PDF/A – TOWARDS A TRUE DIGITAL

ARCHIVAL SURROGATE (DAS) FOR

DIGITAL MANUSCRIPT COLLECTIONS

Rodney Obien and Jeffrey Monseau

Wallace E. Mason Library, Keene State College, 229 Main Street, Keene, New Hampshire, U.S.A.

College Archives, Springfield College, 263 Alden Street, Springfield, Massachusetts, U.S.A.

Keywords: Digital Surrogates, Digital Libraries, Digital Preservation, Manuscript Collections, PDF/A.

Abstract: Digital surrogates provide a non-invasive means to study old manuscript documents that are often too

fragile and valuable for wide public access. These surrogates are generated from web-accessible derivatives

made from high-resolution archival masters; these masters serve as long-term digital preservation copies.

What if there was a file format that combined the functions of digital surrogate, web-accessible derivate,

and archival master? This paper considers the notion of the archival file format PDF/A (ISO: 19005-1) as

digital archival surrogate or DAS that combines the functions of surrogate, derivative, and master. The

paper discusses, furthermore, the versatility of PDF/A in dealing with the complex nature of old

manuscripts, and the possible implications of adapting PDF/A as a DAS standard.

1 INTRODUCTION

Digital surrogacy unlocks many options for studying

rare and fragile cultural artifacts. Often, surrogates

provide the only feasible access to the originals.

Surrogates allow for the “virtual viewing” of an

artifact -- enabling detailed analysis (Nichols, 2007).

They also allow cultural institutions to virtually

bring together materials, “re-construct wholes from

parts and make resources both visible and

convenient for researchers to use” (Butler, 2009).

Nowhere are digital surrogates more valuable

than in the study of old manuscripts. Manuscripts,

like medieval illuminated text, musical scores,

letters, and diaries, offer researchers and scholars a

rich source of historical documentary material.

These handwritten documents are complex, multi-

page objects, bound and unbound, and often too

fragile and valuable for wide public access. Digital

surrogates afford a non-invasive means of providing

access to these unique complex materials.

Several digitization projects have focused on

creating high-quality digital surrogates for

manuscript collections. The British Library

(http://www.bl.uk/onlinegallery/ttp/ttpbooks.html),

for example, is making its manuscript collections

available through the “elegant” Turning the Pages

system, developed by the Armadillo Corporation.

The Internet Archive project

(http://www.archive.org) is committed to digitizing

manuscripts, making them accessible through

document viewers like Djvu (AT&T) and PDF

(Adobe).

The standard process of creating a surrogate is

by first digitizing the “original” manuscript and

creating archival masters, such as TIFF or

JPEG2000 files, and then high-quality web-

deliverable derivates, whether JPEGs, GIFFs, or

PDFs, are made from the masters which then

constitutes the surrogate. This process also follows

the standard two-tier digital preservation model of

master and derivate practiced in the library and

archival professions. Creating, storing and retrieving

these derivatives, though, is a complex and

cumbersome process that often requires expansive

resources and a high level of expertise that can be

daunting to small or large institutions a like.

What if there was a way to simplify this process?

What if there was a means to combine the functions

of an archival master, derivative, and digital

surrogate? What if there was only need for one file

format. And what if this file format existed already

as a digital archival standard, recognized by the

International Standards Organization (ISO)? Such a

file format does indeed exist in PDF/A.

The PDF/A file Standard (ISO: 19005-1) could

234

Obien R. and Monseau J.

PDF/A â

A ¸S TOWARDS A TRUE DIGITAL ARCHIVAL SURROGATE (DAS) FOR DIGITAL MANUSCRIPT COLLECTIONS.

DOI: 10.5220/0002843702340239

In Proceedings of the 6th International Conference on Web Information Systems and Technology (WEBIST 2010), page

ISBN: 978-989-674-025-2

be the solution to simplifying this multi-tiered

process. PDF/A has the potential for electronically

preserving manuscripts and providing high-

resolution, web-deliverable digital replicas.

Effectively, PDF/A could be used as a “true” digital

archival surrogate (DAS) for long-term digital

preservation and high-quality access to digital

manuscript collections.

