|
Good Practices Handbook
3 Practical Guidelines
3.1 Introduction
This chapter presents the core of the handbook. It provides practical
guidelines for organisations and bodies contemplating, or involved in,
digitisation projects. The emphasis is on the cultural sphere; however,
the material is to a large degree relevant to other spheres (e.g. tourism,
general document management).
The material in this chapter is broken down in accordance with the stages
in the digitisation life-cycle. This means that a reader can easily identify
material which is relevant to his work, regardless of how far his own
project has progressed. It is anticipated that many users of this handbook
will be at the first stage of the project (planning); however, at least
some of the material provided here should be of value to any digitisation
project.
The digitisation life-cycle stages are used as the basis for the breaking
down of the guidelines and mirrored in the table of contents.
Each guideline description is structured in the following way:
- A Guideline Title
- An Issue Definition, which sets the scene for the guideline and/or
introduces the problem which the guidelines addresses
- A set of Pragmatic Suggestions which aim to facilitate the relevant
aspect of setting up or executing a digitisation project
- Notes or Commentary, where any additional information is provided.
This is sometimes empty
Supporting references are available on the Internet Web site:
http://www.minervaeurope.org
Neither the guidelines nor the references are exhaustive – however
they provide the most important information needed by a project which
is addressing a particular task or tasks within the life-cycle of a digitisation
project.
3.2 Digitisation Project Planning
Planning of the project is the first step in any digitisation project.
Time spent on planning the project will pay dividends in the easier management
and execution of the project. Normally, the following questions have to
be answered:
- What (work needs to be done)?
- Who (should do it)?
- Where (should it take place)?
- When (will it take place)?
- How (will it be done)?
A digitisation project should have clearly specified goals and objectives
– these will impact directly on the selection, copyright and publication.
The project should have suitable personnel, with appropriate knowledge
and skills, as well as a training plan in place to provide any additional
expertise that the project may require.
A project should not begin until some research has been carried out into
other projects in the same area. Such research will identify issues which
need to be addressed, will stimulate new ideas and areas which might not
yet have been considered, and will add value and credibility to the project
output.
Research will also help to indicate the amount of work which may be planned
for the execution of the project, by meeting or talking with organisations
which have completed similar projects. Such interactions will help to
establish whether your organisation has the personnel, the skills and
the technology infrastructure to carry out the project, or whether significant
training and preparation will be required.
Some time may profitably be invested in ascertaining the copyright status
of the material which is to be digitised. Failure to secure permission
to digitise and to publish on the Web can cause the failure of a digitisation
project, despite any technical expertise and experience.
A technical pilot may also be considered, at the start of the project,
in order to ensure that any anomalies or problems with the technical workflow
are resolved before commencing the main project.
The Reasons for the Project
Issue Definition
Each digitisation project has its own reason for being executed. Often,
the reasons involve providing access over the Internet to cultural holdings
which would otherwise be underused, or protecting fragile holdings from
the wear and tear of hands-on access. In other cases, the projects are
exercises in inter-body cooperation, and involve the establishment of
portals, networks, etc.
The reasons for the project will have a profound effect on the criteria
for selecting the material to be digitised. They will also affect the
project management, the meta-data, the online publication (if any) of
the project output, the quality control etc. ‘Why’ is the
most important question to raise before starting a digitisation project.
Pragmatic Suggestions
- The project must have concrete, explicit aims, and these aims must
be documented.
- The aims of the project should be realistic, when compared with the
resources available.
- All steps of the project should be validated against these aims, in
order to ensure that work carried out in the project contributes towards
the achievement of the guidelines.
- The project aims should document the value which the project will
bring to the institutions involved in the project. If time and effort
are to be invested in the project, the justification for the project,
from an institutional point of view, must be clear.
Human Resources
Issue Definition
Before a project can start, it is important that the personnel required
to work on the project is available. Many cultural bodies do not have
large corps of staff who have a great deal of free time to carry out digitisation
projects, over and above their usual duties. Also, the knowledge required
for digitisation projects may be different from the skills necessary to
carry out the tasks for the daily operation. Hardware and software solutions
required for a digitisation project, therefore, need to be identified.
Pragmatic Suggestions
- Ensure that sufficient staff is available to carry out the project.
- Assign staff to each task or work-package of the project plan.
- Identify training requirements, including information technology training
and education in the handling of delicate artefacts and documents.
- Carry out, if possible, training by using the hardware and software
solution which will be used during the project, before the project commences
(vendors sometime offer technical solutions free of charge for training
or relevant equipment can be hired short-time).
- Aim for a small core of skilled staff dedicated to the project, rather
than a large group of ‘occasional’ staff.
Notes/Commentary
Even if the material presented in this guideline is common to all project
management scenarios, it is worth to reperat this matter: in fact there
is possible risk to irreplaceable artefacts and documents if the resourcing
is not properly handled.
Research
Issue Definition
Regardless of the scope of any particular project, it can be assumed that
similar projects have been carried out in the past. There is a strong
likelihood that information about such projects will be available on the
Internet, or else published in appropriate journals, etc.
Researching the area as part of the project planning process can help
to identify candidate hardware and software solutions, to plan workflow
and process, and to avoid issues and obstacles which have been experienced
by other projects.
Pragmatic Suggestions
- As early as possible in the planning process, carry out research into
any other projects which are addressing similar issues to the project
being planned. This handbook provides a starting point; however the
amount of material available on the Internet is the largest and most
comprehensive resource.
- Research helps avoiding mistakes. It can also put the project team
in contact with others who have completed similar projects, and give
the opportunity to learn from their experiences.
- Having carried out research adds credibility and value to the output
of any project. Assurance that your project has not been carried out
in a vacuum, by taking into account the work of others, enhances the
results of your project.
Notes/Commentary
Many cultural digitisation projects are funded with public funds, and
have a requirement to publish their findings and their reports. Such publication
is almost always on the Internet, as well as using other appropriate media.
Project teams are usually very happy to share their experiences and their
results – this adds value to their work.
Risks
Issue Definition
At the start of any project, plans have to be made to guarantee a successful
outcome. However, the goal is not to eliminate all risks but to prepare
for them by creating a project framework which responds to the unforeseen
in a resourceful and effective way. The aim is to create a project with
staff and procedures that can accommodate changes. Therefore, all project
planning need to have a risk analysis.
