• Sean Quimby, Chair
• Meredith Evans Raiford
• Steven Mandeville-Gamble
• David Murray

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

[This was a fun, well-crafted session with three smooth, polished presenters.]

Using the International Brotherhood of Teamsters as a Case Study, will look at how private donations affect collecting.

Since the 1990s, federal and state funding of universities has dried up, so they “have to be in the development game.”  At the same time, philanthropy is changing.  New philanthropists see themselves as “social entrepreneurs.” They expect to see business plans, want tangible results.

Mission Creep: if you don’t carefully handle the obligations you make to donors, your programs can be distorted, bleeding away resources from core functions of your institution. It’s important to understand the goals your donor has, why she is at the table.  Make sure the donation is a good fit, that both donor’s and institution’s needs are met.  Donors have sued to regain  control of collections (and associated endowments) when the institution isn’t taking care of them , or meeting the donor’s goals.  Lesson: make sure both the letter and the spirit of the donor agreement are met.

GWU and Teamsters see themselves as partners.  Teamsters approached GWU about putting records on permanent deposit, with a $1 million endowment.  GWU had reservations: there’s no labor archives at GWU, no labor historian on campus; had run out of space in archives storage facility.  GWU is “an aggressively pro-capitalist” institution; its labor studies classes (not called that) are all about busting unions.  As such, SMG told Teamsters it was a poor fit

IBT asked for a proposal for creating the necessary conditions: labor archivist, historian, preservation endowment.  SMG wrote the proposal, but had tensions with GWU Development VP who wasn’t willing to consider IBT’s needs/goals.

Labor archivist is embedded in IBT facility 3 days/week, travels to local unions, works with IBT Education dept., as well as providing research and reference service at GWU labor archives.

Lesson: By not just taking the $1 mil, GWU built trust with IBT, which led to much better relationship.

Two articles about “new philanthropy” worth reading:

• Lisa Brower: Paving the road to Hell? Cultural institutions and the “new” philanthropy, RBM {journal} 5 :1 3/20/04
  • TK Reis and SJ CLohesy Unleashing new resources, New Directions for philanthropic fundraising 32

Two cases:

1. GWU has a family collection of 40,000 rare books; it’s a good collection, widely used. However, the University didn’t use the endowment properly, spent money on related items for general collection.  As a response, the family bought new books for collection and put them on deposit rather than donating the books.  Family also gave money for staff salary. Development asked family for $1.5 million to endow curatorship.  Things got ugly between family and development office.  Meredith and Steve contacted the family on their own, built relationship based on genuine interest, not just fundraising.
2. GWU received donation of electronic and paper pollster data.  Pollster data didn’t fit into GWU collection, but they took it and made the data available online. Donor gave money to University to create a society of pollsters and to digitize data: the library was not at the table during the discussions.

Lessons: Ask, Write and Follow-up.

1. Ask: It doesn’t hurt to ask for money  When you ask, consider the donor’s needs, goals
2. Write: Identify the needs of the special collections. What collections should you acquire to improve the special collections as a whole? What do you need for resources to process/digitize new collection? Identify the donor’s needs. Then write the proposal yourself. Work as a partner with your donor
3. Follow-up: Your donor should get a thank you note from you, before she/he hears anything from the development office.  Keep in touch with the donor on a personal level.  Don’t let development office bombard the donor with requests.  Make sure the relationship is open.

The donor is the center. The goal is to help the donor accomplish the change they want to see in the world.

Our job is a different kind of CRM: Charitable Relationship Manager.  We are servants of an institution that existed before we got here and will exist after we’re gone.

“It’s a privilege to help people try to change the world… The best part of my job is going on asking people for money. The worst part is sitting in the office trying to get the money spent… A lot of donors are people who’ve spent most of their lives making money.  I get to sit on a yacht and help them figure out how they want to be remembered.”

Goals are quantifiable (number of prospects/contacts/proposals, amount raised). Work is not quantifiable.

Case study: Plastics Pioneers Association Archives

“Harry,” the donor, approached Syracuse: I’ve got all this plastic stuff in my garage.  I want you to build a website about it.  SU sat down with Harry and his friends to learn about what their needs and goals were.  They built a prototype website for him, he liked it and gave $10k.  Then, more important, he gave them information that led to acquisition of National Plastics Center and Museum collection.

Sean Quimby raised the question: what implication does this have for the rest of the library’s special collections?  Lots of interest from faculty across disciplines were interested in the plastics collections.

