css.php

Brooklyn Daily Eagle

The Brooklyn Daily Eagle: A Case Study

Introduction

Newspapers from the nineteenth and early twentieth century have a troubled future.  Intended from the outset to be “important today, discarded tomorrow,” printed on large sheets of cheap paper that make them expensive to index, hard to photograph, and challenging to preserve (Deegan et al., n.d., p. 3), they are sometimes available to the patient researcher willing to travel to select libraries and scroll through reels of microfilm, but few have been made available online.

The Brooklyn Daily Eagle 1841-1902 Online (Eagle) project is an example of how an important historical newspaper was preserved digitally through outsourcing and by taking advantage of cutting-edge technology.  The Brooklyn Public Library (BPL) realized the need to digitize the Eagle, obtained funding from Institute of Museum and Library Services (IMLS), coordinated with the Library of Congress (LC) to borrow the best microfilms available, outsourced the scanning work to Online Computer Library Center (OCLC), and employed Olive Software (Olive) to quickly process 61 years of TIFF files and bring Phase 1 of the project online (How the repository was created, n.d., para.1-3).

Project Goals and Objectives

The Eagle’s Web site links to a PDF document that describes a pilot project completed in 2001 by the British Library which mirrored the Eagle project.  The two projects are very similar and share common goals, which may be paraphrased as: (1) to use new technology to quickly digitize a large amount of materials; (2) to index them automatically; (3) to deliver them on the web; and (4) make them accessible to a wide audience. (Deegan et al., n.d., p. 1).

While goals are commonly broad and abstract, objectives need to much more concrete and specific in nature (Grassian & Kaplowitz, 2001, p. 136), showing intended effects.  Although unable to find the original IMLS proposal that would list such objectives, Joy Holland, in an e-mail response for this study, noted that “a huge amount of information on Brooklyn in the 19th century was locked away” before the Eagle.  No indexing was available, and patrons had to know specific dates and then start scrolling through reels of microfilm.  The objective was to open up and preserve this unique resource via digitalization (personal communication, May 6, 2009).  BPL’s methods of measuring if this objective was reached will be discussed in a later section.

Planning

A key player in the Eagle project, Susan Benz of BPL presented her insights on planning and managing digital projects in the Special Library Association’s 2006 Museum, Arts and Humanities Division Conference, and the following attempts to summarize her thoughts, as recorded by Gerald Patout of SLA.  Managers should use the grant proposal itself as a guide in the process.  They should “consider the audience, preservation aspects, purpose, mission, timeframe, resources, and standards outlined in the proposal” as they progress through the project.  Outsourcing should be considered, but only after the vendors’ track records are studied carefully.  Communication with other libraries that are doing similar projects is vital.  Managers should be aware of their patrons’ needs and adapt projects to suit them.  And during the project, managers should conduct “outcomes based evaluations” to ensure they are reaching their target audiences.  And of course, managers must ensure patrons are aware of the project (Patout, 2006, para 10).

BPL appears to have learned much by studying the British Library’s pilot project, and capitalized on the partnerships that were developed between OCLC and Olive Software.  As Holland mentions, BPL was the first public library to use Olive software, “and the British Library experience was about the only model around. We were very early on in the curve.”  At the time, Olive had no competitors.  Prior to IMLS funding, BPL had submitted another proposal to an unnamed institution, but it was declined. This proposal also involved OCLC and LC, and both were very helpful with the proposal re-write, especially with the cost estimates (personal communication, May 6, 2009).

OCLC has its own Preservation Service Center and Conversion plant in Bethlehem, P.A. (Deegan et al, n.d., p.5), and uses two different approaches when preserving historical newspapers.  In her PowerPoint presentation comparing OCLC’s work with ContentDM and Olive, Spruill (2006) notes that no matter which approach is chosen,  OCLC supports NARA, National Digital Newspaper Program Guidelines, and DLF (slide 11).

With extensive experience in digital preservation, OCLC was a sound choice.  Olive was chosen because it was the only company around with experience and success in this type of processing.