The PDF/A file standard (ISO: 19005-1) was

developed to archive and preserve Portable

Document Files (PDFs) and digitally-born

documents, such as word-processing and publication

files. PDF/A has grown quickly into an industry

standard for preserving textual documents and is

used in many companies, libraries and archives.

PDF/A is a versatile format. Like its cousin PDF,

PDF/A can preserve the structure, layout, and visual

appearance of a document, has the ability to embed

metadata, and is designed to work with multi-page

documents. The difference is that, unlike normal

PDF, it has the added benefit of backward

compatibility and is completely self-contained.

These attributes gives PDF/A the capacity to deal

with the complex nature of manuscripts, and hence,

the potentially serve as a true digital archival

surrogate.

This paper considers the notion of PDF/A as a

digital archival surrogate (DAS) for digital

manuscript collections. The paper will discuss (1)

complexities associated with old manuscripts; (2)

provide an overview of the PDF/A standard; and (3)

discuss the adaptation of PDF/A as a DAS and its

implications.

2 MANUSCRIPTS – COMPLEX

OBJECTS

Manuscripts vary in form and appearance. They can

include letters, diaries, illuminated text, musical

scores, scientific notations, financial ledgers, or legal

documents, such as this 1727 “Oath to King George

II of England” (Figure 1). They can be bound,

unbound, or rolled like a scroll. However, what they

share in common is they are generally handwritten,

manually produced, and page-oriented documents.

When digitizing manuscripts, one must consider

their inherent complexities.

First, the value in a manuscript is not only

informational but also visual. Manuscripts are

considered textual documents, since their purpose is

to be read. Yet, because of their handwritten nature,

they can be classified equally as visual (see Figure

1). Therefore, manuscripts hold a “dual nature”; they

are both “purely textual” and “purely graphical”

(Nicolas, 2009).

Second, manuscripts require a means to preserve

the context and meaning of a document. This

requires a means of maintaining a visual relationship

between pages. This is no different than a book or a

newspaper. Context and meaning is lost if the

sequential continuity is not maintained between the

pages. An orchestra, for example, will be lost if a

page in a musical score is misplaced or missing.

Figure 1: Handwritten Oath to King George II of England

(1727) with multiple signatures. The figure represents a

four-page document. It is simultaneously textual and

visual (courtesy of Keene State College).

Figure 2: Detail of the Oath to King George II (1727)

(courtesy of Keene State College).

Third, researchers need not only to be able to

read the manuscript but also have a means to closely

examine the handwriting, notations, and other

aesthetic and tactile features of the manuscript. In

effect, the manuscript is not merely viewed as text

but also artifact. It is not sufficient to solely provide

PDF/A - TOWARDS A TRUE DIGITAL ARCHIVAL SURROGATE (DAS) FOR DIGITAL MANUSCRIPT

COLLECTIONS

235

a transcription of the text. The visual appearance of

the manuscript is imbued with meaning (see Figure

2). For example, the researcher may want to study

the quality of the signature to determine attributions

or authenticity.

Fourth, non-linear navigation between pages is

essential. We assume that documents are read

linearly, left to right (or vice versa), and page-by-

page. Consider if the document’s author intends for

the reader to “jump” forward or backward in a

manuscript. A common example is a coda or repeat

symbol in a musical score, directing a musician to

move forward or backward in a piece.

These complexities necessitate different

techniques to allow a digital surrogate to honestly

portray a manuscript. Text must be able to be read,

but the quality of that text must be able to be

evaluated. This requires high resolution and the

ability to effectively view these documents. Page

sequence must be able to be maintained while being

able to be navigated in non-linear ways. This

requires complex methods to ensure that the

document can be navigated.

3 PDF/A – AN OVERVIEW

PDF/A-1 or part 1 was approved as an open standard

in May 2005 and published by ISO in September

2005. Developed by the working group ISO/TC 171

SC2, Document Imaging Application, Application

issues, for which Association for Information and

Image Management (AIIM), or The Enterprise

Content Management Association as it has been

renamed, acts as secretariat, the official standard is

listed as “ISO 19005-1:2005. Document

management – Electronic document file format for

long-term preservation – Part 1: Use of PDF 1.4

(PDF/A-1)” (pdf-tools, 3).