What you can (must) do:

1. Do what you promise (that is, spend the money!)
2. Be enthusiastic and optimistic
3. The donor always wins
4. Don’t take any crap

Question: [For David] Could you explain what you mean by “crap”?

David: We never say “We have no room.”  That’s a bad explanation.  We say instead, “We only take things we really want.”

Question: What do you do when development office or faculty accept gifts that you don’t want, without asking you?

David: Don’t let it happen again.  It’s unethical for development to do that.

Meredith: I talk to the donor and recommend other places, find a better home for the donation.

Sean: Faculty don’t think archivist should presume to select and appraise!

Question: How do you get a place at the tablet so the archivist’s voice is heard?

Meredith: I make a lot of friends on campus – get to know people, take them to lunch, get people interested in and aware of the special collections . Encourage teachers to bring classes in, have events.  You have to build relationships with faculty.

David: Our biggest donors are the campus – the provost, the faculty – and we yell at them.

Question: How do the processors (who don’t have time to process any more) communicate with curators who want to bring in more stuff?

Steve: As AUL for collections, I make sure we don’t take in collections that we can’t process and don’t even have a strategy to process.  There has to be a plan that takes into account special collections’ needs and capacity.  Demand resources for processing from head of library, or whomever.

Meredith: Know what you need, give it in writing to the curators.  Let them know how much money you need to process, what these things cost, how the money will be used.

• Jeanne Kramer-Smyth (Chair) Building Archives Websites That Google Will Love
• Matt Herbison Online Collections Crawlability for Libraries, Archives, and Museums
• Mark A. Matienzo Findability in the Flow: Discovery through Linking

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

If every page on your website has the same name, it’s like a box of chocolates – your user never knows what she/he will get without biting into each one (or drilling down into each page)

Keys to good (for SEO) page construction:

1. Unique page title, fewer than 65 characters
2. That title should have a clear keyword target phrase
3. One URL per page
4. Unique meta description, fewer than 164 characters
5. Content-rich text – not a big image with a few words, but a big (or small) image with a paragraph of explanation

Analytics

1. Where did they come from?
   1. Traffic sources and keywords
   2. When and why do they leave?
      1. Determine your bounce rate (visitor leaves immediately) and exit pages
      2. Where am I on a Google search?
         1. There are lots of good tools, among them Google Webmaster tools

Next steps/action items

1. Look for the websites that rank higher for the phrases you care about.  Do they have better material or are they just playing the game better
2. Look at analytics
3. Bake SEO into discussions of web site (re)design.  Think about SEO ahead of time
4. Follow some SEO blogs, such as SEOmoz Daily SEO blog

[Matt tweeted each one of his slides during his presentation.  Twitter scheduling client?  Brilliant!]

The goal is crawlability

If web crawlers can get to your stuff, your stuff won’t turn up in search engines, and users wont find your stuff

Search engines don’t search (query) your site.  They can only browse or click links.  Content that’s locked in databases (Deep Web/Hidden Web) can’t be searched

Questions to ask about your website:

1. Does it block robots?  Robots.txt tells web crawlers which part of your website to ignore, NOT to crawl. A lot of websites have old versions of these files, inadvertently blocking crawlers from seeing the good stuff.  Example: Minnesota Historical Society robots.txt blocks directory /VisualResources.
2. Using sitemap.xml.  This tells web-crawlers what to look for, what pages to index.  Its like a traditional sitemap for human users, but much more detailed
   1. Functions in a similar way to OAI-PMH
   2. Vanderbilt Television News Archive has more than 800,000 records, most of which have summaries.  The sire has nested branching by month and year.  So top level sitemap points to each year/month sitemap, each of which in turn point to those 800k items
   3. Long URLs? Long URLs contain parameters for search, but too long makes it hard for crawlers to crawl
      1. Resources on NARA website are a “nightmare of crawlability”: no browsing, enormous and unstable URLs, but the URLs don’t need to be that long
      2. Can you use short, permanent URLs (or permalinks)?  If you have a database-driven system, each item probably has a unique identifier that you can use in the URL
      3. Is your content hidden in fancy features (such as Ajax, JavaScript, Flash)?  Then it is hidden from crawlers too
         1. Don’t let anyone sell you a Flash-only site.  It looks beautiful but content is hidden
         2. Do people get trapped too far in your site?

General approaches for beating the system

1. Incoming links highlight resources wherever they appear in your site [unclear how this works]
2. Get to know Google Webmaster Tools!