Material Selection

The Brooklyn Daily Eagle was published continuously from 1841 until 1955, and then was briefly revived from 1960-1963 (Brooklyn Daily Eagle 1841-1902 online, n.d., para 2).  The BPL team broke the project up into phases according to both complexity and need (as described in “Goals and Objective”).  Subsequent phases are out of the scope of this study, but according to the Web site, the newspaper’s layout became much more complex after 1902, and the paper got much longer: 1902 was a logical cutoff for Phase 1.  Though not stated, clearing copyrights most assuredly make subsequent phases more labor intensive.

Many patrons expressed interest in the newspaper and its audience goes beyond Brooklyn.  BPL initially targeted researchers, historians, but the general public has become involved as well.  Covering both national and international news, as well as Brooklyn daily life, the newspaper “provides a window into Brooklyn’s past,” and chronicles the city’s attitudes to contemporary events in the U.S and abroad. (History and timeline, n.d., para. 1).  As such, it was met with excitement when the first Beta version appeared in 2003 (Holland & Benz, 2004, p.1).

Conservation/Preservation

The Eagle is part of the Brooklyn Collection, BPL’s local history division that includes “books, maps, photographs, newspapers, ephemera, prints, and newspaper clippings” that are used by researchers worldwide as well as “local students, genealogists, journalists and filmmakers.” (Brooklyn Collection, 2009, para. 1).  Although the collection has the “entire run” of the Eagle available on microfilm (Brooklyn Daily Eagle Materials, n.d., para 2), BPL elected instead to borrow the “new second generation negative microfilm reels” owned by LC.  (How the repository was created, n.d., para 1).

This extra step ensured that the project started with the best images available, and wisely avoided the costly effort of scanning the Eagle from the collection’s clippings housed in it calls the “morgue.” (Brooklyn Daily Eagle Materials, n.d., para 2)  Besides being financially impractical, such a decision would have had many conservation and preservation implications.  Rather than dealing with originals, the Eagle project borrowed microfilmed surrogates.  But essentially, these surrogates are at this point (and especially in the future) more valuable than the crumbling originals, and extra care had to be taken to ensure they were not damaged or lost in the digitization process.  This was not the first time LC and OCLC worked together on such an endeavor, and OCLC’s Preservation Service Center’s expertise surely weighed favorably in LC’s decision to lend out the microfilms.

In one of its datasheets, Olive claims its ActivePaper Archive software creates “a future-proof format that eliminates the risk of technological obsolescence” (Olive ActivePaper Archive, 2007, p.2).  This tall claim is based largely on its use of industry-standard formats: TIFF, ASCII ( for its XML), PNG, and PDF that allow viewing with no additional software.  Though its “future-proof” claim seems more like advertisement gimmick, Olive’s ability to separate content from all proprietary software by encapsulating and preserving “text, metadata, structure, context, relationships, styling data, file properties, knowledge tags, hyperlinks, graphics, and image maps” (Underlying Technology, 2009, para. 4) into XML is commendable.

But what about the TIFF files?  Did Olive care about preserving the originals after its access files were generated, and the indexing done?  No mention of their preservation was found on the Olive Web site.  The loss of TIFF originals could make BPL more wedded to Olive’s infrastructure.  (As a side note, Olive has functionality to monetize collections, which allow users to order, pay, and download high quality images online, so it does deal with TIFF or near TIFF-quality images in other applications.  This functionality was not enabled on the Eagle site.  BPL mentions that pictures from the Brooklyn Collection may be ordered by phone (Frequently Asked Questions, n.d., para. 11), not via Olive.)

The TIFF files created by OCLC and processed by Olive were intermediary files that made the Web site possible.  Though no longer needed for day-to-day operation, they could be used in the future if another software company develops a new and improved way of archiving historical newspapers.  Preserving such a number of large files is not cheap, but it would be a shame to lose such rudimentary building blocks.  Holland was unsure, but thought that BPL retained the TIFFs in a “storage facility” and that they were backed up regularly along with other library items (personal communication, May 6, 2009).

Imaging/File Naming Standards

Technicians at OCLC used Sunrise scanners to create Group IV TIFF images at 300 dpi from the microfilm reels supplied by LC.  As has been mentioned before, the nature of newspapers (even from microfilm) makes imaging demanding, and the Web site mentions that the images needed to be cleaned up and aligned prior to being sent to Olive (How the repository was created, n.d., para 2).

