All posts by Henry Hebert

CDA Episode 2: Prepping for Ingest

In the first blogpost of our series introducing the Conservation Documentation Archive (CDA), I wrote at length about the needs for creating documentation during our work and the various forms in which it takes. Unfortunately, ingesting those materials into the Duke Digital Repository was not as simple as migrating digital files from one platform to another. A lot of work went into preparing each record, and today I’d like to share some more of the details and considerations that went into that process.

Our “hard copy” documentation has always been stored in a filing cabinet, organized by fiscal year, with the files of each treatment contained in a paper folder. As record keeping practices and technology changed over the years, we began producing our reports and images digitally and saving them to shared network storage provided by the university. The digital files are similarly organized in folders by fiscal year and treatment number.  

Preservation of our born-digital records has long been on our minds: it’s imperative that conservation treatment documentation persists for use by future custodians of the item (see AIC Guidelines for Practice #28 and commentary). Electronic records require regular, proactive maintenance in order to remain usable over long periods of time. The shared network drives are regularly backed up, but the integrity of the files are not checked as part of that process. If files were to become corrupted or deleted, we might not know for some time and our records could be lost. 

Documents that are stored in a human-readable format, on the other hand, have advantages for long-term preservation. Their storage requirements are pretty minimal (a cool, dry place), and they require no active maintenance or special equipment to access them. AIC’s Guidelines for Practice state that if the maintenance of electronic documentation records cannot be ensured, the conservation professional must create and maintain hard copies of that documentation using the most permanent materials available (see Guideline #28, Commentary B). The adage often used in libraries and archives is “Lots of Copies Keep Stuff Safe”, abbreviated LOCKSS.  

American Photocopy Equipment Company (1963)

Since we already have a method for storing hard copy records, we’ve been printing representative copies of our digital records and storing them in the filing cabinet. The reports are black and white electrostatic prints from our office copier (see Batterham, 2008). We also select a small number of representative before and after images, arrange them four to a page and then print them with pigment-based inks on Epson glossy photo paper using an Epson SureColor P800 inkjet printer

The 35 mm slides in our legacy documentation are a bit of an exception. While they can be viewed with just a light source and magnification, special equipment is required in order to work with them easily. The colors of those slides have also shifted over time – and will continue to do so (Weidner 2012). “Benign neglect” isn’t a viable preservation strategy for this media.

Kodak slide and film viewer

The Duke Digital Repository offers a number of advantages over our past record storage strategies. Many of these are outlined in the Digital Preservation Policy, including regular monitoring of records for fixity and version control, maintenance of multiple copies in different locations, and control over who can edit, move, or delete materials and metadata. In addition to providing a better preservation environment, the digital repository also offers an improved mechanism for access by scholars using the collection. Rather than requesting a copy of paper records from our filing cabinet, library users can just view the records themselves online. In order to make that kind of access possible, however, we needed to get all of our documentation in order.

Our records underwent several phases of work to prepare them for ingest into the repository. Thanks to support from the Lyrasis Catalyst Fund, we were able to outsource the digitization of all of our legacy hard copy documentation, as well as hire a student worker and intern to assist with the large volume of file prep and metadata creation.

With digital records for all of our treatments in one place, we were able to perform a comprehensive  review to ensure the files were correctly ordered and labeled. Reports were checked to ensure they were complete, images were rotated to the correct orientation, and a file naming convention was established. Some file formats are better for long-term preservation than others, so the raw DNG files and compressed JPEG derivatives were separated into subfolders; only the uncompressed TIFF images are ingested. Written reports were migrated to PDF format. If the report was born digital (rather than a scanned page), the documents were converted to accessible PDF and saved in the PDF/A format for better long-term preservation. 

Once ingested, each treatment is organized as a single item in the digital collection, with the associated reports and digital images grouped together. We created metadata for each treatment to better identify and organize the records, and allow users additional access points. A title was created for each treatment, which includes the year the treatment was completed, a brief description of the work done, the workflow through which it came to the lab (if known), and the title of the item being treated (for example, “2016 board reattachment and tissue reback for instructional use of Athanasii Kircheri”). We also produced metadata to describe the culture or geography of the item’s origin, its condition, and the treatment techniques used. Whenever possible we adopted controlled vocabulary terms from the Art & Architecture Thesaurus (AAT) or Ligatus for all of these fields. Description for Duke’s other digital collections already rely on terms from the AAT. A unique identifier (either the BibSys number or archives collection identifier) was verified for each treatment in order to link the digital collection of documentation to the catalog record or finding aid for the item. 