Pragmatic Suggestions
- Distribution of digitised images over the Internet is a form of publication,
and is by this reason covered by laws of copyright and Intellectual
Property Right (IPR). Examples of questions in a risk analysis are:
- What could be the consequences of using the material without specific
permission?
- Has attempts been made to find the rights holder?
- If copyright infringement does occur, what would the impact be
on the project?
- For public information the legal value of the information is an important
issue. What steps have been made to guarantee that a digitised source
material is not corrupt and has actually been produced by an authorised
institution?
- The authenticity must also be guaranteed. What actions have be undertaken
to maintain the image files, and what tools have been used?
- Financing the project could be a problem and, therefore, a potential
risk for the possibilities to reach the goals of the project.
- A key question is the level of skill in the project. Is it possible
to hire new highly skilled and experienced persons? If not, will it
have an effect on the work plan of the project?
3.3 Selecting Source Material for Digitisation
The selection of the material to be digitised is an important decision
for any digitisation project. Typically, the ideal choice is to digitise
all the material in a collection or holding; however, this is rarely feasible,
so choices must be made. The criteria for selection will differ, depending
on the goals of the digitisation project; an online resource for schools
may choose to digitise material in line with a syllabus, while a museum
may digitise its best-known holdings in order to stimulate visitor numbers
or its most fragile artifacts in order to minimize demand for ‘hands-on’
examination. These are of course not the only issues to be addressed in
the selection criteria – the reasons for choosing to digitise particular
material will vary from project to proposal, as will the reasons for deciding
not to digitise. Examples of other reasons include legal constraints,
institutional policies, technical difficulty of digitisation, already-existant
digital copy, etc.
Establish Selection Criteria
Issue Definition
When planning a digitisation project, the choice of which material to
digitise is critical. The criteria for selection will depend on the goals
of the project, as well as on technical and financial constraints, copyright
and IPR issues, and the activity of other projects in the area.
Pragmatic Suggestions
- It is essential to establish criteria for the selection of material
to be digitised. The selection criteria must reflect the goals of the
overall project. At least the following criteria may be considered:
- Access to material which would otherwise be unavailable, or of
limited availability;
- Wider and easier access to very popular material;
- Condition of the originals;
- Preservation of delicate originals, by making digital versions
available as an alternative
- Project theme
- Copyright and IPR
- Availability of existing digital versions
- Cost of digitisation
- Appropriateness of the source material for online viewing
- The criteria for selection should be explicit and discussed with,
and endorsed by, all relevant stakeholders, prior to selection or digitisation.
- The selection criteria should be fully documented, so that the reasons
for any decisions to digitise or not to digitise are clear throughout
the project.
Notes/Commentary
Most commonly, cultural bodies have a core of high-value, high-user-interest
material which is, by default, included in any digitisation project which
is meant to represent the institution.
A large proportion of all digitisation projects have online Web publication
as a goal. This means that the copyright and IPR issues which surround
any material which may be digitised must be considered before selection.
Selection against the Criteria
Issue Definition
Having established the criteria against which material is selected to
be digitised, the actual selection process can take place. This guide
suggests how to manage this process.
Pragmatic Suggestions
- Each candidate for digitisation must be evaluated against the selection
criteria. In case that any selection criterion is not met, this should
be noted. In the event that this results in the rejection of important
or critical objects, it may be necessary to review the selection criteria.
Should this occur, the new criteria should be noted.
- Once an object has been selected for digitisation, its details should
be entered into the digitisation management knowledge base (see chapter
Managing Digitisation Projects).
Notes/Commentary
At this stage, the project is engaging with each of the items to be digitised,
for the first time. This is the optimum opportunity for the project to
create a knowledge base of all the items in the scope of the project.
Having such a knowledge base will support the management of the project,
and help to ensure that, for example, the appropriate expert knowledge
is acquired for handling rare artifacts, as well as more mundane issues
such as the location of originals.
3.4 Preparation for Digitisation
An appropriate environment and hardware/software system must be in place
before digitisation can begin. The elements of such an environment include
hardware for the digitisation process itself (e.g. scanners, digital cameras,
copy stands, other hardware), a computing infrastructure to which the
hardware is connected, software for image capture and processing software,
as well as software for meta-data and quality control. The working environment
should be appropriate to the material being digitised, paying special
attention, for example, to light, humidity, vibration, disturbance, movement
of the originals, etc.
Hardware
Issue Definition
The appropriate technical equipment must be in place for the digitisation
to go ahead. Typically this will consist of digital image capture equipment
(digital cameras, scanners for books, documents or microfilm, audio and
video hardware, if appropriate) connected to an appropriate computing
platform (computer, operative system, network, etc). Two different digitisation
methods, using different hardware, can be distinguished: scanning and
the use of digital cameras.
Pragmatic Suggestions
- Appropriate hardware must be installed and its quality and functionality
controlled before digitisation begins.
- Relevant test targets should be used for the evaluation of digital
image capture devices.
- No source material should be present until the hardware environment
has been fully established and tested with non-sensitive materials.
- Most digitisation projects will require a flatbed scanner, for material
which is not harmed by being pressed flat against a hard surface (e.g.
unbound printed material and manuscripts).
- The largest possible scanner should be acquired by the project. The
folding or mosaiced scanning of materials should be avoided. The project
should bear in mind that the transportation of large (e.g. A0) scanners
is not trivial.
- Usually, a flatbed scanner should only be used where the material
is already flat, and will not be damaged by being held against a hard,
flat surface. A scanner with a book cradle may be appropriate for many
bound articles, up to the appropriate size limits. Most digitisation
projects will require a digital camera, for capture of material which
cannot be flattened or held on a scanner book cradle.
- If a scanner is used, it should ideally be at least as large as the
item to be scanned.
- Image capture (by scanner or digital camera) should be carried out
at the highest reasonable resolution. This will often result in very
large master files; smaller files can be extracted from the master,
for purposes such as Web delivery. However, a higher-quality image can
never be derived from a lower-quality image.
- The definition of a ‘reasonable’ resolution will depend
on the nature of the material being scanned, and on the uses to which
the scanned image will be put. For example, if the scanned images are
only ever to be used as thumbnails, this can allow scanning at a low
resolution. Equally, the resolution must capture the most significant
details of the item – if scanning at a high resolution yields
no more information than at a lower resolution, the high resolution
scanning is difficult to justify.
- Image capture should create a file format which is loss-less, i.e.
not compressed. Typically, the Tagged Image File Format (TIFF) is used.