File naming details in this process were not found, but given the limitations of the 8.3 naming convention, the very specific nature of the file exchange (OCLC to Olive), and the number of files involved, it might be speculated that the file naming system involved a carefully worked out convention that included, at a minimum, the issue date and page number.  In 2002, large files were commonly transmitted via FTP, so we might also speculate that the TIFF files were transmitted rather than physically shipped to Olive, and the process could have been done in batches, so that Olive did not need to wait for OCLC to have everything complete before beginning their process.  (A quick calculation shows that over 22,000 issues of the Eagle were printed in 61 years, and if each issue typically contained 20 or 30 pages, the number of TIFF files sent to Olive may have approached one-half million.)   Today’s Sunrise scanning machines image reels of microfilm almost automatically, and a technician can scan up to 8,000 pages per day (High performance microfilm scanners, 2006,  para. 6), but in 2002, the technology may not have been as robust, and the subject matter may have been more challenging to image.  Group IV compression was used on these TIFF files, presumably to save some space.  A typical page of the Eagle’s has a complex layout, but is textual in nature, and has headings of different heights, and has a lot of noise.  Resolution higher than 300 dpi was unnecessary.

Olive took the TIFF files from OCLC and generated PDF, PNG, JPG, and XML index files “produced by OCR” and XML index “location files, which include bitmapping information” (How the repository was created, n.d. para. 2).  A quick look at the file names on the Eagle Web site reveals that the 8.3 convention is used only for the issue’s individual pages (e.g. “pg011.png”).  The PDFs contain an entire issue, some in excess of 20 pages, and file sizes not uncommonly exceed 25 megabytes.  They have long and meaningful names that include date information (e.g. “002-BEG-1902-12-16-001-SINGLE.PDF”).

Periodicals are by nature organized by date, and a little detective work provides a glimpse of the folder structure used by Olive.  The following path points to the eleventh page of the December 4, 1902 issue of the Eagle: “BEG/1902/12/04/1/Img/Pg011.PNG.”  These large-dimensioned PNG files are not very readable, and are more intended to give the user a feel for the whole page, and how the columns appeared together.  Olive allows the user to click on an individual article, and a popup window displays a zoomed-in, very readable PNG file of the article itself.  These files have a nine character file name starting with “ar” and seven numbers that seem to be internally generated by the system (e.g.”ar0010201.png”).  As with the PDF file names, the 8.3 convention is not used here, and there is really not a reason for it – these access files are not intended to be used outside the Web site.  From the popup, the user is allowed to view the article as a PDF, but this function merely downloads the whole issue as a PDF, not the individual article (Brooklyn Daily Eagle (1841-1902) online, n.d.).

An inspection of the PDFs reveals that they were generated directly from TIFF files: the application in the properties window is listed as “ActivePaper TIFF Converter for Windows.”  The PDFs are of very high quality, and can be zoomed in to see much detail.  Although the Web site says that Olive generated JPGs for the site, none were observed in this study.  Thumbnails were also in PNG format.  In terms of image dimensions, there is not much consistency in size.  The PNG full page images are all around 1600-1650 pixels on the long dimension.  Article PNGs vary widely in size, as do thumbnails.  The PDF full issue files also vary in size.  They are composed of very high quality images that are around 24-25 inches by 17-18 inches – probably the original size of the newspaper.

PNG format was a good choice for the Web access files.  Besser (2005) writes that support for PNG is growing, and that it is designed to eventually replace the GIF format.  It supports compression, and provides more than GIF’s 8 bit, 256 color palette (The image, para. 30).  Although imaging for the Eagle is done in gray-scale, the PNG format choice wisely looked towards evolving standards.  Besser notes that the PNG format may in the future replace the current TIFF/GIF/JPEG standard itself, serving as both archival and access formats (para. 29).  Though PDFs are proprietary, they are the de-facto standard for web file downloads, and this format was an equally good choice, given their ability to combine numerous files into one package.

Quality Control

Holland notes that the BPL team oversaw the Eagle’s progress, did quality control and ensured that the Web site would “look the way we wanted it,” which, to her chagrin, is actually not the way it looks today.  Apparently a recent Olive software upgrade was not popular at BPL.  Quality control training was also necessary for team members, though no details were provided (personal communication, May 6, 2009).