Why are we interested in SEO? Because we want to be found

The difference between Search and Discovery (Marchionini 1995)

[Good bullet list of differences.  Full article here: www.ils.unc.edu/~march/getty.pdf]

Discovery happens elsewhere (Lorcan Dempsey) – people are going to learn about your collections and resources on places other than your site

Go where your users are

URLs are the currency of the web (L Dempsey again): URLs and the pages themselves can be shared by and among users for information, for influence and for good will

Think about how links impact they way people find your resources.

1. Look for backlinks – that is, find out where people are coming from.  Google analytics is useful for showing this
2. Follow your nose: how do you link out to other resources? How is your information related to other pockets of information on the web

The Facebook “Like” button is more than a silly thing folks can do on FB.  You can add a Like button to any page so users can easily link to your site on FB.

You can “Like” real things.  The Open Graph protocol, designed for pages about real world thing

Wikipedia: there are legitimate concerns about how reliable the content is, but it has proven to be a community that is responsive to people’s concerns.  People often complain “Wikipedia is not authoritative. It doesn’t cite sources!”  The Wikipedia page on Yale University cites over 114 references.  Categories in Wikipedia allow editors to group like sites.

Know your community: Wikipedia has standards that you should know and follow. See policy on “Subject and culture sector professionals,” which encourages us to improve Wikipedia by adding information to pages and linking to our sites from Wikipedia

Linkypedia new tool to see how/where your site is linked on Wikipedia.  [Wow!]

Q&A

Question: If you are in a small shop with minimal IT support, how do you prioritize what to do (if you can only do 1 or 2 things)?

Matt: Do two things first: (1) Add intelligent, respectful links to your site from Wikipedia and (2) improve your sitemap.xml

Jeanne: Fix your page titles

Question: Two comments:

1. Google can access a well-built Flash website – it is challenging, but it can be done
2. Making a website usable and accessible to users with disabilities will make it more crawlable

Jeanne: Adhering to Section 508 will help your SEO, and improving SEO makes your site more usable for the disabled

Question: What about non-HTML content, such as wiki pages, on your site?

Jeanne: Ultimately it’s all HTML (except for images), but wiki pages are great for SEO because they are so well-structured

Mark: The web is getting more data-intensive, so we want to have multiple representations of our date, such as both human-readable and machine-readable.  If you can provide information as structured data, it may allow you (and enable others) to do more experimental things

Question: What about CONTENTdm?

Matt: Item-level search engine hits often come from known-item searches.  Any time you can aggregate about that item-level, you are more likely to be found. For example, CDM default is to show 20 items on a page.  If you can group those 20 items [with a collective description?] that group is more likely to be found, and then users can choose the individual thing they want.

S.402 Archivist 2.0: Policies, Partnerships, Predictions and Possibilities

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

1. Nan Rubin, Preserving Digital Public Television
2. Lisa Schmidt, It’s Different with Digital: Influencing Curation Policy When There is No Mandate to Curate
3. Joanne Kaczmarek, From Here to Infinity! The Future of Archives

Nan Rubin, Preserving Digital Public Television

[I had laptop issues and didn’t take notes.  Here’s the link to the site she discussed]

www.thirteen.org/ptvdigitalarchive

Lisa Schmidt, It’s Different with Digital: Influencing Curation Policy When There is No Mandate to Curate

Archivist as policy maker:  “Stop disciplining data and start herding it” (Bailey?)

Archivist as policy maker: take responsibility, speak IT language, and play consulting role

At MSU, the Archives offers guidance and creates policy on a number of projects

1. Digital Curation Planning Project: Assess digital assets, tech infrastructure and state of digital curation needs and practices
   1. Began with questionnaire, followed up with in-depth interviews to assess current state and offer recommendations
   2. Identified 4 types of content:
      1. i.      Publications
      2. ii.      MSU records of historic interest
      3. iii.      Research data
   3. Archives can provide advice for all content types, and has policy enforcement authority for some types (such as publications)
   4. Archives assists University Relations office, which has servers bursting at the seams with digital photos and video
   5. Spartan Archive Project: NHPRC-funded (3 year grant) to develop a permanent electronic records archives; beginning with three electronic records series from the Office of the Registrar, then will extend to other university records of enduring value
      1. MSUA does have a mandate to collect this
      2. Records Management Training: University Archivists makes presentations and leads workshops on best practices for managing electronic records
      3. Tiered Storage Initiative at MSU: Archivists provides guidance, ensure trusted digital repository space not just cheap storage. Will connect with records retention schedules, so folks can throw whatever they want into storage

Conclusion: we have traditionally been stewards of records of enduring value, so we can do this for digital records too!  We can be effective consultants because we are part of the institution but not invested in processes at the unit level.  Get to know your records managers.  By taking on this policy making role, the archivist becomes more of a strategic asset to her/his institution.

