I should add, not just signs of mere shifts of intellectual  
vogue. Both in critical theory and in the development of database  
concept, we find an astonishing coincidence of change towards  
interpretative flexibility, genuine fascination with the open work (to  
quote Umberto Eco), and attempts at separating the structure of the  
input from the structure of its derivative output. The cultural  
consequences of this change in operative preconditions are enormous.  
They obviously affect not only the relation between author and text on  
one hand and the text and its readers on the other, they have profound  
bearing on information processing of any form. This paradigmatic shift  
in information processing has had tremendous organizational and  
societal consequences. Newly emerging organizational structures, new  
administrative and productive processes, new forms of resource  
allocation have been some of the most important consequences of this  
change in information processing.

The most prominent consequences of changes in software architecture  
and social organization can be observed in the large industrial and  
financial corporations of the late 20th century Real time production,  
lean production, logistics, supply chain management, human resource  
management, all these tasks have been considered since the middle of  
the 1970s the most important challenges faced by late modern  
enterprise. All these processes are determined by or at least  
supported by database management systems. As a consequence of the  
implementation of relational databases and other enterprise  
application software which reside upon such databases, the entire  
corporate world of the late 20th century has become subject to the  
manager’s disposal and command.

The database-driven enterprise is represented and available in all its  
operational processes and structures. Public administrations and even  
universities are joining the club of Oracle and SAP customers. The  
operations and logistics of a supermarket, the accounting of a bank,  
including its grid of automatic tellers, as well as large  
technological systems in the telecommunications or transport sector,  
depend on the consistent interaction of databases, the recombination  
of data which is often stored in distributed data servers.

The search society and its culture of recombination had started to  
emerge even before Gilles Deleuze was able to coin a confusing term  
for its adaptive and distributed mode of operation. It is a society  
which operates in real-time mode, a culture which thinks of itself as  
a project with continuous change management, a society which has to be  
understood in terms of the permanent fluctuation of its conditions and  
relations, of its writing and reading, of its calculations and  
decisions, and indeed, of its practices of search and query.

---

Further reading including bibliographical notes

· David Gugerli: Die Welt als Datenbank. Zur Relation von  
Softwareentwicklung, Abfragetechnik und Deutungsautonomie, in:  
Gugerli, David et al. (eds.): Daten, (Nach Feierabend. Zürcher  
Jahrbuch für Wissensgeschichte, Bd. 3), diaphanes: Zürich - Berlin,  
2007, pp. 11-36

· David Gugerli, Suchmaschinen. Die Welt als Datenbank. Suhrkamp  
Verlag, Frankfurt am Main, 2009



#  distributed via <nettime>: no commercial use without permission
#  <nettime>  is a moderated mailing list for net criticism,
#  collaborative text filtering and cultural politics of the nets
#  more info: http://mail.kein.org/mailman/listinfo/nettime-l
#  archive: http://www.nettime.org contact: nettime@kein.org