The Digital Dungeon

An article I had been working on for a bit with Heidi Winkler finally got published in the second volume of the International Journal of Librarianship in a special issue on Data Librarianship . Abstract This article reviews the influence that massive digital libraries like the HathiTrust Digital Library can have on local, smaller institutions’ digitization, preservation, and curation programs. The history of HathiTrust’s digital preservation efforts as a Trusted Repository is reviewed. A case study is presented showing how one academic library made difficult digital stewardship decisions in a modern world of globally federated preservation initiatives. The authors introduce the concept of deselection as part of the digital curation process and discuss how digital collection administrators can refine their local digital preservation efforts to better reflect the realities of constrained human and financial resources. Read more...

This month is all about self-evaluations for me and my employees.  Because of this, I have been thinking about how a manager is supposed to show their work and their worth. The easy answer is to say that if the employees are doing well, then the supervisor is good. It could be that the employees are doing well despite a bad supervisor. An employee doing badly is also not a sign of a bad supervisor. So what tangible thing can I say makes me a good or bad supervisor? Throughout the year, I try to focus on the actions I take to make my employees' lives at work better. I try to give them direction, advice, and help make things easier. I also try to champion them. Things do not always work but I adjust. When I sit down to write my own evaluation, though, I end up writing about documentation. To me, that is a concrete indicator of a good supervisor. They care enough about the work, and their employees, to write things down and make a record. I want to challenge everyone to write...

In the last few weeks, I was asked to write a small paragraph explaining my area to others in the library.  I was happy to do this, as many people say they don’t know what my people do.  It’s sometimes hard to explain to others what we do without going into overtly technical topics and terms.  If we have done our job right, we’re practically invisible, which is the way it should be.  Anyway, writing the description made me realize why there is often a mis-match between what we do and what people think we do.  I’ll let you read the description yourself.  I’ve underlined the important bit. “Digital Resources is primarily an Open Access publisher.  We publish both born digital items (produced by students or faculty), and we scan to publish or republish old items. We curate digital collections through the whole digital life-cycle. Our work is a bit different from other departments because the more work we finish; the more work we create in having to m...

I've been reading "An Emergent Theory of Digital Library Metadata: Enrich then filter".  I'm about 1/3rd of the way through, and so far I am convinced that libraries need to make it easier for patrons to add or suggest changes to metadata.  I'm convinced enough that I will add the functionally to my list of future directions for the collections I manage.  However, in light of the recent national conversation about fake news, I do question whether or not communities can effectively actively police incorrect content.  Wikipedia is used an example of how crowd generated information can work, but Wikipedia is also often the first search result in almost any search, meaning that it not only has a high chance to be seen, but also a high chance to be edited if it's wrong.  People online love correcting others. The problem I see about applying that model to library metadata is that there is almost no way for library data to be as popular as Wikipedia, so it will hav...

The Bureau of Indian Affairs is one of the oldest Bureaus in the United States.  It was established in 1824 by Secretary of War John C. Calhoun . While the history of the organization has been controversial, their records are open to the public.  This collection brings together letters distributed from the Bureau of Indian Affairs starting in 1832 and going on into 1966. View the rest of the collection:  http://bit.ly/2h0hKvW 

As I am looking around for the next digital content management system,  I decided to gather some usage statistics about the systems I'm most interested in. These are all just estimates as I have to rely on how many institutions self report their usage. The exception is CONTENTdm, which I had to go by the number they use on their marketing materials. CONTENTdm: Used by over 2,000 (number supplied by OCLC website) DSpace  : Used by over 1,500 institutions. Fedora Commons : Used by about 300 institutions.   Hydra  : Used by about 50 institutions. Omeka: Used by over 287 In looking at moving to Fedora Commons or Hydra, I was surprised by how few people are using Hydra.  I believe it's because it's notoriously hard to set up, which is why IMLS is funding Hydra in a Box , which should be a turnkey solution.  I'll be interested to see how the numbers change after that. We are also thinking of trying out Omeka, and it's good to see that Om...