Joanne Kaczmarek, From Here to Infinity! The Future of Archives

JK’s motivation: to break out of her own world view and freely imagine future possibilities.  Hasn’t had a lot of opportunity to work with physical records, but has a growing appreciation of them.

JK worldview: technology will continue to influence how we manage information.  The temperament of archivists is (rightly) difference from that of technologist.  All archivists do not need to become technologist; we just need to be able to speak with them; it’s OK to be who we are.

To get an idea of what the future will be like, JK talked to young people (the 20 and under crowd) because the future is theirs.  A lot of people under the age of 20 don’t have a concept of “the wet signature,” don’t rely on paper documents as evidence/representation of past events.

What are their views on technology and on archives?  JK conducted a survey of 10-25 year olds.  Used Survey Monkey and her 15 year old niece, who said, “Don’t ask a lot of questions Aunt Joanne. They aren’t going to answer them.”  43 respondents, most of whom where friends or knew Jennifer, JK’s niece – mostly from the Midwest, but a few from around nation and world.

Which technologies do you use daily or weekly?  Facebook, phone use is high; Twitter, letter-writing low.  How do you store your digital media?  40% said “personal computer.” Only a few used print or web.

How do respondents see technology in their grandchildren’s time: One said that technology will impair their grandkids’ social skills.  They expect digital content, representations of their lives, to endure for grandkids.  Automation will require us to leave our homes less and less.  “A chip in people’s brains that allows them to project whatever they want to a computer”

What is an archives?  “A place for old stuff, old historical stuff, maybe on a particular topic.”

In summary, technology is not going away; people will be doing more to keep stuff themselves, but will rely on “safe” places; archivists shouldn’t lost sight of our foundational perspectives, but we don’t have to do it alone.

[JK then showed a video she'd made, highlighting some of the themes of her talk]

Q&A

[Didn't hear the first question. ]

Discsussion between NR and JK about rights and whether the sounds and images in JK’s video were infringing or protected by fair use.

Question for Lisa: could you give an example of paper-sparse office?

MSU business offices are looking to use less paper, not go fully paperless.

[Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

There was an interesting discussion from three panelists, during which I took no notes.  Here are some notes from the Q&A]

Question/comment from Joanne Murray of Drexel College of Medicine Legacy Center: There is money out there for marketing and assessment of your educational programs.

How do you learn how students are learning? Video games, social media?

NY State Archives: [missed this]

Michigan: try to incorporate 21^st learning skills

LOC: we focus on inquiry-based instruction because that’s what teachers are interested in.  We also track what state standards are addressed by our materials [may have misheard].

NYSA: using video games for learning – the games look good, but feedback from students and teachers was that participants (also?) want copies of documents that they can hold in their hands

LOC: Social media – send out an RSS once a week or so, in the process of starting  a blog

NARA: RSS feed for document of the month, uses Facebook

Q: How can archivists make Teach American History  repository visits more meaningful for teachers?

Library of Congress: we work with our curators to bring “the real stuff” to teachers.  Break down LoC’s resistance to working with teachers.  Ask teachers about their theme/focus to ensure that the materials they see are relevant to them.  We go to TAH conferences to see how it works and do outreach

Michigan: observed TAH workshop, “it was like watching paint dry.” Instead we get them out of doors, move them around, look at historic structures after studying the documents related to those structures.

Library of Congress: teachers love to move around, they hate to sit, they hate being lectured to

NY State Archives: remember that teachers may not know what an archives is.  If students are learning skills (e.g., map reading), teach them the skills not just show them the content.

NARA: takeaways are important.  We take TAH participants into the stacks.

Question: To what extent have your programs been successful (if it is even a goal) at considering the needs of teachers?  Not just showing teachers how to use systems, but to improve those systems for them.

NARA: huge priority.  We don’t just put up the documents we think should be online, but find out what teachers want.  NA has seen a reduction in onsite researchers, is focusing on online users

Library of Congress: when we first put up American Memory, we surveyed its users: teachers were the #1 user group.  Some staff “don’t understand why we are dealing with these kids.”  They are future users.

Michigan: we understand that kids are the key to the future.  Our 2 main user groups are teachers and genealogists.  We try to find materials that cross over, so the things we do for genealogists will add value for teachers also.

NY State Archives: trying to alter the vocabulary for search, tagging things in ways that users will understand.  We make it easy for archivists to find things, but do we use vocabulary a 13 year old would use.  Unfortunately, we do not consider educational programs in our digitization plans.