The Eagle site mentions that OCLC had to clean up and align the TIFFs (How the repository was created, n.d., para 2), and there must have been a procedure that allowed Olive to request a cleaner image if one failed to be processed correctly.  But this is supposition.

Outsourcing

The British Library’s pilot project demonstrated the OCLC/Olive partnership worked well with this type of historical newspaper digitalization project, and BPL wisely followed its outsourcing model.  The complexity of digitizing and cataloging this amount of data would have been challenging to do in-house.  Spruill (2006) seems to infer that ContentDM is a viable DIY option (Slide 10), but such a choice would have had massive implications for BPL, including adding and training staff, buying new equipment, and addressing the challenges of OCR for historical newspapers, and would have probably degraded the final product.  The project end-dates would have been very difficult to determine.  Outsourcing allowed for a concise timeframe and proven outcomes, and by proposing the British Library’s approach, BPL provided a proof of concept and was able obtain the IMLS grant.

Research has not shown any problems associated with the outsourcing, although such problems are often not advertised.  Few projects progress seamlessly.  Prior mention of the unpopular software update by Olive, and issues of performance, detailed later in the “Evaluation” section, may hint at some friction in the outsourcing effort.

Storage

Olive shuns the use of relational databases: it claims that this model does not perform well “with the volume and complexity of digital newspaper archives.”  Instead it relies on a file repository, and drives its processing with XML.  File locations are embedded within the markup, and data are “organized in a logical file-system hierarchy.”  This system provides great flexibility, as the files can be distributed simply across multiple hard drives and storage media.  (Deegan et al., n.d., p. 7)

The Eagle is hosted on Olive’s Web servers (J. Holland, personal communication, May 6, 2009).  It would be interesting to know what models are used, and what contingency plans are in effect if the files on the web server(s) are lost, but this information was not found.  Presumably the outsourcing contract between Olive and BPL would include a service level agreement (SLA) that would stipulate such plans.

Cataloging/Metadata

Olive generates its metadata according to a document type definition (DTD) that it has developed, which is a hybridization of Dublin Dore, News ML/NITF (News Industry Text Format) and “Preservation Mark-up Language.” (How the repository was created, n.d., para 2).   Its ActivePaper Archive software generates metadata in “three layers,” one based on each of the three standards.  The first two ensure that the metadata is based on “an open, integrative platform.”  The third standard (PRML) is used to map the layout of a document, and record the coordinates of each of its pieces (Deegan et al., n.d., p. 7).  This structural metadata is vital in the process that Olive employs to automatically generate the descriptive metadata that is available for keyword searching.

Automated indexing is not a new concept, but the complex page layouts of historical papers make the task particularly challenging.  Among the issues faced are the number and variety of information objects on a page, each of which may jump to subsequent pages.  (Deegan et al., n.d., p. 8)  And of course there are problems with image quality.  The authors note that when microfilm is scanned to TIFF files, the quality is almost always poor.  Newspapers were typically still attached to their “binds” and the image often became “skewed” when originally microfilmed, resulting in “rotated or curved characters and images” (p. 11).  A quick look through some of the Eagle’s issues (particularly the older ones) indeed reveals that some articles are partially or totally unreadable.  All newspapers contain “noise” that makes their column layouts difficult to calculate.  Paper decay, dirty scanner lenses, deterioration of the microfilm and poor quality scanning all contribute to the distortion of the images which Olive attempts index. (Deegan et al., n.d., p. 12).

The most complicated part of Olive’s processing is its ActivePaper Archive software’s ability to ingest the structural metadata that its “zoning engine” has previously created, pinpoint the coordinates of an article using its “segmentation engine,” optically discern the textual content, and generate descriptive metadata in XML about the particular article in the form of keywords.  (Presumably a list of stop words is employed, though no mention of this was found.) This OCR process is hampered by large font titles of various sizes, “garbage elements” in the text itself, “broken vertical and horizontal lines,” and scratches in the microfilm (Deegan et al., n.d.. p. 14-15).  Olive employs artificial intelligence and “fuzzy logic algorithms” to overcome poor image quality and relies on “OCR-generated word patterns” that it stores in the XML itself to aid searchability on the user interface.  Olive claims that this provides better search results (Deegan et al., n.d.. p. 6-7).  And though it is not clear if this is done concurrently or recursively, Olive generates three PNG files: a low resolution thumbnail, a screen-size, mid-resolution page image, and a smaller, higher resolution article image, and stores their physical path locations within the XML.  And finally, the TIFFs for each issue are combined into a PDF using ActivePaper TIFF Converter for Windows, stored in the repository and its path likewise recorded in the XML.

