the australian newspapers digitisation program: helping communities access and explore their...
TRANSCRIPT
1
Helping communities access and explore their newspaper heritage.
Rose Holley – Manager Newspaper Digitisation Program http://www.nla.gov.au/ndp [email protected]
Australian Media Traditions Conference23 November 2007, Charles Sturt University, Bathurst
2
Status of the ProgramNovember 2006 Minister for Arts and Sports approval
Budget approval -$8 million for 3 million pages over 4 years
Contracts signed with digitisation suppliers
April 2007 program pilot phase commences
3
Content and CoverageNational Content
Initially a title from each state
Focus on major titles from each state first
Anticipated that ‘regional’ titles may be contributed later
Coverage: published between 1803 – 1954
(out of copyright)
West Australian
Northern Territory Times
Courier Mail
Advertiser Sydney Gazette
Argus
Mercury
Canberra Times
4
First Newspaper
• First page of first Australian newspaper ever published
The Sydney Gazette and New South Wales AdvertiserSaturday March 5 1803
5
Through 150 years
• Up to 1954 (when Copyright applies), and later if agreement with publishers.
The Argus 22 August 1945
6
Relationship - ANPLANWebsite: http://www.nla.gov.au/anplan/
7
Keep Up to Date with Progress• Website: http://www.nla.gov.au/ndp/
8
National Help• NLA working with State and Territory
Libraries as part of ANPLAN. • Libraries suggest titles and dates and
provide microfilm for digitising.• ANPLAN members and other stakeholders
will provide feedback on the search and delivery prototype.
• Developing model for national contribution of regional newspapers.
9
Process in briefNational sourcing of selected newspaper microfilm
masters.
Masters scanned by Contractor, Sydney to tiff files.
NLA perform quality assurance, add metadata.
Contractor, India process tiff files - OCR, zoning, xml markup.
NLA QA files, ingest to system, create derivatives for delivery.
10
LogisticsAustralia (State Capitals – Sydney/Canberra) USA (Virginia) - India (Hyderabad, Chennai)
11
6 Month Progress• IT Infrastructure and storage implemented at NLA
• Content management and ingest software developed by NLA to support workflow
• Quality assurance and production software developed by US/India contractor
• Pilot data sent to contractors to test workflows, systems and software against agreed project spec.
12
Next 6 months• Acceptance of pilot data then commence
production phase (3 million pages)
• Development of search and delivery prototype
• Public launch of service with a good body of content in 2008
• Progressive addition of content – national program ongoing
13
Technology – internal NLAOld newspapers being processed and delivered
using latest digital technology
• NLA developing in house:– Ingest and storage system– Workflow and content management system including
quality assurance module– Search and delivery system
• NLA providing:– System Infrastructure
(storage, backup, disaster recovery)
14
Infrastructure and Storage
Online Storage – 70 TB:• Working space for images in processing 40TB for 1 million pages• Search and delivery derivatives 30TB for 3 million pages• XML files, database systems and indexes 1 TB
Offline Storage – unlimited for master images on tape.
15
Establishing Workflows
16
Technology - external
• Scanning microfilm using Flexscan/Eclipse scanner and latest software (nextstar) from NextScanwww.nextscan.com
20,000 pages a week.
17
Scanning Contractor
18
Digital Images returned to NLA
19
Quality Assurance at NLA
Use 2 widescreen monitors placed vertically. Can view complete page within context of issue.
Add metadata, sort out missing and duplicate pages within an issue.
Prepare batches to send for OCR.
20
Metadata
21
Page verification
22
23
Technology - external
Software developed to:• Zone areas and articles on a page • Flag continuing articles across multiple pages• Categorise articles on a page• OCR text on a page• Re-key headings and first 4 lines of text.• Deliver XML files (ALTO) and METS/MODS
files.
24
India Facility - Hyderabad
25
26
Quality Assurance
27
OCR Accuracy
28
Batch reporting
29
Acceptance Criteria
30
Prototype Development
Under discussion:• Derivative sizes and zoom technology
testing• Search and Browse features• Results and refinement of results• User interaction with source (web 2.0)• Interface design
31
Digital Newspaper Searching
• Newspapers full text searchable• Image captions searchable• Search across multiple papers e.g. by
persons name.• Refine searching by:
– Date– Newspaper title– State published
32
Refine search by categories
• News• Advertising• Birth Death Marriage notices• Obituaries• Editorial commentary and letters• Shipping News• Arts and leisure• Detailed lists, results, guides
33
Search Illustrations
Categorised as:• Photo• Cartoon• Map• Graph• IllustrationCaptions searchable
Canberra Times 26 July 1928 page 6
34
Browsing and Viewing
• Browse papers page by page• Zoom in and out of image
– to read small text– to view context of article within page layout
• Print article or entire page or issue
35
Zoom technology
36
Testing derivative sizes and zooming
37
Prototype wireframe
38
Other features
Under discussion:• OCR correction by users• Personal annotation of articles by users• Tagging results• Creating public sets (for historical events)• Clustering results• Searching across other relevant resources (paid
subscription services, international resources, other digital resources)
39
Prototype release
• To be released to stakeholders who have given microfilm content
• Stakeholders able to view their data• Feedback on data quality and search
functionality• Amendments made and then ‘search and
delivery version 1’ released to a wider group for testing and feedback before public launch in 2008.
40
Pilot Data• Canberra Times• Sydney Gazette• Northern Territory Times• South Australia Advertiser• Hobart Town Gazette, Courier, Colonial, Mercury• Melbourne Argus• Perth Gazette• West Australian• Brisbane Courier Mail(12 titles, 8000 issues = 50,000 pages = 500,000 articles)
41http://www.nla.gov.au/ndp