- If a digital camera is used, a project shall choose the most powerful
and flexible one which can be afforded. The limitations of the digitisation
hardware cannot be overcome by any subsequent processing. It should
be noted that ‘digital zoom’ does not provide a better quality
picture; it merely displays less pixels per unit of view. In order to
capture detail, three parameters are most important – the number
of pixels in the image, the bit-depth, and the quality of the optical
lens being used.
- It is important to have appropriate stands for holding material while
it is being digitised.
- A digital camera with a dedicated copy stand should be used. The camera
should be tripod-mounted, and have supplementary lighting, filters,
etc, as appropriate. Consultation with an experienced digital photographer
with a background in similar projects is advised, if at all possible,
before setting up the hardware environment.
- The photographic plane and the plane of the material being digitised
must be exactly parallel, if the image is not to be distorted.
- Appropriate lighting must be part of the photographic set-up when
using a digital camera; it is very rare for ambient light to be sufficient.
- Suitable filters should be used in order to reduce colour distortion.
- A computer with significant storage should be connected to the devices.
This computer should be backed up very regularly – this requirement
reflects the high costs in time, technology and possible wear on the
originals, of the digitisation process.
- If an item must be scanned in multiple parts, an overlap of several
centimetres should be provided, in order to ensure that there are no
gaps between the parts. The same settings, light, etc should be used
for all parts, in order to avoid any ‘patchwork’ effect.
Notes/Commentary
The hardware used is a major constraint on the quality of the end result
of any digitisation project. Unless the project is digitising only flat
materials which can be scanned without damage to bindings, frames or the
source material itself, the use of a digital camera will be important.
While an analogue camera can be used, and the slides or prints scanned,
the advantages in terms of time, effort and quality of a high-specification
digital camera are many.
If the project has a limited life-span, renting hardware may be appropriate.
Another alternative is the use of external agencies to carry out the digitisation
on behalf of cultural bodies involved in the project.
Software
Issue Definition
Having created a digital version of the object, the resulting file is
likely to require processing before it can be used. Colour may need correction;
extraneous detail may need to be cropped (removed) from the edges of the
image, etc. Also, the master files are typically very large, so a smaller
file in a compressed format will often be needed (e.g. as a thumbnail
image, or for Web delivery).
Pragmatic Suggestions
- When the scanner or the digital camera is turned on, a calibration
routine should start automatically.
- Suitable image processing software will be needed to utilise the master
files for whatever the purpose of the digitisation project may be. While
digitisation hardware will typically be provided with some software
included, this is usually not of sufficient power and flexibility for
many projects.
- The requirements on the software depend on the aims of the project.
It is worthwhile to note that, once the master files are not modified
in any way; various different types of software can be used to process
them. However, the cost in time and effort may be significant, and will
usually overshadow the cost of a more powerful software package.
- The project should acquire the most appropriate and powerful software
package which it can afford.
- As an absolute minimum, the software must be capable of:
- opening very large image files;
- modifying the resolution and the colour depth;
- saving multiple different versions, in different file sizes;
- selecting and copying a part of the image, and saving this as
another file;
- exporting images in different file formats, including the Web
standards JPEG and GIF.
Several free software packages provide this level of functionality;
however investing in a commercial product is likely to pay dividends
in time, effort, documentation and technical support.
- In the event that the digitisation project has an OCR component, the
choice of software is also critical. Any OCR exercise has a certain
amount of manual editing and correction; the manner in which this is
supported by the software product in use can have a significant effect
on the time and effort required by the project. Better OCR packages
may enable review and editing on a single screen, suggest possible corrections
for mis-read words, support the use of multiple text columns (e.g. newspaper
layout), etc.
The evaluation of multiple OCR packages is likely to be worthwhile,
if the project exceeds, for example, one person-year in size
Notes/Commentary
The right software will save a digitisation project a large amount of
time and effort. If the project is of significant duration (e.g. more
than two persons for more than six months), evaluation of several software
packages may be worthwhile, in order to establish the best match for the
requirements of the project.
Environment
Issue Definition
Many rare or delicate materials require a particular environment. It is
critical to any digitisation project that the digitisation process has
the least negative effect on the source materials. An appropriate digitisation
environment is important to many digitisation projects.
Pragmatic Suggestions
- The environment in which digitisation takes place is of considerable
importance.
- Expert opinions should be sought in order to ensure that all aspects
of handling of original material are addressed as well as possible.
These include the environment for digitisation.
- The area used for digitisation should be dedicated to the digitisation
project for the duration of the project. Excessive movement, rearrangement
etc. of the workspace can lead to damage, loss or other negative effects
on the source materials, as well as to loss of time by the project.
- If the source materials have particular requirements in terms of light,
humidity, etc., then these should be replicated as closely as possible
in the digitisation environment. For certain materials, such as leather
documents, a short-term increase in humidity may assist in relaxing
the materials prior to flattening for photography or scanning.
- In almost all cases, direct exposure to bright light (e.g. sunlight)
for extended periods is not recommended. Smoking, eating and drinking
in the vicinity of the items should of course not be permitted –
keep coffee away from the work area!
Notes/Commentary
Depending on the size and budget of the project, a dedicated digitisation
environment may not be feasible. However, the aims outlined here, to minimize
movement, disruption and handling of the materials, should be kept in
mind.
As with the handling of heritage material, no references should be taken
as a substitute for discussion with those whose responsibility includes
the care of the material.
3.5 Handling of Originals
This section considers how a digitisation project should treat the material
which is being digitised. In many cases, the source material is rare or
valuable; the negative effects of digitisation on the source material
must be minimized.
In every case, it must be emphasized that the specialist knowledge of
the individuals who are responsible for the source material on a day to
day basis will be valuable to the project team.
Moving and Manipulating Original Material
Issue Definition
In many cases, the material to be digitised is of particular sensitivity
or fragility. Replacing hands-on access with online access is often an
important reason for digitisation projects in the first place. It is critical
that any digitisation project takes steps to ensure that no damage is
done to the original material during the digitisation process. These steps
may range from the use of the correct hardware to the establishment of
a suitable micro-climate or the movement of the digitisation centre of
operations to the location of the material, rather than vice versa.
Pragmatic Suggestions
- Consult the person usually responsible for the source material, before
moving or handling it. Include any information on appropriate handling,
in the digitisation project knowledge base (see chapter Digitisation
Process Management)
- Be prepared to be flexible – an inconvenience to the digitisation
project can be overcome, while damage to a unique artifact may be irretrievable.