Q. Do you collaborate with archivists at colleges that train teachers?

Library of Congress: just ask us.

Q: Person from the IMLS recommends CHNM site called “history matters.”  “Very readable and fun.” http://worldhistorymatters.org

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.  This is a short post as I lost battery power midway through this final session, which was a shame as the last two presentations were both very good  and very interesting.

Works at a small institution that has little funding, particularly for digital preservation, and limited IT support.  Currently in place: Archive-It, CONTENTdm, local storage, and OCLC’s Digital Archive (for remote distributed storage).

Approach was to look at what others were doing, identify how others were transforming the same type of files that SLNC had (about 20 file types).

Tools needed to be open source, easy to use (GUI not command line), and versatile.

Transformations need to have minimal loss of quality or degradation, [other factors, moving fast]

No tools to convert .mov to .mj2, so they tried converting .mov to .avi, and there was quality loss.  Fewer problems converting image files, though there were some issues.  Converting text files didn’t produce “any noticeable loss of content,” but did lose formatting/layout (e.g., tables).

Most free OSS tools required some use of command line.

Or an alternate title:  “How I got pneumonia while failing to herd cats.”

Research builds on JD’s PhD research on repository copyright practices and their impact on users.

Research plan was to identify main issues of concern to users (historians and genealogists) through focus groups and follow=up interviews.  Recruited though listservs, reading rooms, fairs, societies.  Getting even one focus group together proved impossible. Conducted 16 telephone interviews.

Preliminary findings: users don’t distinguish between archives and other sources; most of them associate copyright with citing sources properly.  Feelings about copyright range from fear of lawsuits to willful disregard (“If it is something I want, I am going to take it.”)

Response to technical measures (such as watermarks) ranged from using image with watermark, removing watermark from image, looking elsewhere for clean image.

For further investigation: methods of collecting robust data from users regarding their copyright practices; what copyright information should we provide to our users? How and where should we present that info to users in the “unmediated online environment”?

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

Preservation recommendations – all the technical papers are on the IPI website

Other resources: glossary, newsletter archive, recommended reading

DP3 Newsletter: quarterly update that also includes research findings, news and events

Imagine embedded metadata as a jello salad [image].  What’s the benefit to embedding metadata in image files – both short-term (use) and long-term (preservation)?

Allow curators to more efficiently collect metadata from faculty and students; make metadata more interoperable

http://metadatadeluxe.pbworks.com

Workflow tool, with basic input panel in Photoshop

Allows image creator to add contextual data, which they might track in a notebook or spreadsheet, to the system for embedding in the image.

Next steps: continue to work on these workflow tools; use VRA core for metadata schema;

“What do I mean by malleable?” A tendency not to return to original shape and not to break.  Allow users to come up with additional presentations of transformation o finding aids, trying to make something better suited for a use case.  Leverage use knowledge and contributions.  Use off-the-shelf wiki technology.  Not too worried about EAD now.

Implementation of a Commons.

Archives/Guides are non-rival goods.  Using guide “A” doesn’t preclude using guide “B.”

Established a finding aid space in wiki. Developed prototype layout.  Added in elements of functionality.

Finding aid wiki has lots of features familiar to wiki users: page history, export to Word/PDF, tagging, auto-generated table of contents.  New things: adding Google map to biographical note.

In the inventory, default arrangement is by date; you can sort list by folder type.

There’s a bibliography of works, which students or whoever can build.

http://wiki.millersville.edu/x/KANV

Screenshot: a policy is a set of rules

A policy can be applied to a record series

We can use any client, even a simple drag/drop client to ingest records.  When record is moved the policies that have been set up for it are enforced

Global priority: sustainable digital preservation and access

How much information is there?

ISO/IEP activities, 2009-10: industry collaboration workshop, international workshop

Educating stewards of public information in the 21^st century (ESOPI-21) – prepare the next generation of public information stewards by building on existing dual degree program at UNC-CH, including experiential component where students work in state or government archives

Interviews with advisory group, a theme has emerged: digital curation governance. What does effective governance require?  This is the kind of question MPA students ask. These stewards ensure the quality, interactivity, etc of public information.