Delivery/Authentication/Availability

The Eagle is delivered online, free of charge to anyone with an internet connection.  No authentication or login is required.  Large, high quality PDFs of each issue are available for download.  Individual articles in PNG format may be saved, printed, sent by email, or saved as HTML.  A researcher may save articles in a personal collection for later retrieval, but as there is no login functionality, access to personal collections is limited to the particular computer it was saved on.  A public computer in a library would amass collections of its various users if its temporary files were not cleared.  A quick look at the Eagle’s FAQ page confirms that cookies are used (and must be enabled) for this functionality to work (Frequently Asked Questions, n.d., para. 9).

The user interface is provided by Olive.  It allows for date and keyword searching.  Browsing is available by date range.  It provides easy navigation via tabs and links (Benz, n.d., Slide 3).

Copyright/Intellectual Property Issues

Everything contained in Phase 1 of this project was published prior to 1903 and is in the public domain.  This made the first 61 years of the Eagle easier to process, as BPL did not need to clear copyrights.  Users can freely download an entire issue of the Eagle as a high quality PDF or individual articles as PNG files “for non-commercial educational, personal or scholarly purposes.”  BPL requires written approval and will possibly charge a fee for commercial users.  Other interesting restrictions include: “deep linking” to a specific digital object from another Web site, “framing” the Daily Eagle Web site within another Web site, and “harvesting” of any Daily Eagle data for use on another Web site or product.   BPL goes on to assert that the “graphic elements and the contents of this web site and database, excluding materials that are in the public domain, are copyright © Brooklyn Public Library 2003.” (Terms and conditions of use, n.d. para 5-6).

Evaluation

BPL reserves the right to gather user data for “internal, statistical purposes and in order to improve our service to the public.”  However, it guarantees that such statistics are anonymous, and will not be shared, except with “sponsors and business partners” (Terms and conditions of use, n.d. para 5-6).  Surely IMLS requires some kind of evaluation once the project was complete and nearly quarter of a million dollars spent, and Holland & Benz (2004) confirm that Eagle traffic was monitored quantitatively: in the first year 60,000 visits were logged per month.  (A visit was defined as “a session of 15 minutes or more.”) (p. 1).

Evaluation should also focus on user satisfaction, and of particular interest is the FAQ page, which seems to address issues that users have complained about: system slowness, system freezing up, searches not working, partial images retrieved, etc.  These types of issues arise often when users have slow connections, older computers, or when net traffic is particularly high.   But the Eagle is a huge repository, and in this case study it was observed that searches in general do take a lot of time, especially if not well formed, and multiple users doing many searches may possibly degrade performance.  This type of quantitative data would hopefully be captured and evaluated too, both by BPL and Olive.

Did BPL triangulate this data with more qualitative information?   Holland & Benz (2004) suggest so.  They note that users of all backgrounds have responded favorably to the Eagle online, “from novices to advanced researchers” — many use it to do genealogical research, while others appreciate its historical data, and use it in schools and colleges.  The authors go on to claim that the Brooklyn Collection in general has benefitted from the Eagle’s success, showing a significant increase from visitors outside the metropolitan area — the most frequently asked question is about the status of Phase 2 (p. 1).  This data was gathered through surveys and unsolicited e-mails. In her PowerPoint presentation, Benz (n.d.) provides copies of the “Online Usability Assessment” and “Functionality Problems (Use of Web Site) Form.” (Slides 19-20).  Some of the content is chopped off, but what is available makes it clear that much evaluation took place once the Web site went live.

Funding and Maintenance

In 2001, IMLS awarded BPL a two-year, Phase 1 grant for $239,412.00 (Grant awards, n.d., para.2).  This was a one-time allotment, and since then BPL has “picked up the tab” for day to day maintenance, including Olive’s monthly fee for hosting the site (J. Holland, personal communication, May 6, 2009).  There is also a place on the Web site where patrons may make personal donations to the project, but research did not find how successful that has been.