- If necessary, bring the digitisation equipment (e.g. digital camera)
to the source item, rather than transporting the item itself.
- Avoid unbinding of bounded books and records. Use instead of a flatbed
scanner a scanner with a book cradle or a digital camera.
- Always remove staples, paper clips, and other fasteners; they can
damage both the digitisation devices and the source material.
- Expert advice (e.g. from the curator of the item to be digitised)
should be sought before any handling of the original.
- This advice should be sought prior to digitisation, ideally at the
time that the article is selected for digitisation. The advice should
be recorded in the digitisation management knowledge base, and consulted
before movement or digitisation of the article. If necessary, the expert
should be briefed on the capabilities of each possible hardware solution.
Notes/Commentary
While much of this material is quite obvious, it is important to establish
and maintain a discipline while handling the source material.
3.6 The Digitisation Process
This chapter provides some practical guidelines for the actual digitisation
process. The technical solutions for digital capture can differ. Scanners,
digital cameras or software applications for optical character recognition
(OCR) are areas covered in some detail, as being most relevant to the
largest number of projects. The digitisation of transparent originals
like microfilm is not considered.
Using Scanners
Issue Definition
Flatbed scanners are a very common digitisation tool. The most common
A4 and A3 models are relatively cheap, require limited skills to use,
and can manage a fast throughput of material, once a workflow has been
established. Larger models (up to A0) of flatbed scanners and scanners
equipped with book cradles are very expensive and thus require long-term
projects/programs, high-volume digitisation, or oversized source material.
Pragmatic Suggestions
- Only scan material on a flatbed scanner which will not be damaged
by being pressed flat onto a hard surface. Consult the experts, if in
doubt.
- Ensure that the glass scanning plate is completely clean at all times.
This both leads to better image quality and also protects the source
material from soiling.
- If possible, scan only items which fit on the flatbed scanner or the
scanner equipped with a book cradle in one piece.
- If it is necessary to scan an item in multiple parts, ensure that
there is sufficient overlap to allow the image to be reassembled, during
post-processing (by using mosaicing software).
- Test the scanner, and its output, on non-sensitive material before
beginning to scan original source material. Train users with the same
non-sensitive material.
- Establish a file-naming convention for the files produced by the scanner,
for example by using the existing cataloguing system or giving them
meaningful names. The file name should allow mapping between the file
and the source item.
- In order to maximise the portability of files across computer platforms,
a file name with a maximum of eight characters, followed by an extension
of at most three characters, should be adhered to.
- Before establishing workflow or work-batching process, carry out some
end-to-end scanning and image processing, in order to ensure that the
end result of the workflow will be what is anticipated.
- Scan at the highest resolution and bit-depth that is feasible given
the reasons for the project, the limitations of the scanner, the conditions
for data storage, and the attributes of the source material.
- Scan with the maximum appropriate colour depth, given the same limitations.
- Back up the hard disk where the data is stored, on a daily basis.
- Quality control of the digital image and of meta-data is important
– at scanning time is the most convenient time to address any
quality issues. The following points may be borne in mind:
- Establish minimum resolution and colour parameters (mainly the
spatial resolution and the bit-depth) for groups of items to be
scanned.
- Examine the scanned output on screen, on paper and in any other
format that you expect it to be used for (e.g. on a mobile device).
- Ensure that the screens (monitors) being used are reliably calibrated.
Avoid having other material on and around the screen, which may
affect the perception of the item
- Master images must be created with visible scaled rulers, and
colour or greyscale images must include also a standardised colour/greyscale
reference target.
Notes/Commentary
Scanning is in itself a relatively simple operation. However, in order
to increase efficiency and minimize errors, having a workflow system in
place will be worthwhile.
Scanning of oversize items, or very high quality scanning, takes a significant
investment of time and effort per item. This can be reduced by using hardware
appropriate to the item (e.g. a larger scanner, a book cradle); in the
event that large hardware resources are not available, allow plenty of
time. Training on oversize or irregular materials should not be neglected.
Using Digital Cameras
Issue Definition
The use of digital cameras is becoming increasingly common in digitisation
projects. This reflects their flexibility in terms of being able to digitise
non-flat objects, such as bound books, folded or wrinkled manuscripts,
and 3D objects. However, a scanner equipped with a cradle is normally
to prefer when digitising bounded books and over-sized material like maps
and drawings.
Pragmatic Suggestions
- Consider renting a high-quality camera, if the scope of the project
is limited.
- Put up the digital camera on a motorized carriage on a column and
place the items to be digitised on a steady copy stand board with specially
tailored lights.
- Organise training from a specialist digital photographer – the
difference in quality between pictures taken by an amateur and the same
photos taken by a specialist can be striking.
- Ensure that backgrounds will show the item clearly.
- Avoid changing the light conditions between shots, and between images
of different parts or sides of an item – this can lead to erroneous
impressions of colour variation.
- Use apochromatic lenses and appropriate lens filters to combat colour
misregistration and image distortion.
Notes/Commentary
The increasing use of digital cameras in digitisation projects reflects
their availability as a mainstream consumer product, and the resulting
decrease in price. However, there remains a significant difference, in
both price and quality, between specialist digital cameras and mass-produced
low-end consumer products.
Software Applications for Optical Character Recognition (OCR)
Issue Definition
Many digitisation projects involve the digitisation of printed documents,
such as books and newspapers. This occurs most often (though not exclusively)
in tandem with the use of scanners. The use of OCR software is a popular
way to extract the information from such scanned information, and to open
opportunities for processing the information. OCR software recognises
the letters and numbers which make up the scanned image (bit mapped image
file), and exports them as ASCII text files, rather than as image files.
This enables searching, indexing, format conversion, and other data processing
operations to be carried out.
Pragmatic Suggestions
- Evaluate multiple OCR software offerings before selecting a particular
product. While OCR software is often included with the sale of a scanner,
more powerful software is typically sold separately.
- A major element of any OCR project is the identification and manual
editing of mistakes, ambiguities and locations where the text could
not be processed. An OCR package which provides a friendly user interface
for carrying out this task can save considerable time and effort.
- OCR works best with documents which are in good condition –
folding, wrinkling and discoloration of the source material will increase
the number of errors and faults in the OCR process. Pre-treatment, where
possible, of the source material should be carried out to avoid this.
- The use of image processing software, to remove discoloration and
improve contrast, before the use of OCR software, should be considered
for material which is not in perfect condition.