1. No one really thinks about why the item is being scanned When people talk about scanning something or putting it online, their reasons for doing so can be shockingly vague.  People will say things like "to make it more available", or "to make it easier to search", but often what they really want is something completely different.  I have had a faculty member initiate a scanning project for photographs only to find out that what they really wanted was to blow up the images so he could see them better.  After the first few test items, and him complaining about how he couldn't see the images well enough, it became obvious what he wanted.  We changed the project to scan the images at a higher resolution to make it a bit easier for him. 2. No one thinks about what, exactly needs to be captured during the scan When people first start out, they think of scanning as a pretty simple process of just taking a picture of the item and putting it online.  However, the...

In the last week, I've been looking at how to make our Institutional Repository better.  One of the best ways we can provide value to our users is to help them identify the impact of the value of their work in the IR.  Along that line, I wanted to know how we could find and count citations to works in the IR.  It would be great if we could say that things published in the IR get cited more. I found this post in Stuart Lewis' Blog about displaying citation counts in DSpace from Scopus  , and in the next few months we may try to implement this.  However, this wasn't exactly what I'm looking for.  This shows the number of citations based on the DOI.  What I want is if someone cites the URI from DSpace, I want to be able to find that and record that somehow. I have not found an automatic way of doing it.  So far, what I've managed to do is do google searches for the first part of our handles, and the first part of our IR's URL. So, our IR's URL ...

ARL released an extended and updated version of their " Strategic Thinking and Design Initiative " report. While the report is a long 126 pages, I've provided a few highlights that I think could also serve as strategic directions for digital libraries: Recognized Shifts that are occurring -  Page 19 (PDF page 21)    Within two decades, the research library will have transitioned its focus from its role as a knowledge service provider within a single university to become a collaborative partner within a broader ecosystem of higher education.  Research libraries will be even more intimately engaged in supporting the full life cycle and activity range of knowledge discovery, use, and preservation, as well as the curating and sharing of knowledge in diverse contexts of the university’s mission and of society more broadly. ARL—the organization—will enable and catalyze research libraries to leverage and mobilize individual assets toward the collective advanc...

The conference this year was exceptional. The keynote address by Trevor Owens was particularly good and helped me at least understand what is going on at a national level for digital libraries and introduced the list of awarded grants for the National Digital Platform . Reading through that list is an eyeopener for seeing the tools and initiatives people are working on. Just some of the highlights: Memorandum of Understanding Workshop: Creating a Process for Successful Digital Collaboration - Brett D. Currier, Rafia Mirza, Peace Ossom Williamson We have already started using these at our institution.  The materials presented were exceptional. Envisioning a Geospatial Data Portal and Curation Network -Kathy Weimer, Douglas Burns, Joshua Been, Kim Ricker, Cecilia Smith They brought up an interesting point that a lot of free governmental data is location based, and GIS knowledge is really needed to allow access to it. The Training Games: A Method for Digitization Skill...

Book Liberator- DIY A simple book scanner with two digital cameras, a cradle, and a way to hold the pages down.

Someone asked me to provide some feedback about some equipment I had used. After I sent him a reply, he suggested I put this information on the blog. So, I’ll put this out there. This review is for the I2I CopiBook from Iimage Retrieval. Performance The Scanner performs wonderfully. It has an artificial limit of 300 DPI, but can go up to 380 DPI (once the software is upgraded). The color is very accurate. We’ve been happy using the CopiBook for books, and for maps. I know people that are using it for newspapers as well and I would not hesitate to use it for newspapers as well. Marketing I actually bought the machine based on a recommendation from someone else, so I never saw the machine marketed. The machine is well made, and works wonderfully with ambient light as long as there is no overhead lights. It works off of ambient light better than some lighted machines. I know that Iimage Retrieval is selling some extra lights to buy with the equipment if a suitable location can’t ...