PDF/A is based on PDF Reference, Third

Edition, version 1.4. This means its files can be read

with Adobe Reader Version 5.0 or higher. Not all

the parameters outlined in version 1.4 are used in the

PDF/A standard. Some things have been excluded,

such as transparency, sound, and movie actions.

The purpose of the exclusion is to make sure that

the digital material can be preserved. Most of the

exclusions are simply unable to be included due to

their nature. The use of multi-media: Audio and

Video or JavaScript or other executable files that are

external are examples of this. Anything that lies

outside of the document and is dependent on

existing technological standards would limit the

preservation life of the document. Encryption is also

prohibited since it depends on a key that may be lost

or forgotten. LZW Compression (a lossless data

compression) is excluded because it is proprietary

and therefore dependent on the whims of a company.

PDF/A-1 also mandates that the file follow

certain rules. Again, these are to ensure that

preservation of the digital material is ensured.

Foremost, all fonts are to be embedded within the

document, and that they must be legally (some fonts

are not in the common domain) able to be embedded

for universal rendering. Color spaces are also to be

specified in a device-independent manner and

PDF/A requires the use of Metadata about the

document structure, creation and provenance. This

includes basic metadata entries like title, author,

creation date, modification date, subject, etc. For this

purpose PDF/A uses Extensible Metadata Platform

(XMP), though other metadata standards such as

Dublin Core can be used if they are embedded

within the XMP framework. In fact, Dublin Core

elements are automatically created in Acrobat,

Adobe’s PDF software, when a PDF is made. XMP

is a subset of Resource Description Framework

(RFD), the language created by the World Wide

Web Consortium for the development of the

semantic web. The metadata can then be used by

content management systems to tie documents

together and to provide search capabilities.

3.1 PDF/A-1a & 1b

The PDF/A standard comes in two different

compliance types, PDF/A-1a & 1b. PDF/A –1a

adheres to all requirements as outlined in the

standard. It is also known as a “tagged” PDF/A. It

ensures preservation of a document’s logical

structure and content in natural reading order.

Advantage to this level of compliance is that it can

be read with basic text editing tools such as MS

Notepad. PDF/A – 1b is for minimal compliance to

the standard. It ensures that the text can be correctly

displayed, but does not guarantee that extracted text

will be legible or comprehensible.

3.2 PDF/A-2

The development of the PDF/A standard is not done.

There is a new version of the PDF/A standard under

development based on PDF Reference versions 1.5 –

1.7. It will be called PDF/A-2. All PDF/A-1

documents will be compatible with PDF/A-2, though

PDF/A-2 documents may not be completely

compatible with PDF/A-1 due to the added features.

PDF/A-2 plans on the inclusion of JPEG 2000 image

compression, more sophisticated digital signature

WEBIST 2010 - 6th International Conference on Web Information Systems and Technologies

236

support, OpenType fonts, 3D graphics, audio/video

content, and consistency with other PDF file formats

like PDF/X (printing), PDF/E (engineering) and

PDF/UA (universal accessibility). PDF/A-2 is set to

be released in late 2009 or early 2010 (Fluckinger).

4 PDF/A = DAS

A digital archival surrogate (or DAS) is a hybrid of

archival master and high-quality digital surrogate.

PDF/A has the potential of acting as a true DAS that

can deal with the inherent complexities of

manuscript material. PDF/A can preserve the

structure, layout, and visual appearance of a

manuscript. The manuscript can be read as a textual

document or viewed as a visual object.

PDF/A, via Adobe Acrobat Reader, provides for

a variety of ways to view the document – single,

multi-page, and/or side-by-side. This eliminates or

greatly reduces any potential loss of context or

meaning.

The Acrobat Reader browser allows for non-

linear navigation. The Reader makes it simple to

page forward or back, or jump to a selected page.

The thumbnail function also allows for a simple

means to navigate quickly between several pages.

The magnification function of Acrobat Reader

allows for close, detailed examination of the

document. The high-resolution of the images offer a

high-quality replication.