- The availability (or not) of dictionaries in the language of the source
material, as part of the OCR package, should be verified.
Notes/Commentary
Relevant products in this market include
- OmniPage
- TextBridge and
- Adobe Capture.
- Abby FineReader
The last of these has excellent editing and fault resolution functionality.
3.7 Preservation of Digital Master Material
In the longer term, it is an important goal of any digitisation project
to protect and keep accessible the data which it has created. This involves
dealing with the inevitable obsolescence of digital file formats and various
types of computer storage media.
Preserving the digital master material and corresponding meta-data helps
to avoid having to re-digitise any items, thus protecting the fragile
source material and avoiding repetition of the labour-intensive digitisation
process including generating meta-data.
File formats
Issue Definition
The digital output of the digitisation process is usually a master file
in uncompressed TIFF format with some meta-data embedded (see chapter
The Scope of Meta-data Used for Object). The file format as well as the
compression used will have a major impact on the usability of the digitisation
output. At this time, issues such as file format, standards file size,
network transmission time, and different kind of outputs (monitor or printer)
need to be taken into account.
Pragmatic Suggestions
- Before deciding on a file format, take into account the relevant standards,
the established global user base and the degree to which file formats
are supported by software in use by your organisation and your target
audience. The size of the global user base is a good indicator of the
future, ongoing, support for a particular file format. It also indicates
the likelihood of sustainable migration paths, when file formats change.
- The default digitisation output file for digital images is Tagged
Image File Format (TIFF). Unless your project has a clear, justified
reason for using some other file format, digitisation output, and so
master files, should use this format.
- The output file will typically be quite large. It is common to have
a large master file, which is stored locally but not transmitted over
the Internet. From this master file, smaller versions can be created
using image processing software, either in TIFF, or more commonly in
a delivery format such as JPEG 2000, PNG or GIF (see the section on
image standards, later in this document).
- More information on file formats is provided in the survey of standards
provided later in this document.
- Regardless of how attractive a proprietary or national format may
appear to be from a technical standpoint, it is important to bear in
mind that failure to use standard formats and media will act as a major
obstacle to international exchange of raster image files and corresponding
embedded meta-data as well as the creation of networked resources.
Notes/Commentary
File format choice must be governed by the imperative to create the highest
quality digitisation output, and by the availability of migration paths
for future preservation of the digital master. The role of standards in
this area is very great.
Media Choices
Issue Definition
The issue of media choice is an important one for projects which wish
to maintain their digital collections over a several-year period. Important
projects such as the UK Domes day book initiative have been lost due to
media obsolescence.
Pragmatic Suggestions
- The output of the digitisation project will be held on server machines,
including those which serve digital content to Internet users. However,
these machines need to be backed up. Also, if a server is not dedicated
to a digitisation project, the digital content should be stored on removable
media, separate to other data on the server.
- All master files (including meta-data) should normally be backuped
on two kinds of media separately stored.
- Currently (early 2003), the use of CD-Rs as a common backup medium
is in the process of being replaced by the use of DVDs. DVDs offer significantly
larger storage, and the hardware needed to read them is becoming ubiquitous
on new PCs and laptops. DVD writers remain more expensive, but are already
well within the means of all but the smallest projects.
- However, DVDs are not expected to replace magnetic tape media like
Digital Linear Tape (DLT) as the storage medium of choice for backup
of computer storage, in the near future. Both of these technologies
should be seriously considered as candidates for preservation of digital
content.
- Regardless of the choice of medium, it must be borne in mind that
the medium will become obsolete in near to mid-term future. Within five
years, migration to new storage media is likely to be a necessity.
Notes/Commentary
The rapid change of media layouts, driven primarily by the consumer electronics
industry, has had major effects on digitisation projects in the past.
However, the increasing trend to store data “on the Internet”
on large server machines, and as data on mobile hard drive units, facilitates
the migration of data from place to place and from medium to medium. Once
servers are backed up and migrated to new servers over time, the dependence
on removable media as the only record of a digitisation process can be
expected to decrease.
In the meantime, the issue of media selection is still an important one.
There is no indication that the limits of compressed, small-footprint
digital storage are being reached.
Migration Strategies
Issue Definition
As noted above, the choice of file format and storage medium must take
into account the feasibility of moving data to a new file format and/or
a different storage medium, in the foreseeable future.
Pragmatic Suggestions
- Examine the relevant standards for file formats and storage medium,
as noted in the previous two guidelines. Compliance with standards is
a reasonable indicator that a particular format or medium will have
some support into the future.
- Proprietary file formats and non-standard media formatting should
be adopted only with great care.
- Migration from one format to another should avoid migrating from a
lossless file format (e.g. TIFF in the image domain) to a lossy one
(e.g. JPEG), for master digital material. Once information is lost,
it cannot be replaced.
- Bear in mind that any choice of file format and/or storage medium
will become obsolete in the foreseeable future (possibly less than five
years, probably less than ten years).
- The size of the market for storage media provides an indication of
how likely it is that migration from one medium to a new one will be
feasible, as the medium becomes obsolete.
- Having created the digitised material, storage media (e.g. CD-R, DVD)
should be refreshed periodically (once every two to three years), to
combat data loss. This involves copying all data to new media.
- The status of digitised material, including when it was last refreshed,
should be recorded in an appropriate log.
3.8 Meta-data
The area of meta-data is one of the most actively researched and dynamic
in the whole digitisation area, as well as in areas such as information
retrieval, Web searching, data exchange, enterprise application integration,
etc.
The selected meta-data model is of particular importance as it influences
the choice of attributes to describe an object. Related to this is the
choice of a standard model, as it is described in the following chapter
on standards.
The Scope of Meta-data Used for Object Description
Issue Definition
Before selecting a meta-data model for a digitisation project, the material
to be described with the meta-data should be reviewed. This will help
to identify existing meta-data models, as well as to pinpoint any omissions
or gaps between what is covered by an existing meta-data model and the
important meta-data for your project.
Pragmatic Suggestions
- The use of appropriate meta-data is very important for enabling search
and retrieval of material from digital collections. This is even more
the case when searching across multiple collections, stored in different
locations, is the overall objective (logical union catalogues, virtual
combined museums, etc.).
- There exist already many meta-data models. Therefore, each project
has to choose as meta-data model based on its own goals. It is advisable
to avoid creating a new one, unless the requirements of your project
are badly underserved by all existing standards.