Staffing

The OCLC Annual Report 2001-2002 (2002) shows a picture of the seven members of the project.  Their names were Jeffrey Marable, Judith Walsh, Sofia Sequenzia, Joy Holland, Susan Shields, Susan Benz and Siobhan Reardon (p. 15).  Susan Benz was the manager of the team, and was hired by BPL especially for this project.  Some part-timers were hired to help out with “Selected Subjects” (J. Holland, personal communication, May 6, 2009).

Collaboration

The Daily Eagle was a huge collaboration effort, as has been previously stated.  BPL, IMLS, OCLC and Olive each played important roles.  Research has found no details on how well the team members worked together, or if there was any evaluation of this type of collaboration.  Noted in other sections is BPL team collaboration in quality control and in the supplemental timelines, historical contexts, and “Selected Subjects” pages.

Documentation

The user interface comes with Olive’s standard set of help screens that guide the user through search scenarios and describe and solve typical user problems.  The user interface may be customized for the particular client: BPL added its own FAQ page, presumably with questions that have come up after the launch, and which pertain specifically to the Brooklyn Daily Eagle.  There was also a page called “How the repository was created,” with a link to the British Library’s pilot PDF, as well as the page listed in the next section.

Timelines/Historical Contexts

The Web site provides a structural timeline broken down into five year intervals, each section discussing changes in the paper’s layout and how advertisements and articles were organized.   It also has separate pages that cover genealogy information and immigration.  These are what Benz (n.d.) calls “Value Added Content” in her PowerPoint presentation. (Slide 21), and it was all created by BPL.  Also in this category is the “Selected Subjects” page which provides for very limited subject searching.  Extremely time-consuming and expensive, subject indexing is out of scope for Olive, and BPL staff worked together to compile a very limited list of hyperlinks to interesting stories, which may be browsed by subject category.  Each of the twenty-three subject categories has from five to thirty hyperlinks below it, along with a one-line synopsis of the article. According to Holland, no thesaurus or controlled vocabulary was used do the subject indexing (personal communication, May 6, 2009).

Possible Improvements

More subject categories would of course improve the browsing experience.  Hyperlinks within the articles themselves to other instances of keywords would be fantastic, though it may not be technically feasible.  The How the repository was created page feels a bit thrown together: why link to another project’s documents, rather than writing your own?  The Web site has little technical information on the project itself, and any account of the project work flow is sketchy.  Project members are not listed or given credit.  But it must be acknowledged that the target audience clearly is not as interested in the methods used to create the repository as it is in the actual content.

The Eagle will of course be improved with the addition of Phase 2 data, and this depends largely on funding.  Certainly the interest and technology are available, and subsequent ActivePaper Archive projects have dealt with copyright issues successfully.  But it all costs money.

Conclusions

This case study has demonstrated how BPL was able to manage a very complicated digitization project by selecting a prior project, and following closely in its footsteps.  Using the British Library’s project as a proof of concept, the library was able to clearly present its goals and objectives to IMLS, and obtain funding.  It negotiated with LC to borrow the best microfilms and outsourced the scanning to OCLC.  Olive did the re-imaging, and created a repository of metadata on which was built the Eagle’s user interface.   BPL customized the site with historical context information, and subject browsing links.  Once live, the Eagle was evaluated both qualitatively and quantitatively, and BPL responded to the results by added further customization in the form of FAQ, as well as an additional genealogical page.  The Eagle project depended on partnerships, collaborations and good management skills, and the overall enthusiasm with which it has been received is proof of their worth.

References

Benz, S. (n.d.). Brooklyn Daily Eagle online: Steps towards success in newspaper digitization. Retrieved May 5, 2009 from worldcat.org/arcviewer/1/OCC/2005/12/27/0000016179/viewer/file184.ppt

Besser, H. (2005). Hubbard, S with Lenert, D. (Eds.).  Introduction to Imaging, Revised Edition. Retrieved May 9, 2009 from The Getty Standards and Digital Resource Management Program of the J. Paul Getty Trust Website at: http://getty.edu/research/conducting_research/standards/introimages/image.html

Brooklyn Collection. (2009). Accessed April 21 from Brooklyn Public Library’s Brooklyn Collection Web site: http://www.brooklynpubliclibrary.org/brooklyncollection