Ultrafiche is an amazingly dense version of microfiche. The images are so small that they might be the size of some small fonts (8-10 size font is probably close). We had an ultrafiche project proposed, and soon realized that none of our equipment could do it. Because the images are so small, you need something that can scan at a much greater DPI, or zoom in closer than most equipment will do. However, it’s not impossible. There are two pieces of equipment that claim they can do Ultrafiche. I haven’t seen them work in person, but I might be looking into it in the future. Here’s the links to the two machines for further reading: ScanPro 2000 ($8000) ST2000 ($95,000)

I’ve been considering how to digitize and display objects, mostly because I have an object project that is coming my way. Our original plan was to put a 3D scan of the object on-line, but any 3D scanner that we could afford wasn’t going to be good enough, so I’ve been looking at different options. Option 1-Just pictures We could just take a few pictures of the object with a neutral background and good lighting. In this case we would treat the collection more like an image collection. Option 2- Flash application My husband and I were looking for cameras and we found this site for Olympus cameras . The cameras are ok, but what I was really surprised about was the display! You can click on the camera and move it around, seeing all different sides. Some of my colleges assumed that it was a 3D model, but I could tell that it was just a series of images taken and strung together using a flash application. At work, we are working on implementing this for at least our first object collection. ...

Toward Digitizing All Forms of Documentation Abstract Large-scale book scanning projects are delivering unprecedented access to the majority of library holdings by giving users unparalleled access to vast collections of books. However, these efforts have focused on typical bound books, which leaves many forms of documentation out of the digitization framework. We present techniques that we have developed to digitize numerous other forms of documentation, including deteriorated manuscripts and photography. These technologies are engineered in conjunction with budgetary and physical constraints often placed on digitization projects. http://www.dlib.org/dlib/march09/landon/03landon.html

http://www.maximumpc.com/article/news/patent_application_reveals_googles_mysterious_bookscanner I’ve had this link sent to me from a variety of different places. I wanted to share it here. I think I have been as curious as the rest of the world as to what exactly Google was doing. I didn’t expect them to be using infrared light. It should be interesting to see what happens with this.

Going Gray? Comparing the OCR Accuracy Levels of Bitonal and Grayscale Images by Tracy Powell and Gordon Paynter . Both from the National Library of New Zealand. The authors processed the same items in bitonal (black and white) and grayscale, then OCRed them. They then counted up the OCR errors. The theory was that Grayscale images would give better results because it gave the OCR software more information to work off of. They were surprised by the outcome. From their experiment, it seems as if there is no difference in OCR quality. This is interesting to me mostly because I had heard that bitonal images gave better OCR results.For a review of how OCR software works, there is a great explanation on Mustek.com: Understanding OCR .

The Joint Conference on Digital Libraries is in Austin this year. It’s an international conference “focused on digital libraries and associated technical, practical, and social issues”.

I haven’t updated in a while. I’ve been training students, and teaching the students to train other students. The lab that I run doubled the size of it's workforce in a month. We went from 18 people to 36. I had estimated taht, in a perfect world, we would be able to scan 128 books a day. After the last month, I don’t feel that this number is practical. Our lab has been able to reach 57 books finished (from beginning to end) every day with two Kirtas APT 2400 book scanners. There are many reasons why I don’t think we’ll get close to 128 books finished or scanned a day. The biggest reason is that our workforce consists of mostly students, and I have no guarantee that I can keep someone at the machines even 80% of the time. I can’t make transitions between users seamless, so there will always be downtime when one person leaves and another person comes in to replace them. But, 57 books finished a day is a good amount. That’s 1710 books a month.

I am involved in a project that has 14,000 books in it. The books tend to be relatively short, which means each one scans in about 15 minutes (on average). With two Kirtas machines, and the Atiz scanner to back up the machines by putting the items that won’t scan automatically on it, I have estimated that we could finish 128-158 books a day at only 80% efficiency (running 18 hours a day, both machines are being utilized 14 hours of the time- all week). That means, if we can make sure the machines are being used 14 hours a day, then we could get finished with our 14,000 book project in less than a year. 109 days at 128 books a day... So, four months. Originaly, we expected the project would take us 45 years. That's quite a difference. We haven't reached 128 books a day yet due to staffing issues. In the next month, those staffing issues should disapear, and we'll see if it really works in practice like it does in theory.