Digitizing Documents and Information
Uppladdad av Wout Dillen
vyer
kommentarer
Relaterad media
In this class in the Information Retrieval 2 course, we take a step back from the topic of machine learning and information retrieval algorithms, and focus instead on how physical objects (and the data they possibly contain) are digitised, and what effect this process of digitisation can have on the information we are ultimately able to retrieve.
The live lecture started with a quick introduction to digitisation itself, focusing on how textual and visual information can be translated into a digital medium, and stored on a computer in a binary format. Since this was a recap of a lecture Wout has given in the context of a different elective course in the program (Technologies for Digital Libraries 1), this part of the lecture was not recorded again. Instead, wel refer you to the other video lecture below.
After studying the very basics of digitisation, the second part of the lecture emphasised that we should always be critical when we digitise documents, by focussing on what can go wrong when we rely too heavily on algorithms. This issue was illustrated by referring to the case of the Xerox scandal, where it was discovered that a software bug caused certain models of Xerox scanners to seamlessly replace numbers in the original document with different numbers in their copies -- meaning that the digital copy unintentionally contained information that was different from its physical original.
In the third part of this lecture, then, we were reminded that the digitisation process itself is not always as straightforward as we might think, and that it leaves a lot of room for agency, interpretation, and subjectivity. Focussing on the field of 'imaging' (reproducing an object's form, physical, and graphical features by creating an image of the object), this part of the lecture was divided into two case studies -- each detailing some of Wout's findings from his research into digitisation practices at a selection of European national libraries as part of the DiXiT project. By first looking into the different digitisation services offered by the National Library of Sweden (case study 1), and then looking at how such services are navigated and put into practice for a specific project (case study 2), this part of the lecture argues that the quality of a digital reproduction is always the result of a process of negotiation between a number of different agents, all of whom place their own demands on the original and/or its digital reproduction. As such, it is argued that we should never take the quality of a digital reproduction for granted (even when it is digitised by a trusted organisation such as a national library), and that we would instead benefit from a better and more nuanced understanding of the workflows, decisions, and negotiations that influenced the digitisation process.
Finally, in the fourth part of this lecture, we concluded by focussing on the 'Digitisation as Preservation fallacy'. Building on some of the materials from the earlier parts of this lecture, it was argued that digitisation never leads to the 1:1 reproduction of a physical document, but that, instead, digital reproductions are always reductive, biased, and finite. As a result, simple digitisation can never suffice as a comprehensive digitisation strategy, and a digital reproduction should always be regarded as a complement to the original, and never as a replacement of that original.
This lecture mentions the following video lectures:
The live lecture started with a quick introduction to digitisation itself, focusing on how textual and visual information can be translated into a digital medium, and stored on a computer in a binary format. Since this was a recap of a lecture Wout has given in the context of a different elective course in the program (Technologies for Digital Libraries 1), this part of the lecture was not recorded again. Instead, wel refer you to the other video lecture below.
After studying the very basics of digitisation, the second part of the lecture emphasised that we should always be critical when we digitise documents, by focussing on what can go wrong when we rely too heavily on algorithms. This issue was illustrated by referring to the case of the Xerox scandal, where it was discovered that a software bug caused certain models of Xerox scanners to seamlessly replace numbers in the original document with different numbers in their copies -- meaning that the digital copy unintentionally contained information that was different from its physical original.
In the third part of this lecture, then, we were reminded that the digitisation process itself is not always as straightforward as we might think, and that it leaves a lot of room for agency, interpretation, and subjectivity. Focussing on the field of 'imaging' (reproducing an object's form, physical, and graphical features by creating an image of the object), this part of the lecture was divided into two case studies -- each detailing some of Wout's findings from his research into digitisation practices at a selection of European national libraries as part of the DiXiT project. By first looking into the different digitisation services offered by the National Library of Sweden (case study 1), and then looking at how such services are navigated and put into practice for a specific project (case study 2), this part of the lecture argues that the quality of a digital reproduction is always the result of a process of negotiation between a number of different agents, all of whom place their own demands on the original and/or its digital reproduction. As such, it is argued that we should never take the quality of a digital reproduction for granted (even when it is digitised by a trusted organisation such as a national library), and that we would instead benefit from a better and more nuanced understanding of the workflows, decisions, and negotiations that influenced the digitisation process.
Finally, in the fourth part of this lecture, we concluded by focussing on the 'Digitisation as Preservation fallacy'. Building on some of the materials from the earlier parts of this lecture, it was argued that digitisation never leads to the 1:1 reproduction of a physical document, but that, instead, digital reproductions are always reductive, biased, and finite. As a result, simple digitisation can never suffice as a comprehensive digitisation strategy, and a digital reproduction should always be regarded as a complement to the original, and never as a replacement of that original.
This lecture mentions the following video lectures:
- What is a Digital Scholarly Edition? (Digitising Cultural Heritage Materials)
- Data Formats and Technological Foundations for the Web (Technologies for Digital Libraries 1)
These slides were originally designed by Wout Dillen, as part of a class on ‘Digitizing Documents and Information’, that was part of the ‘Information Retrieval 2’ course in the University of Borås’ International MA programme on ‘Digital Library and Information Services’ at the Swedish School for Library and Information Science. This class took place on Thursday 8 November 2022.Wout currently works as a Senior Lecturer at the University of Borås.
This work is licensed under a Creative Commons Attribution 4.0 International Public License. All works of other authors cited, linked, and referred to here are their intellectual property and are used for academic purposes only.
- Taggar
-