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

• Ann Holt (Pennsylvania State University)
• Corinne Rogers and Elizabeth Shaffer (University of British Columbia)
• Sam Meister (Digital Archive of the Birth of the Dot Com Era Project) and  David Kirsch (University of Maryland)
• Mary M. Manning (Texas A&M University) and Judy Silva (Slippery Rock University of Pennsylvania
• Gordon Daines (Brigham Young University)
• Ken Thibodeau (National Archives and Records Administration)

Access, Communication, and Learning: Physical and Digital Potentials for Knowledge-Making in Archives

Ann Holt (Pennsylvania State University)

“I am a Doctoral student in art education at Penn State. You are in the right room.” Used knowledge of the history art education as archival processor at Penn State. Became interested in archival research process itself, fragmented communication between researcher and archivist. Communication about research is vital to learning.  Last November, assisted grad course on history of art education: 5 grad students who had no archival research experience.  Goal: cultivate an awareness of the value of archives. Created a secure blog which students could use to share and discuss digitized images from the collections.

After initial class visit to archives, most communication took place on the blog. It provided a place for students to post photos they took (mostly with camera phones) and for AH to present information about doing research (e.g., posting finding aids, linking to Special Collections web site).  Benefits of blogs: author-generated categories and tags. Students requested material through blog posts, so classmates could see what they were using.  At the end of the class, assessed what students had learned about archives, how their attitudes about archives had changed

Professor liked that blog created a research community, multiple perspectives.  Students thought that blog transformed individual research into “a community of sharing… a place to voice ideas.”  At the end students hosted a conference and added podcasts of conference to the blog.

How to preserve a million digital records when everyone has a million other things to worry about

Corinne Rogers and Elizabeth Shaffer (University of British Columbia)

Background: research conducted as third phase on InterPARES project

Key findings:

1. there is no principle of original item in electronic records
2. preservation can begin at creation
3. [missed this]

Site: BC institute of Technology, largest post-secondary institution in British Columbia.

Goal: create policies that fit with existing records management rules, and with organizational culture so they would be implementable.   BCIT is [was?] a paper-based organization with centralized paper records despite multiple campuses.  In the paper-based model, policies, procedures, retention were applied to paper records, which were sent to the archives for preservation.  In the digital office model, IT department would provide storage for digital records, ad hoc.  Relationship between RM and IT was not hostile but not cooperative.

Institute faced issues of privacy and security, FOI, electronic discovery.  The archives was “temporarily” closed and the archivist let go, with duties going to RM office.

Methodology to develop policies and procedures

1. Analyze recordkeeping environment
2. Draft policy for long term preservation
3. Interview stakeholders
4. Analyze interview data in order to understand what was actually being done (in comparison with what procedures said should be done)
5. Drafted procedures to accompany policy

Sherwood Archive Project: Preserving “at risk” private digital business records that are of public interest

Sam Meister (Digital Archive of the Birth of the Dot Com Era Project) and  David Kirsch (University of Maryland)

The record of business at risk.  Three elements of the business sector threaten record keeping:

1. Startups are worried about staying open, not documentation
2. Nature of shareholder capitalism demands that managers act in ways that benefit shareholders.  Harder to convince them to keep records
3. Once private value of records is no longer worth the cost of keeping, businesses destroy them.  IT makes this easier

Ways to address these challenges: treat records as an asset of bankrupt companies.

Dot Com Archives works with Sherwood, a company that buys assets of troubled businesses

Methodology: collection assessment, interview key Sherwood staff, and field observation (including site visits to failed firms).

Interim results: produced selection criteria for the records of failed businesses.

Environment is highly variable in terms of time and access. Sherwood staff may have months or only weeks to assess and acquire records of failed businesses  Access issues are common to electronic records, but have new twists – e.g., IT staff who maintained records may have moved on from failed company to new jobs,

Next step: create partnership between Sherwood and storage facility, which maintains confidentiality of private data. The balance of privacy and providing access

The Blended Archivist: Balancing Contemporary Roles in American College and University Archives

Mary M. Manning (Texas A&M University) and Judy Silva (Slippery Rock University of Pennsylvania

Surveyed college and university archivists, sorted them by institutional size.

Hybrid archivists: more than half of archivists surveyed had library duties

Archivists without library duties spend more time on archival duties than do those with library duties

Hybrid archivists were more likely to be tenure track, more likely to conduct research/scholarly writing

Hybrid archivists self-identify as (in descending order):

1. Archivist
2. Archivist/Special Collections Librarian
3. Archivist/Special Collections Librarian/Other Librarian

Identifier #3 was most common among archivists from medium-sized universities.

Top responses to the question What are the principal challenges? were similar among hybrid and non-hybrid archivists.  Hybrids worried more about quality (doing lots of things but none of them well)

Top 4 responses to the question: what are the principal benefits? were similar among hybrid and non-hybrid archivists.  They differ in that hybrids are more likely to benefit interaction with faculty and students, while others mentioned knowing the institution, increased visibility

Re-engineering Archives: Business Process Management (BPM) and the archival quest for efficiency

Gordon Daines (Brigham Young University)

How tools from Business Process Management (BPM) can help archivists improve archival business process, particularly backlogs.