PDF/A offers a means to embed documentary

information regarding the original manuscript using

Extensible Metadata Platform (XMP) functionality.

Essentially any metadata standard, such as Dublin-

Core, could be used to create a record containing

information on the creator, date, description, format,

identifier, language, publisher, relation, rights,

source, subject, and title or any other data that was

needed.

4.1 Concerns Regarding Adaptation

The potential obstacles to PDF/A acceptance as a

DAS are file size and efficiency as a web-

deliverable surrogate, and questions of conversion

and migration. Another concern, albeit minor, is the

limitations caused by the complexities of manuscript

material in what sub-class of PDF/A can be used for

DAS.

4.1.1 Web Deliverability

PDF/A files are web-deliverable and can be read

with Adobe Acrobat Reader or a plug-in version of

the reader for a web browser. Since PDF/A is

backward compatible, it provides an added

advantage if a user is without the latest version of

Adobe Acrobat Reader. The problem for the use of

PDF/A, though, is that some of the files can be quite

large, which can make it hard for their transfer over

the web.

With the proliferation of high-speed, broadband

Internet access, file size is less of an issue. Once a

100MB+ document would be a daunting task to

deliver via the Web, this, though, has changed over

the past few years. The substantial increases in

electronic transfer rates, sending megabytes per

second, have made large document delivery,

although not instantaneous, quite tolerable.

The next generation of PDF/A, PDF/A-2, will go

further to change attitudes towards web delivery

with its use of JPEG2000 technology. JPEG 2000 is

an ISO standard that has been published as ISO/IEC

15444. JPEG2000 allows for LZW (lossless)

compression of large, high-resolution images.

Compression will reduce file size significantly,

which in turn will exponentially increase the

efficiency of web delivery for PDF/A files.

4.1.2 Future Conversion & Migration

When adapting a new file standard, questions of

future conversion and migration need to be asked.

PDF/A is an ISO standard with a requirement for

backward compatibility. It will be readable and

accessible by future versions of Acrobat Reader or

other PDF viewers.

PDF/A files can be converted to other formats,

like TIFF, using commercial software programs like

Adobe Acrobat or freeware such as MyMorph

[http://docmorph.nlm.nih.gov/docmorph/] developed

by the United States National Library of Medicine.

Pages from a multi-page PDF/A can also be

abstracted and converted using available software.

Future conversion and migration of PDF/A, if

needed, will not be an issue as long as PDF/A

continues to be an ISO standard, and as long as long

as PDF technology continues to be an industry-wide

standard.

4.1.3 Use of PDF/A-1b as DAS

The strong visual nature of manuscript material

would make the sub-class PDF/A-1b the only

current option for DAS since text attraction (via

OCR) would be made difficult due to the

handwritten characters. Handwritten text, currently,

is very difficult to convert using any standard OCR

capture program. PDF/A-1b, of course, meets all the

minimal but important standards for the format.

PDF/A - TOWARDS A TRUE DIGITAL ARCHIVAL SURROGATE (DAS) FOR DIGITAL MANUSCRIPT

COLLECTIONS

237

4.2 New Digital Preservation Model

The adaptation of PDF/A as a DAS standard has the

potential of shifting the current digital preservation

model. The current model, as espoused by many

leading institutions, calls for a two-tier system. The

top tier directs for the creation of archival masters.

The widely accepted standard is TIFF or JPEG2000,

and some instances for PDF/A for text-based

documents. Reference versions, the lower tier,

include JPEG or GIFF for visual material, plain text

in ASCII or UTF-8 character encoding, XML, PDF,

or PDF/A for text or textual documents (CSU

Libraries).

Tier 1

Preservation

Tier 2

Web Derivate

Visual Materials TIFF

JPEG2000

JPEG

GIFF

Textual

Documents

PDF/A PDF

PDF/A

TXT

Figure 3: Two-Tier Digital Preservation Model.

PDF/A can eliminate the two-tier system by

creating a single tier. PDF/A, as DAS, serves as

both preservation and derivative. This has great

implications for file management requirements and

would eliminate complicated file-naming

conventions.