It’s important to note that some information or records are not represented in the CDA due to privacy or copyright issues. In all of our digitized reports, the signatures of employees have been redacted. While the email communication between conservators and curators that provides additional context for treatment decisions has been preserved, those files are suppressed from public view. Closed collections and items not owned by Duke University Libraries that we have documented (typically for exhibit loans) are also suppressed. 

There was a lot of discussion in the media at the start of 2024 about copyright protections due to a certain culturally significant character entering the public domain. Shout-out to Jennifer Jenkins at Duke Law School for compiling a yearly roundup of works entering the public domain – it’s a really great read. Many of the items that we have treated are still under copyright protection, but there is a good fair use argument for making our treatment images of them available. The images are only used for scholarship, only small portions of copyrighted text are represented, etc. We ultimately decided to suppress certain treatment images of photographic collections, though, as they are fairly high resolution reproductions of the entire work. In those cases our reports are still available. See, for example, this condition assessment for an exhibit loan of a print from the Tom Rankin collection

To date we have ingested around 1,400 records into the repository, which covers up to the 2016/17 fiscal year. At the moment we are hard at work preparing records from the next 3 fiscal years for ingest. Our eventual goal, once we are caught up, is to have a regular, annual release of records from the previous year.

We have heard that some of our colleagues are already using the CDA for research and instruction. In the next episode of this series, we will dive into the interface and look at some of the ways this collection can be used for research.


References:

Announcing the Conservation Documentation Archive

Today we are excited to publicly announce the launch of The Conservation Documentation Archive (CDA). This is the culmination of several years of work to digitize and make available all of the conservation documentation that has been produced as part of caring for Duke’s collections for the last 26 years. Over 1400 records have been ingested into the Duke Digital Repository at this point, with smaller batches of records expected to be added annually. This work was generously supported by a 2021 Lyrasis Catalyst Fund grant and was awarded the inaugural Sandy Nyberg Award. We hope that this archive will become a valuable resource, not only for researchers to access additional information about objects in the collection, but in documenting the standard preservation practices of our institution and the profession at large. A great deal of thought and effort went into building the CDA, so over the coming months we will publish a series of blog posts discussing in greater detail some of our motivations and processes for creating this repository collection. 

In the first installment, we will travel back to February 2020, when staff from Conservation Services began discussing this project, to examine the scope of records that needed to be digitized and problems that the CDA is attempting to address. 

Browsing interface of the Conservation Documentation Archive showing a grid of thumbnail images and descriptions of each treatment

Readers who are unfamiliar with the details of our work may be asking, “What is all this documentation and why do we need to save it?” The rationale for creating and maintaining documentation is laid out in The American Institute for Conservation’s Guidelines for Practice (see numbers 24-28), one of the core guiding documents for our profession. The purpose of this documentation is to be an accurate and permanent record of our examination, testing, and treatment for any of the objects that come under our care. This could be when an item will be altered as part of conservation treatment, but documentation is often created for condition assessments, like collection surveys or prior to an exhibit loan. Our records attempt to describe the collection material, establish its condition at the time of examination, and help future custodians in their work with the item. When possible, the records we create include both written reports and images. 

The top of a treatment form, showing several of the bibliographic data fields: call number, lab number, barcode, title, author, publication date, etc

The format and detail of the treatment reports has varied considerably over the years, depending on the type of object or collection, circumstances, and who produced it. Most of our records were produced by department staff over the years, but some of the records come from vendors or conservators in private practice. Reports typically start with a number of basic fields with identifying information from the catalog or finding aid and name of the examiner and a date for the report. The item’s dimensions are measured and recorded on the form. The report might also include a statement about the justification or goals for treatment. 