Brooklyn Daily Eagle Materials. (2009). Accessed April 21 from Brooklyn Public Library’s Brooklyn Collection Web site: http://www.brooklynpubliclibrary.org/brooklyncollection/eagle.jsp

Brooklyn Daily Eagle (1841-1902) Online. (n.d.). Retrieved March 15, 2009 from Brooklyn Daily Eagle (1841-1902) Online Web site: http://eagle.brooklynpubliclibrary.org/Default/Skins/BEagle/Client.asp?Skin=BEagle

Connecting through images: Susan Benz. (2007, March 15). Library Journal.  Retrieved May 6, 2009 from http://www.libraryjournal.com/article/CA6423409.html

Deegan, M., King, E. & Steinvil, E. (n.d.).  British Library Microfilmed Newspapers and Oxford Grey Literature Online: Project Report.  Accessed March 28, 2009 from Olive Software Web site: http://www.olivesoftware.com/documents/olive_software-british_library_archive_case_study.pdf

Frequently Asked Questions. (n.d.). Retrieved March 29, 2009 from the Brooklyn Daily Eagle (1841-1902) Online Web site: http://eagle.brooklynpubliclibrary.org/Default/Skins/BEagle/Client.asp?Skin=BEagle

Getting started with Active Paper Archive. (2007).  Retrieved March 20, 2009 from Brooklyn Daily Eagle (1841-1902) Online Web site’s Help page: http://ffx3.digitaleditions.com.au/Default/Skins/APAAGESMHDEMO/Help/archivehelp.htm

Grant awards. (n.d.).  Retrieved April 11, 2009 from the Institute of Museum and Library Services (IMLS) Web site: http://www.imls.gov/results.asp?State=0&city=&description=on&inst=&keyword=&program=gt_1007&sort=year&year=6

Grassian, E. & Kaplowitz, J. (2001). Information literacy instruction: Theory and practice. New York: Neal-Schulman.

High Performance Scanners. (2006). Retrieved April 15, 2009 from the Sunrise Web site: http://www.sunriseimaging.com/ckv.htm

Historic newspaper: Brooklyn Daily Eagle online (1841-1902). (2005, March). Journal of Academic Librarianship. Retrieved March 25, 2009, from Library, Information Science & Technology Abstracts database.

History and timeline. (n.d.). Retrieved March 15, 2009 from Brooklyn Daily Eagle (1841-1902) Online Web site: http://eagle.brooklynpubliclibrary.org/Default/Skins/BEagle/Client.asp?Skin=BEagle

Holland, J., & Benz, S. (2004, August 15). Unanticipated outcomes. Library Journal. Retrieved March 25, 2009, from Library, Information Science & Technology Abstracts database.

How the repository was created. (n.d.). Retrieved March 29, 2009 from the Brooklyn Daily Eagle (1841-1902) Online Web site: http://eagle.brooklynpubliclibrary.org/Default/Skins/BEagle/Client.asp?Skin=BEagle

OCLC annual report 2001-2002. (2002). Retrieved April 1, 2009 as a PDF from OCLC Web site: http://www.oclc.org/us/en/default.htm

Olive ActivePaper Archive. (2007). Retrieved April 30, 2009 from a PDF datasheet on OliveSoftware Web site: http://www.olivesoftware.com/products/activepaperarchive.asp

Patout, G. ( 2006). 2006 MAHD conference highlights. Retrieved April 11, 2009 from the Special Library Association (SLA) Web site: http://units.sla.org/division/dmah/dmah_36_4%20Summer%202006.htm

Spruill, D. (2006). The solutions in context.  Retrieved April 3, 2009 from a PowerPoint presentation on the Online Computer Library Center (OCLC) Web site: www.oclc.org/us/en/memberscouncil/meetings/2006/october/spruill_debra.ppt

Terms and conditions of use. (n.d.). Retrieved March 29, 2009 from the Brooklyn Daily Eagle (1841-1902) Online Web site: http://eagle.brooklynpubliclibrary.org/Default/Skins/BEagle/Client.asp?Skin=BEagle

Underlying technology. (2007). Retrieved April 30, 2009 from OliveSoftware Web site: http://www.olivesoftware.com/products/technology.asp