- Time spent modelling the important characteristics of the material
being digitised, and identifying its key attributes and descriptors
will be well invested. Such a model can then be compared with the scope
and features of existing meta-data models.
- Possible controlled vocabularies (e.g. to describe a location, or
an artist) should be identified. Several such vocabularies already exist
and can greatly increase the success of searches, etc. See the section
on meta-data standards and controlled vocabularies, below, for details.
Notes/Commentary
The Making of America II project (Library of Congress) used three categories
of meta-data:
- Descriptive – for description and identification of information;
- Structural – for navigation and presentation;
- Administrative – for management and processing.
Of particular importance are the meta-data models to be selected for
a digitisation project – the choice of which set of attributes
that will be used to characterize the works and items to be digitised,
the resulting images, the description of the undertaken processes, techniques
and technology, the rights management, etc.
The National Library of Australia has a powerful model for this.
The plethora of existing models and competing standards for meta-data
has led to projects which focus purely on translating from one standard
to another.
Appropriate Meta-data Standards
Issue Definition
Certain important standards already exist for meta-data. In the bibliographic
domain (and increasingly in non-library cultural domains), the Dublin
Core standard is of great importance.
Pragmatic Suggestions
- Review existing meta-data models and standards before creating your
own.
- Creating a totally new meta-data model for cultural collections should
be avoided.
- The meta-data work carried out by similar projects in the past is
likely to be relevant to your project – meta-data models travel
well between projects in the cultural area.
- Unless your project has good reason not to do so, the Dublin Core
fields should be included in the meta-data model. While museums may
find the CIMI model better fits their holdings, a common core set of
attributes should be aimed for, which will enable cross-collection searching.
- If a proprietary meta-data model is to be used, a mapping from this
model to the Dublin Core should also be developed.
- While a naming scheme or national naming convention may be very useful,
a full meta-data model is better, both in terms of the amount of data
that can be stored about an item, and also to enable more powerful searching
and interoperation with other projects and other countries.
Notes/Commentary
There is an impressive number of existing standards, covering various
aspects of meta-data. However, there is also significant overlap across
standards, and a very large population of institution-specific models,
where sectoral or cross-domain models have been neglected.
3.9 Publication
At this stage of the project, the digital master material has been created
and stored/backed up. A suitable meta-data model has been identified,
and the meta-data associated with each article has been created.
Preparation for publication involves processing the newly-created material
prior to publication. Typically, publication means display on the Internet,
and processing means reduction in image/audio/video file size, quality,
and downloads, to fit the operational characteristics of the Internet.
Image Processing
Issue Definition
The TIFF files created during the digitisation process are typically very
large (a few to many megabytes). Such files are not appropriate for Internet
publication, due to the great length of time that they would require to
download to the end user.
Pragmatic Suggestions
- Create delivery versions of master material. As a minimum, there must
be one delivery version. A second version, a ‘thumbnail’,
may also be useful, depending on the layout of the Web site on which
the material is to be published.
- Delivery versions are created by opening the master TIFF file in
an image processing package, and exporting it in JPEG, PNG file format(see
‘Image Standards’, below).
- Typically, colour resolution can be reduced, to 256 colours. If this
shows an appreciable loss of quality, a higher colour resolution can
be used. Choosing the right colour resolution usually requires some
subjective decision to be made.
- An image created at 72 DPI will show at approximately its original
size on many computer monitors. This makes 72DPI a reasonable choice
for many images which are to be viewed on-screen. For lower resolutions,
a subjective decision of ‘acceptable quality’ will be required.
- Choosing file format, colour resolution and pixel resolution involved
deciding on what is ‘acceptable’ quality. A balance must
be found between quality and file size.
- In general, the total image files on a Web page should not greatly
exceed 100 kilobytes. Larger images can certainly be published; however,
these should be accessed via a link from the Web page, with suitable
warning text that the download may be prolonged.
- Unless material is being streamed, video and audio material will typically
involve large file sizes, with the file downloaded before viewing offline.
However, the download time can be adjusted by changing the frames per
second of the video, the sampling rate of the audio, etc.
Notes/Commentary
Decisions regarding image processing depend to a large degree on personal
judgement. The guidelines provided here may be considered too strict or
too lax, depending on the project and the end user audience.
Image processing software such as Paint and Paintshop is freely available
online. More powerful image processing software may save sufficient time
and effort to justify the expense of the software.
Audio and video editing software is also available freely online. Equally,
audio and video hardware is usually supplied with the software required
to edit and process the data created.
3D and Virtual Reality Issues
Issue Definition
The guidelines provided above for image publication are not immediately
applicable to digital renderings of 3D and virtual reality material. However
the balance between quality and file size is a common one on the Internet.
Pragmatic Suggestions
- Viewers for 3D and VR material are not yet widely distributed with
operating system software. This contrasts with image, audio and video,
which are commonly provided with Windows software.
- Ensure that viewers for any 3D or VR material are readily available.
Make the viewer software available from the same site as the material.
This helps to overcome any issues with other software download sources
becoming unavailable.
- Evaluate multiple viewers before endorsing one or another. Compatibility
across file formats and viewers is not as standardized as in the still
image domain.
- Modern PCs, with a focus on games, will often have hardware accelerators
and increased graphics memory. This can have a profound effect on the
VR viewing experience.
Notes/Commentary
A VRML viewer which has been successfully used in one of the reference
projects (the Irish ACTIVATE project) is the Blaxxun Contact viewer.
Online Publication
Issue Definition
Many digitisation projects in the cultural area lead to the creation of
online cultural resources, usually a Web site with images, meta-data,
3D artifacts, etc. They range from the simplest content sites to complex,
software-driven portals and viewing engines. A large body of knowledge
covers the creation of Web sites; only a few guidelines are provided here,
as well as links to examples of Web sites nominated as best practice examples
by Minerva partners.
Pragmatic Suggestions
- Web sites should be easy to navigate – links to the front page
or to a table of content should be available throughout.
- Due attention should be paid to universal access and to the utilisation
of Web sites by the partially sighted and other disabled persons.
- Web pages should be short enough to minimize the amount of scrolling
necessary by the user.
- Images should be small enough not to disrupt the browsing experience.
Larger images should be linked to from the Web pages, with a note to
the effect that the image is large and download may be slow.
- The use of animations, pop-ups, pop-unders, Flash and similar technologies
should be treated with care. It should be possible to bypass lengthy
introductory animation sequences.