Next we try to thoroughly describe the object (and accompanying items), including the format, structure, style, and decoration. The report includes what materials are used and if there are any distinguishing characteristics or marks. We also try to capture any condition issues that are observed, including damage or degradation, evidence of past treatment, and risks of additional damage or loss from use. We will note the methods of examination, including any testing and their results. This information informs our proposal for treatment. 

a checkbox form of various condition issues for objects, physical damage, chemical damage, biological damage, historic evidence
Condition section of an Object Treatment form

The proposal for treatment is often a list of potential options, ranging from minimal intervention to very extensive treatment. We typically list the materials we will use, any alternative approaches that might be possible, and the potential risks. The proposal will include an estimate for the treatment time to help with setting priorities and workload. At this stage in the process, we will hold a meeting with the collection curator and other stakeholders to discuss the various treatment options and arrive at the best course of action for the item. This section of the report includes space to document the date of the meeting, the names of staff in attendance, and signatures of the conservator and curator or collection manager.

The remainder of the report describes the treatment itself, such as procedures or techniques and the location and extent of all alterations. If the treatment carried out is in any way different from the proposal, we will note why. This section documents any material that may have been added or removed, including the manufacturer or source for added materials. We list any adhesives or other substances (cleaning agents, solvents, poultices, etc) used in the treatment, including their chemical name and manufacturing source. This information will be most helpful if the item needs treatment again in the future or if any of the alterations need to be reversed for some reason. The date the treatment was completed and the time spent are also recorded here. 

Two photos of the same book in 3/4 view before and after treatment. Color checker and label positioned in the photo by the book.
Before and after treatment photos

We produce photographs of the item before any alterations are made and after treatment is complete. In some cases we will photograph the object during treatment, too. We follow the AIC Guide to Digital Photography and Conservation Documentation for our photography workflow. The photographs typically show the item positioned on a neutral gray backdrop and include a target or photo checker card and printed label. The target or color checker provides a standard of comparison to help capture scale, direction of illumination, and true color of the item. These targets help us to create consistent before and after images, so that one can more easily compare the changes that occurred during treatment. The label in the photo includes a unique identifier for the item (the lab log number) and date of the photograph. This practice ensures we can always identify the item in the photograph, even if the file names or image metadata are altered or erased. 

the large filing cabinet where all the conservation documentation folders are stored

 

Conservation Services has been storing legacy documentation produced over the last 26 years in a large filing cabinet. The records for each treatment reside in their own paper file folder. The formats and media of those records have changed with the available technology. In addition to paper reports, our archive of physical media contains 35mm color slides, and inkjet photographic prints. The born-digital documentation is saved in a variety of file formats on networked storage. Reports tend to be saved in Microsoft Word or PDF format, while the images are saved in an archival raw format (DNG), as well as derivative TIFF and JPG versions. The TIFF acts as a preservation-friendly file format, while the JPG is a compressed format that is much easier to scroll through or post on the web. We have been printing paper copies of the forms and representative images as a backup for several years. Some of the older treatment folders hold very small fragments (like remnants of original sewing thread) that could not be reincorporated into the object during treatment. Our current practice is to encapsulate small fragments and store them in the enclosure with the item. 

example of an open treatment folder showing a paper report and clear plastic sleeve with twelve 35mm color slides.

The conservation documentation that we produce has enduring value for both collections research and the history of the library. One of the key principles of the Duke University Libraries Strategic Plan is support and advocacy for openness. Our department has always made our records available to anyone who asked (assuming that access respects donor agreements for restricted collections and confidentiality), but previously there hasn’t been a good mechanism for researchers to know that an item has been treated or that these records exist. Library staff in other departments may not even be aware that we have a cabinet full of reports sitting in our lab. 

side by side bar graphs showing 29 of 37 research libraries generating conservation documentation, but only 11 of 37 always or regularly including in archives
Survey data from McCann (2012)

This is a fairly common situation across our peer institutions. A 2012 survey by Laura McCann at NYU Libraries indicated that a majority of conservators at research libraries are producing documentation for special collections treatments, but fewer than half are depositing those records into the institution’s archives. Maintaining records of previous treatments is important for making decisions about the item’s care in the future. It also becomes an important record if an item is lost, destroyed in an accident, or becomes inaccessible for other reasons. Improved access to our documentation might help us to evaluate different treatment methods or materials. It might also aid future scholarship into the history of the conservation profession, providing a record of accepted practices for different time periods, and giving more context to our thought processes and rationales for certain treatment decisions.