A single manuscript document may consist of

several pages. Each page requires an individual

archival master and a corresponding web-access

derivative. Each file will also require a unique file

name and a means to document the structure of the

original manuscript.

Consider a manuscript collection of 1000 letters.

Each letter consists of four pages. Each manuscript

has a unique filename, individual file folders with

separate sub-folders for archival masters and

derivates. The collection would generate 4000

individual archival masters and 4000 derivates for a

total of 8000 individual files. There would be a total

of 1000 folders and 2000 sub-folders.

The DAS model would create a single four-page

file. Each DAS would have one unique file name

and no folder. Each DAS would contain descriptive

data, unlike the previous model.

The model changes even more if PDF/A-2 is

adapted. Not only will there be a reduction in file

management requirements, one would also see a

significant reduction in file storage requirements if

compression is brought into the equation. This

could have major impact on storage cost. JPEG2000

technology has a 1:50 compression ration. For

example, in our 1000 manuscript digital library, each

DAS averages 160MB in file size. The entire library

would total 160GB. The use PDF/A-2 with LZW

compression would reduce the library to 80GB. That

is a significant reduction in storage and a potential

savings in storage cost.

Figure 4: Two-Tier System.

Figure 5: DAS Single-Tier System.

5 CONCLUSIONS

PDF/A has the promise of serving as the standard for

a digital archival surrogate (DAS) for old manuscript

collections. It has the versatility to deal with the

complexities of manuscript material, allowing for

navigation within in a manuscript while easily

maintaining page integrity. In addition, it has the

WEBIST 2010 - 6th International Conference on Web Information Systems and Technologies

238

added advantage of allowing researchers to being to

drill down into the image and study the manuscript’s

meaning as an artifact. PDF/A is able to accomplish

this without the expertise and expensive software

that causes large and small institutions to question

the worth of digitizing manuscript material.

The next stage of research requires testing of

PDF/A as a DAS through case studies using

manuscript materials. The materials should consist

of a wide-berth of document types, including, but

not limited to, illuminated text, musical scores,

handwritten letters, and personal diaries. Section 4

of this paper offers a useful model for PDF/A to be

tested and compared against current digital

preservation and access standards for efficiency and

effectiveness in file management, organization, and

storage, web deliverability, and usability,

particularly in regards to researchers and scholars.

These tests should help verify PDF/A’s viability

as a true digital archival surrogate. Adaptation of

PDF/A as DAS would have a major impact on

digital preservation policies by simplifying the

digital preservation model and creating a new

standard that would ensure the tong-term

preservation and digital access to the world’s

precious manuscript collections.

REFERENCES

Butler, S 2009. “Exploiting special collections: using

digital methods to enhance their research and learning

potential.” SCONFUL Focus, 45, p. 81-86.

Colorado State University Libraries 2009. CSU Digital

Repository Preservation Format Support Policy.

Colorado State University Libraries, viewed 21

October, 2009, <http://lib.colostate.edu/repository/

preservation. html>.

Fluckinger, D 2009. PDF/A Takes Root as Digital

Archiving Standard (4 April 2009), PDFzone, viewed,

21 October 2009, <http://www.pdfzone.com>.

Nichols, SG 2007. “An Artifact by Any Other Name:

Digital Surrogates of Medieval Manuscripts.” In

Archives, Documentation and Institutions of

Social Memory: Essays from the Sawyer Seminar.

edited by Francis X. Blouin, Jr. and William G.

Rosenberg. University of Michigan Press.

Nicolas, S, T Pacquet and L Huttle 2003. “Digitizing

Cultural Heritage Manuscripts: the Bovary Project.”

From ACM Symposium on Document Engineering,

p. 55-57.

PDF-Tools.com, 2007. “White Paper: PDF/A – The

Basics.” Understanding PDF White Papers, v. 2.0

(January 22, 2007): p. 1-9, pdf-tools.com, viewed 21

October, 2009, <http://www.pdf-

tools.com/public/downloads/whitepapers/whitepaper-

pdfa.pdf>.

PDF/A - TOWARDS A TRUE DIGITAL ARCHIVAL SURROGATE (DAS) FOR DIGITAL MANUSCRIPT

COLLECTIONS

239