- Web sites should ideally be multilingual, with at least the host country
language and one or two other languages (commonly including English,
as the de facto online language standard) supported.
- Links to external resources should be verified on a periodic basis,
in order to minimize dead links and the annoyance associated with these.
Notes/Commentary
The actual process of making material available on the Web is one which
is widely understood and documented. This handbook does not provide guidance
on how the create Websites, program in HTML, build Web-enabled databases
and carry out the other tasks which are needed to create and maintain
a Web presence. It is anticipated that many of the cultural institutions
which utilise these guidelines will already have some Web server functionality
availability, which they will exploit for their digitisation project.
3.10 IPR and Copyright
The publication of any material online must be accompanied by some consideration
of the Intellectual Property Right (IPR) associated with the material.
For material which is in the public domain (such as particularly old books
or newspapers, or material placed explicitly in the public domain), there
is relatively little difficulty. However, many cultural institutions derive
revenue from the use of images of artifacts or images in their collections,
and so are defensive of copyright. Material, the copyright of which is
held by third parties, can only be published with the consent of such
third parties.
Fortunately, a range of technical options are available to protect the
copyright of material placed on the Internet. These are surveyed here.
Establishing Copyright
Issue Definition
The initial step when exploring the copyright situation for a cultural
item is to establish the ownership of that copyright.
Pragmatic Suggestions
- Establish the legal situation with regard to copyright and publication
in the country where the project is being carried out. Each country
has its own copyright laws, usually dating back to at least the 19th
century. Such laws usually apply to all forms of publication, including
online publication. They may, or may not, cover the act of digitisation,
which may be construed to be an archiving process, or may be considered
copying.
- On no account should online publication go ahead without copyright
being sought.
- Certain items, e.g. old newspapers, have clear copyright rules governing
them. Typically these allow free copying once the papers are of a certain
age. Items which fit into this category can be freely digitised and
published.
- For items whose copyright is vested in the institution carrying out
the project, internal permission will be required for digitisation and
online publication.
- For items whose copyright is held by a third party, such as the lender
or donor of a collection of historical items, that party’s permission
must be sought, in writing. Only when such permission has been received,
should publication go ahead.
- Securing permission to digitise and publish may involve payment. The
amount of payment must be balanced against the value of including the
relevant item(s) in the online resource.
Notes/Commentary
The copyright situation varies from country to country.
Safeguarding Copyright
Issue Definition
The publication of items online on the Web is an open invitation to make
copies of the items. It is infeasible to prevent some level of copying
of material displayed on the Web. However, there are a number of possible
procedures which can be considered, each of which has some effect in the
safeguarding of copyright.
Pragmatic Suggestions
- Establish whether or not copyright must be safeguarded.
- Agree the procedures to be used to safeguard copyright, with the copyright
holders.
- The following procedures are among those which could be considered:
- Addition of a visible watermark or copyright stamp on each image.
- Addition of an invisible digital watermark on each image. Such
marks can be used to prove the ownership of a ‘stolen’
image, as well as to track the use of the image across the Internet.
- Encryption of images, with the issuing of the appropriate key
only to registered users. This, of course, reduced the value of
the online image to the rest of the public.
- Restricting publication to low-resolution images, such as 75 dpi
for screen viewing. This restricts the degree to which images can
be used in other domains, such as printing, clothing, etc.
- Restrict publication to only small parts of an image.
- Display images only to registered, authorized members of a particular
community.
- Test the results of the copyright protection process using the first
few items, in order to ensure that the process does not have any unexpected
or unwanted effects.
Notes/Commentary
The approach which is most appropriate for any one project will depend
to a large degree on the goals of the project and the cultural institution,
as well as on the nature of the material. In general, the publication
of a small selection of images, at low resolution, is a common approach
for online galleries and museums. The relative uniqueness of many cultural
holdings provides proof of ownership of copyright in many situations.
3.11 Managing Digitisation Projects
The success of any project, including digitisation projects, is influenced
to a large degree by the management of the project. This section provides
a small number of guidelines specific to the management of digitisation
projects in particular.
Digitisation Process Management
Issue Definition
A typical digitisation project will involve dozens, hundreds or even thousands
of items to be digitised. In order to achieve an efficient project, it
is important to establish a work-flow that maximises the through-put of
the digitisation team. In addition, information resources such as a digitisation
project knowledge base will be of significant importance.
Pragmatic Suggestions
- Establish and document each of the steps that an item must go through
during the digitisation process. These will include, for example,
- retrieval from storage / usual location
- cleaning or preparation
- scanning or photography
- return to usual location
- file naming
- file storage
- creation of online delivery versions of large master files
- backup of servers / storage media
- Develop a digitisation project knowledge base that can be used to
track the object through the digitisation process, and enables the status
of the project to be reviewed at any time. This knowledge base may take
the form of a database (e.g. in MS Access, Oracle, MySQL, etc), or may
use a simple spreadsheet or even a collection of documents. The important
issue is not the format of the knowledge base, but the process which
ensures the recording of actions which are carried out.
- The name, identifier and other relevant information for each item
to be digitised should be entered in the digitisation project knowledge
base, as soon as the item has been selected. The status of the item
(i.e. which step it is has last completed) must also be recorded, on
an ongoing basis.
- Procedural choices must be made – for example, should items
be collected at the digitisation workstation at the start of each day,
each week, or on a per-item basis.
- Articles which require similar activities or hardware setups should
be digitised together. This reduces time spent setting up digital cameras,
configuring scanners, etc. The parameters for hardware setup should
be documented, in order to allow any digitisation to be replicated in
the event of file loss, etc.
- The location, phone numbers and backup staff of key service delivery
personnel (e.g. IT support) should be noted at the start of the project,
and remain available throughout.
Notes/Commentary
The larger the project, the more worthwhile it is to establish a process
and workflow. The efficiencies which this introduces will greatly repay
the time spent setting them up. The references below include some projects
which concentrate purely on this aspect of digitisation.
Team Development
Issue Definition
Digitisation projects often expose the staff of cultural institutions
to new technologies for the first time. Such technologies include digitisation
hardware, Web publication, image processing, meta-data tagging, database
development and population, etc.
Pragmatic Suggestions
- If possible, include at least one person with appropriate information
technology skills in the project team.
- Assess the state of knowledge of the personnel to work on the project,
and the IT skills that they will need, well in advance of the project.
Identify training needs and fill these before the project starts.