With this summary of what we are trying to preserve and why out of the way, next we will look at some of the work that went into digitizing the legacy records and creating the necessary metadata for ingesting everything into the Duke Digital Repository. We’ll be taking a break from blogging in December, so look for our next installment in January 2024. 


References:

Automatic Text Generation Fail

Everywhere you turn lately it seems like tech companies are adding some kind of AI feature to their software. Sometimes these updates are incredibly useful. Other times… not so much.

MS Word automatic alt text generation identifies scroll as toilet paper

We discovered that Microsoft Word will now automatically generate Alt Text (alternative text) descriptions of the images you insert into your documents after it described an Ethiopic scroll as a roll of toilet paper. Clearly the robots have some training to do on cultural heritage materials.

Accessibility settings interface in MS Word, highlighting the checkbox to turn on automatic alt text generationAlternative text is important for making documents more accessible and we make every effort to add it when we can. I’m sure this automatically generated text is accurate enough for stock images and maybe will improve over time, but for the treatment and handling documentation we create it’s really only good for a laugh. We have turned it off rather than risk some ridiculous description being saved in our work. Just open the Options menu and look for either Accessibility or Ease of Access (depending on your version) to silence the machines.

Woolf’s New Home

Virginia Woolf’s writing desk has been an important fixture in the library’s exhibit suite since it opened after the renovation in 2015. The standing desk with an angled writing surface was designed and commissioned by Virginia Woolf in the early 1900s and resided first in Asheham, then in Monks House. In the 1920s the desk was painted by Woolf’s nephew, Quentin Bell, and the legs were later shortened by his wife, Olivier. The plinth on which the desk now rests restores the writing surface to its original height.

Virginia Woolf writing desk positioned in the library exhibit suite

The desk has been a focal point of the Michael and Karen Stone Family Gallery, a smaller gallery space positioned at the back of the suite of three rooms. Recently, Woolf’s desk has been relocated to a more prominent home: the window alcove between Rubenstein Library’s Photography Gallery and Reading Room.

Woolf desk positioned in the reading room windowThis space is the perfect size for the desk and will allow more visitors to safely experience and learn about this important object.

Prior to moving the desk, we spent a couple of months monitoring the environment inside the alcove. We have been carefully looking at the environmental conditions inside each of the exhibit galleries for a long time, so we have a good understanding of the intervals and degree of change in temperature and relative humidity in those spaces. The Stone gallery is very stable, being positioned behind doors and the other exhibit rooms. The photography gallery experiences more fluctuations due to its proximity to exterior doors. Longtime readers may recall our 2018 experiment to monitor the environment inside frames.

We have environmental data from the center of the reading room, but none from the alcove itself. It seemed best to position one of our Onset HOBO MX1101 data loggers right at the height of the desk, rather than on the floor and I didn’t want to cause any damage to the wood paneling on the walls. The MX loggers have magnets on the back, which are extremely useful for deploying them in the stacks, but no metal surfaces are available nearby. Luckily I was able to find some plastic hooks that could be used to attach the HOBO directly to the glass.

plastic hooks with suction cups for attaching to glass

The hook pivots and, as you rotate it down into position, creates additional suction inside the cup. The manufacturer indicates these hooks can hold up to 5lbs, so they are strong enough to hold a data logger. The suction does diminish over time, however. I found that they failed after about 2 months, so I did have to reset the hook about half-way through the monitoring period.

Environmental monitors attached to glass with suction cup hooks

The MX data loggers have built in mounting loops, but they weren’t big enough to fit the end of the hook – so I created a short chain with textile tapes to put them in the right orientation. Anticipating that photo gallery visitors or researchers inside the reading room might be curious about a small device hanging on the glass, I attached small labels on either side of the data logger to explain its purpose.

After collecting environmental data for the space for several months, we were assured that the alcove was a good space for the desk to live. We hope that more visitors to the library can enjoy this important piece of literary history.

Defy Inflation by Cruising

Much of the news this week is dominated by either underwater ship wrecks or inflation. After doing a little research about an early 20th century literary magazine that came across my bench, I discovered that one advertisement serendipitously intersects both of those topics.