What is BPM? Improving business processes to be more efficient and make customers happy.  BPM for archives means we serve patrons better.  BPM is holistic: is looks at the entire business process rather than sub-components.

A simplified version of archival business process: 1. Select/appraise, 2. Arrange/describe [missed last two]

In order to redesign appraisal activities Perry Special Collections created a process model of relationships. Shows both existing process and proposed improvements

Use cases include graphical and textual  descriptions of relationships and activities

Original Order: Lost In Cyberspace?

Ken Thibodeau (National Archives and Records Administration

Will describe a current research problem to get feedback

About 6 years ago, when talking with staff about arrangement, KT was challenged by a colleague who said arrangement of electronic records is unnecessary, since it doesn’t matter where you put them.  KT dismissed this at first, then considered that  where arrangement was once a solution to a problem, it’s now a requirement.  We can create virtual filing cabinets in cyberspace, but this is self-limiting.  Original order express links among records, the way they are created and used – an archival bond, an organic relationship.  KT: perhaps we should talk about archival links not an archival bond.  Arrangement  is a method of preserving original order, but only part of it.

[great graphic showing archival relationships that I can’t explain or summarize]

There are tools for analyzing the content of many documents based on content analysis. Also, tools for analyzing communication, interaction.

Conclusion: let’s preserve as much as we can of the archival bond (or archival links) including processes of record creations, subsequent uses o records, their content, their creators (the actors who produced them)

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

Persistent Digital Archives and Library System (PeDALS)

Richard Pearce-Moses (Clayton State University)

Research goals/questions

1. Curatorial rationale
2. Test middleware
3. Can we stop dealing with items and start dealing with chunks?
4. Evaluate LOCKSS as the basis of a storage network. It was designed for serials collected by many institutions.  Concerned about its capacity and performance
5. Build a community of shared practice that meets the needs of a wide range of repositories

Secondary goals

1. Remove barriers to preservation by keeping costs as low as possible,
2. Investigate the use f open source software

First lesson: don’t start a major research project in the middle of a recession

Fundamental assumption: majority of records in agencies recordkeeping systems are not archival

Metadata is frequently limited – varies widely in quality and quantity

Results

Curatorial rationale

1. Describe before ingest not after
2. Data wrangling: preprocessing to normalize records
3. Ingest into LOCKSS

How much variation would there be in the workflow and business rules?

1. Workflow requires some tweaks
2. Preprocessor needs to change some record formats (such as turning .pst to XML)
3. Metadata mapping – different records require different metadata

Archivists work with rules not records. Some archivists struggle e with is

Test middleware

1. Rules and workflow can be automated
2. Code can be reused for common records series and also for similar records
3. Were successful at normalizing and enhancing metadata
4. Test BizTalk – it’s good for many small xml files, nit for large digital objects. It doesn’t serve as integrated workflow platform.

Community-based Development of Preservation Services

Chien-Yi Hou (University of North Carolina, Chapel Hill), Richard Marciano (University of North Carolina, Chapel Hill), Caryn Wojcik (State of Michigan)

DCAPE: Distributed Custodial Archival Preservation Environments (dcape.org)

There’s a session about it on Saturday morning at 8 a.m.  It is a partnership/consortium of six state archives and many cultural institutions.  Members are contributing records and advice.  It is developed by archivists and funded by NHPRC.  What is unique is taking generic middleware components and creating an integrated design that works for archivists.  Middleware was developed for large scale visualization of scientific data.

One major goal of DCAPE is to implement a design that provides essential information for archivists.

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

Archival quality is an archival value along with provenance and integrity

As early as 1939, AQ was a distinguishing characteristic of archives

The research environment is that we “take what we can get”

Hathi Trust: 27 partners, 6 million volumes

Google hysteria pervades environment, partly because Google isn’t telling us that they are doing

Hathi Trust is focusing on preservation, so the user interface isn’t very good yet.  (Screen shot of Hathi Trust vs. lousy Google books OCR)

Right now, archival quality in large-scale digitization is material-centered, one-size-fits-all, compromise is failure.  We are moving toward a user-centered model, where “good enough” is a value and users can “vote with their feet.”  [Move to a better service?]

Use cases

People are reading things online, but print-on-demand services are widespread – possibility of preservation/access copies

“Non-consumptive research” – a strange concept wherein you don’t actually read the text, just process the data.