- IT skills are not the only ones which may be needed. Specialist skills
may be needed, as noted above, in the handling of delicate documents,
artifacts, etc. Appropriate training maybe available from the individuals
whose responsibility includes the source material.
Notes/Commentary
It is better to have a small core of skilled personnel working on a project
than a larger population of occasional participants. However, while developing
and using a particular skill is efficient for the project, staff may prefer
to be exposed to the full digitisation life-cycle. Digitisation and meta-data
tagging is not in itself particularly rewarding work – exposure
to other elements of the project will increase staff satisfaction.
Staff Training
Issue Definition
Unless the staff working on the project has significant experience from
prior projects, there will be a requirement for staff training. This will
include two quite different areas – the technology to be used, and
the handling of the source material.
Pragmatic Suggestions
- Do not assume that no staff training is required, nor that archives,
library or museum staff automatically has all the relevant expertise.
- Ensure that the training requirements of the staff on the project
are identified at the start of the project, i.e. already in the planning
phase. These training requirements should be included in the digitisation
project knowledge base, and acted upon before the training is needed
in the project.
- Certain training, such as the use of the digitisation technology,
may be to learn ‘on the job’; other training, such as handling
of source materials, requires training in advance.
- A smaller core of personnel, who are trained and develop experience
during the whole project, is to be preferred to a larger, more casual
group which changes its membership more frequently.
- Technology training may be well delivered from another project in
the same institution; alternatively an outside digitisation agency may
be able to provide training.
- Curator training may best be provided by the individuals who are responsible
for the care of the original material.
Notes/Commentary
A lack of staff training can lead to unfortunate and irreversible accidents
or incidents early in the project; the same may result at any time if
staff is removed from the project and new personnel start to work. A small,
well-trained core is a desirable aspect of such projects.
Time invested in training at the start of the project should be repaid
in extra productivity and less problems during the life of the project.
Working with Third Parties for Technical Assistance
Issue Definition
It is often appropriate for a digitisation project to engage the services
of one or more third parties during the project. The services which are
most commonly provided include the actual digitisation itself, the management
of the project, integration with third party systems, software development,
etc. This allows the cultural body to concentrate on its own areas of
expertise, without need to train and retain staff with advanced IT or
other skills.
Pragmatic Suggestions
- As with any other project, the relationship between technical partners
and other project members should be governed by clear, strict contracts.
A documented and signed specification of the products or services to
be provided should be agreed before any work is carried out.
- The work being carried out should be reviewed on a regular basis,
to ensure that what is being delivered is in fact what the project wants
or needs.
- While the use of third parties can be convenient, it should be borne
in mind that any expertise or experience to be gained during the execution
of the outsourced work will be lost to the cultural institution at the
end of the project. This also applies to temporary staff who is employed
for the duration of a project. It may be better to dedicate a long term
member of staff to a project, while replacing him in the short term
with a contractor.
Notes/Commentary
Certain large projects, such as the French national digitisation programme,
have identified a preferred supplier, the relationship with whom may stretch
for several projects and several years. Having established a working relationship
with a supplier, the value of changing supplier between projects may need
to be questioned.
Working with Third Parties in Cooperative Projects and Content Sharing
Issue Definition
Many digitisation projects are either cooperative efforts which involve
two or more cultural bodies, or else EU-funded Framework projects, which
almost always have multiple partners in multiple countries. The guidelines
for establishing and managing multi-partner projects are many, and go
beyond the scope of this document. However, a few pointers are included
Pragmatic Suggestions
- Ensure that all partners are aware of, and have endorsed, their roles
and responsibilities within the project. Refresh this knowledge on a
regular basis.
- Establish a common mode of communication across partners, and ensure
that all partners receive the information which is aimed at them. Electronic
mail is ideal for this purpose, so long as partners read and reply to
such mail.
- Subcontractors should be governed by strict commercial contracts,
with their deliverables clearly and unambiguously defined.
- The IPR of all partners should be clearly documented and formally
signed by all partners. A partnership agreement which clearly states
the IP Rights covering material which is being brought to the project,
and material which is created by the project, should be agreed in advance
of the project commencing.
- Each partner should have a clear role in the project – if a
partner’s role is not clear, review whether or not the partner
is necessary to the project.
Notes/Commentary
The notes above are only a small part of the possible material that could
be provided on the establishment and management of multi-partner projects.
Partners and suppliers are a major source of delay and confusion within
a project – clear agreement and common endorsement of the roles
and responsibilities of all partners at all times can help to avoid this.
Costs
Issue Definition
Building a digital collection is normally very expensive. Projects, therefore,
have to take into account all start-up and infrastructural costs as well
as costs for running the project. That means costs for initial planning,
data specifications, tracking and documentation systems, staff training,
and so forth, as well as the incremental cost for digitising the actual
source material.
Pragmatic Suggestions
- Digitisation projects should consider the following costs involved
in the design, implementation and maintenance of a digital collection:
- Staff development
- Facilities needed
- Operational costs
- Costs for storage and for delivery systems
- Staff development includes salaries for project management, Web programmer,
educational officer, technical support, etc. but also travel costs and
training
- Costs for facilities are often connected to questions concerning required
functionalities and the need for tradeoffs. Projects have for instance
to decide whether keeping costs on a low level is more important to
the overall success of the project than achieving the highest possible
standard for image capture.
- Operational costs to consider are:
- Time for handling source material (from shelf to image capture
device and back again) as a percentage of total salary cost per
day.
- Preparation of source material (conservation, cleaning etc).
- Capture time (from set-up to naming and saving) provided as a
percentage of the operators total salary costs per day.
- Cataloguing and handling of meta-data as a percentage of total
salary costs.
- Hardware and software cost per digitised item (preferably based
on depreciations or replacements costs rather than acquisition costs.
- Quality assurance time as a percentage of salary costs.
- Hardware and software maintenance.
- Technical support time related to capture.
- Project management time related to capture.
- Training related to capture.
- Be aware of that image capture often is the least costly part of a
digitisation project. On average, one third of total costs are connected
to digital convention, slightly less than one third to creation of meta-data,
a bit more then one third to administrative and quality assurance tasks.
The rest is long term maintenance costs.
- Storage costs to consider should normally be total costs for maintenance
per gigabyte.
Notes/Commentary
Sustainability in the long term is often pushed down the list of priorities
by more immediate and pressing concerns. Regardless of the quality and
robustness of the digital resources created by a digitisation project,
they will not last long if the project in mind cannot find funds for their
maintenance.
|
|
|