This copy of The Bookman came in for some minor repairs before going on exhibit. The covers are the main advertising spaces for this publication and mostly feature some pretty dull descriptions of books available from George H. Doran or Harcourt, Brace and Company. It being June, the image of a steam ship and “Ideal Summer Vacations” advertised on the back really caught my eye.

Eight days in Bermuda for only $90 sounds really nice, but was it a good deal in 1924? The Bureau of Labor Statistics’ CPI Inflation Calculator estimates that sum to be the same as around $1600 today. That seems like a reasonable amount to spend on a long cruise; however, after a quick search I discovered that many of the major cruise lines today are offering the same voyage for less than half of that price. Cruises (at least to Bermuda) have beat inflation!

I’m sure the accommodations on the modern vessels are a lot more comfortable than a hundred year old steam ship, too. In reading about the ships listed in the advertisement, I discovered that they both ended up sinking. The Fort St. George was destroyed by British aircraft during WWII, while the Fort Victoria only sailed another 5 years from the date of this ad before being struck by another ship and sinking in New York Harbor. The wreckage was later dynamited to prevent it damaging to other boats. Luckily all of the Victoria’s passengers were rescued by the Coast Guard before she sank.

Large Marge Rides Again

This spring has been pretty busy and keep forgetting to post an update on “Large Marge“, our new standing press. She had been in storage for a while before arriving here and needed a bit of TLC before we could start using her. We also needed some help from our colleagues in Facilities to create a modification for the base.

After vacuuming off the cobwebs and giving all the parts a quick brushing, I wanted to address the rust that had started accumulating on the base and support rods.

Bottom of press with some rust
Before
Bottom of press after cleaning and waxing
After

The rust was pretty superficial, so I was able to remove it with steel wool. I applied a thin layer of Bowling Alley Wax to all of the surfaces to prevent further oxidation.

Before and after of bottom of press showing rust removal
Before and after scrubbing and waxing.

The next thing that needed addressing was lubricating the threads of the large screw. The grease inside the threaded flange was still functional, but all of the lubricant exposed to air had completely dried out. I found the easiest way to remove it was to scrape it off the threads with a micro-chisel.

The press screw showing dried lubricant being scraped off
Scraping off the old grease

Once the screw was clean, I added some new lithium grease and then raised and lowered the platen a couple of times to spread it across the threads evenly.

side by side images of the press screw after cleaning and then with new lithium grease applied

The most important addition to Marge was a modification to the base. These presses were designed for edition work with many books being stacked between pressboards and loaded in at once. This means that with the platen in its lowest position, there is still about 30″ of daylight between the base and platen. Since we are only pressing one item at a time, we needed to raise that bottom surface up.

full image of press with new wooden platform and stack of press boards. Thankfully one of the carpenters in Duke’s Facilities Management Department was able to build a box that fits exactly inside the gap. The box is reinforced to withstand the full strength of the press, even when someone is using the extended arm to tighten it. An added bonus is that all of the pressboards can be stored inside.

I’m looking forward to putting this press to use, particularly for building clamshell boxes for folio and double folio-sized books. Books of that size are often very heavy and difficult to move around, so they can benefit from the added protection of a cloth-covered enclosure.

Stamps in Your Library Passport

On Tuesday, Libraries Assembly put together a really great passport event for staff to learn more about the different departments and groups within our organization. Eighteen groups volunteered to set up a table with information and activities:

  • Adopt-a-Highway Team
  • Center for Data and Visualization Sciences
  • Conservation Services
  • Data and Reporting Learning Group
  • DivE-In Council
  • Divinity Library
  • East Campus Libraries  (Lilly and Music)
  • Family History and Genealogy Research Guide and the Genealogy@Duke Team
  • Ford Library (Business)
  • Knitting Club
  • Law Library
  • Libraries Assembly
  • Libraries Summer Camp
  • Medical Center Library and Archives
  • Munch & Mull Digital Scholarship Group
  • Where in the World am I from (International Area Studies)

We were there to distribute our new bookmarks, branded buttons, and instruct visitors in simple pamphlet binding.

The library staff who attended were given a passport with space for each group to stamp (or in our case draw a little doodle in highlighter). They were then able to enter their completed passport for prizes in a raffle.