Validation: we need two things that don’t exist now – (1) measurements and (2) validation of those measurements from the user’s perspective

Traditionally, quality is a bar over which all things must pass.  If it doesn’t pass the bar, it fails.  The only debate is, where is the bar?  Now, we want to redefine quality as absence of error relative to use.  We don’t have quality metrics, just personal reactions (and blogs that discuss those reactions).

Google provided quality data to University of Michigan but it is not publicly available.  Michigan is measuring errors in its Hathi Trust material. Amazingly low level (< 1%) in scanning.  There’s a 2% error rate in post-scan work.  Low but when it’s bad, it’s really bad.  It’s bad if it’s an error in the book you want.  Collectively, not so much

Two examples of scanning errors in Google books project: warped page, thick text, source crop, scan crop, blurring (either in source or in scan), thumbs on the page, bad digital cleaning (erased text).

Michigan has created a 17-part, 3 tier error model:

1. Data level
2. Page level
3. Volume level

Use this model to estimate the error incidence in various clusters of Hathi Trust content.  We must convince folks that statistical sampling is a viable way of measuring truth. Manual review QC is incredibly difficult.  Goal is to identify the limits of manual review and identify potential for automated processing of quality review.  Another goal: mechanism for branding items to show that they’ve passed QC test.

Long-term implication of research: theoretical research into digital pres is part of archivist’s job.  Have the surrogates become the de facto archive?  When it’s all digital, what happens to the original books, documents.

Q&A

Question: Any patterns among errors found during quality control? E.g., length of books, front or back of book

A: Michigan’s process was to sample 20 consecutive pages, to find human error.  Google is highly experimental with its post-processing, so the QC helps Google check its post-processing algorithms. We hope to dig more deeply into the relationship between problems with the source and problems with the scan.

Disclaimer (borrowed from David Weinberger): Live-blogging. Getting things wrong. Missing points. Omitting key information. Introducing artificial choppiness. Over-emphasizing small matters. Paraphrasing badly. Not running a spellpchecker. Mangling other people’s ideas and words.

AVC was the only archivist member of the Task Force

Where the task force is now: concluded work in February 2010, trying to move the conversation from the BRTF to the broader community.

Trying to create a sense of urgency: “I worry about these groups that come together and do all this work” and then no one pays any attention.

4 topics:

1. Research data
2. Scholarly discourse
3. Collectively-produced web content (e.g., Wikipedia.  “We tried not to go there” but had to.)
4. Commercially-owned cultural content

In order to have access to information tomorrow, we have to pay attention to preservation today

So what do we mean when we talk about the economics of digital information?  We often focus on the technical issues, but the question of the economic problem raises some hard questions to the community at large.  It really comes down to: “Who’s going to pay for it?”

Content now gets paid for through federal grants, ads (Google), subscriptions, donations (Wikipedia)

The Task force was charged with analyze evaluate best practices, make recommendations.
Deliverables were an interim report in December 2008, and the final report in February 2010 Interim report focused on current practices, majors themes and systemic challenges. Final report was a structural analysis of the 4 common scenarios, with recommended actions and next steps.  An invitation-only symposium was a conversation about those 4 areas, attended by smart people

What does economic sustainability require?

1. Recognition the benefits of long term access and preservation
2. incentives for decision makers to act
3. means of selecting the information we want to preserve
4. mechanisms to support allocation od resources
5. appropriate organization and governace

What are the roadblocks to digital preservation

1. long-term preservation is funded by short-term allocations
2. there is  alack of alignment between stakeholders, roles and responsibilities.
3. Inadequate institutional/community incentives

Alignment challenges: there are many stakeholders, we will only succeed if we align their interests

Three big challenges:

1. We must prioritize the preservation despite other pressing short-term needs
2. We are complacent that current practices are good enough.
3. We can’t be scared that the job is too big to take on.  If we don’t do it, who will?

Task Force recommendations on the 4 areas of focus are spelled out in final report

For near term action, we need to build organizational partnerships, invest in the capacity to support stewardship.  We need to create a public policy that sustains and encourages digital preservation

Some actions since report was issued:

1. NSF requires recipients of new grants to have a data management plan in place
2. LOC is creating a National Digital Stewardship Alliance (this month)
3. University of  Toronto is building a new digital curation institute

Recommendations for research

1. Selection of digital material.  We don’t know what makes sense for long-term preservation.
2. We need to know more about funding models, how to leverage economies of scale to lower the costs of preservation.
3. What are the stakeholder

Final Report (pdf) here; other Task Force info here: http://brtf.sdsc.edu