It was a really great way to spend an hour and interact with our peers. We have welcomed many new staff over the last year, so it was nice to meet some of the new folks. It was also a great opportunity to catch up with colleagues from all of the different libraries around campus. Hopefully this will become an annual event!

Handsome Residences of Durham

by Rachel Penniman

We recently got this 1881 map of Durham in the lab and spent time comparing old streets and buildings to what’s currently on those locations.

We were especially interested in the large plot of land on Dillard Street owned by Julian Carr. The map shows many little winding paths on the property.

We wondered if it was a park or a cemetery, but no, it was just a huge personal estate. We found some great information on the Open Durham website including pictures of the beautiful original homes built on that site.

https://www.opendurham.org/buildings/waverly-honor

Waverly Manor (Courtesy Durham County Library, via Open Durham)

The original large homes are long gone, replaced by a number of commercial buildings in the 1920s, then a surface parking lot in 2008, and now a large hole in the ground.

A mixed-use development is currently planned for the block. Likely not as handsome as some of the previous structures, but at least a residence once again.

Hogarth Gets a Spa Day

For the last few months, I have been working on cleaning and stabilizing a very large (25″ x 19″) and fascinating book.

This binding contains a collection of  83 engravings, in various sizes, by William Hogarth (1697 – 1764). This item was formerly owned by Frank Baker (1910-1999), a faculty member at Duke. Hogarth was an English artist known best for his satirical works depicting morality and  social criticism. These works were first executed as paintings and then sold as engravings by subscription. The prints are remarkable and capture so many small details of English life in the 18th century. Interestingly, Hogarth’s work was so widely pirated that he fought to obtain copyright protection and the Copyright Act passed by Parliament in 1735 is known as the Hogarth Act.

In order to make these prints available to patrons, the book needed quite a bit of cleaning and mending. Several of the pages at the front and back were detached.

Before and after surface cleaning

The paper was so covered in surface grime that your fingers would become black from just turning pages, so I spent several weeks just surface cleaning everything. The resulting change was pretty dramatic.

With the tears along the edges mended, and the loose sheets reattached, this items is a little less daunting and safer to handle.

Many of the prints are large enough that they are simply sewn into the binding, but the smaller prints are mounted at the corners to larger sheets. Some of the smaller prints had become detached. Only a few of the prints in the volume have hand-applied color like plate 5 from ‘A Rake’s Progress‘ (above).

Loose prints were reattached with wheat starch paste.

During my initial review, I thought that the first  of four plates in the Election series was torn along the top (or gutter of the binding).

Upon closer inspection though, I discovered that the Election plate was whole – this torn stub was from something else.

Dr. Baker’s typewritten inventory doesn’t list another print in this location and the numbers penciled on each page aren’t interrupted, so it seems like this one has been missing for quite a while. Looking at the details of exposed wooden rafters that are depicted, this fragment could be from one of Hogarth’s more famous works, “Strolling Actresses Dressing in a Barn”.

It is hard to know for sure, but we will be noting the fragment’s location in the catalog record.

Apart from mending and reattaching the prints themselves, I also spent some time flattening the original interleaving. The binder had included sheets of thin, laid paper by affixing them to the verso of each leaf using dots of red wax. The interleaving had become very badly creased, torn, and in some cases was missing entirely.

During treatment, I flattened and repaired the interleaving as much as possible. New loose interleaving sheets were  added for the openings where original interleaving was missing or had major losses.

The top engraving above, titled “Tailpiece, or The Bathos”, is Hogarth’s last engraving, published just eight months before his death. It depicts the figure of Time exhaling his last breath among ruins. In the advertisements for this print that ran in the St James’s Chronicle for April 14, 1764, Hogarth wrote that it should “serve as a Tail-Piece to all the Author’s Engraved Works, when bound up together”.

The previous owner who had these prints collected and bound honored Hogarth’s wishes.

Quick Pic: Matching Outfits

We just love this illustration of a Dacian warrior and his horse in armor from Thomas Hope’s Costumes of the Greeks and Romans (1962 edition).

Illustration of Dacian warrior riding a horse, both in chainmail armor

At first glance, though, it kind of looks like they were wearing matching crocheted outfits. I don’t know that such a thing has been done before, but based on the number of images I can find of crocheted chainmail and horses in sweaters, it seems entirely possible.

Crochet armor, horses in sweaters