Digital Accessioning Wins, Fails and Dragon Tales: Part 1

This is a long due post about the Digital Accessions Program (DAP), the new service DPS started building in spring of 2024 and has finally been able to launch (check out our wiki for more). DAP is a central service that works with our library partners to safely extract digital content from legacy computer hardware or cloud storage as part of the library's process of evaluating and accepting digital collections. In our DAP Lab we’ve been able to gather equipment needed to safely extract data from the most common computer media found at Harvard Library, including 5.25-in and 3.5-in floppy disks, Zip disks, USB thumb drives, and most types of internal or externally connected hard drives. As shared in the Library’s BDS Vision Year One Progress Report, DAP has been able to coordinate with partners across the Library to test workflows, and we collaborate with stewards to create outputs that help them make decisions and reach their collections processing goals (see our NDSA Digital Preservation 2025 presentation for more information on this part of project formation).

Enter the Two Headed Dragon

Helping people is the key goal, because data extraction isn’t as easy as plug in and drag/drop. The drives are hard to find, the cables are esoteric and confusing, and these can cause drive damage if they’re plugged in wrong. Finally, special hardware tools and methods must be used to prevent your current computer from altering the files. As media is replaced with new storage methods, it becomes obsolescent and fades from cultural memory. Case in point: our most recent summer intern had never seen a floppy disk before. And with changes to cloud-based Microsoft Office 365 word processing, even the trusty standby of “well, it’s the icon for the save button” is sadly no longer relevant. I can still see a floppy icon when I habitually hit Ctrl-S while typing in MS 365 in my browser, because a helpful (and perhaps a little bit snarky) pop-up message featuring a floppy icon says "we're saving your changes as you go” (maybe to help old timers like me feel more secure-old habits die hard).

So collective memory fades, media becomes obsolescent, and then the physical media itself is getting older and is starting to degrade with the help of decades spent in poor storage environments with high humidity and non-ideal temperatures. This can happen with old magnetic tapes like VHS, the binder can begin to chemically crack from the mylar base like old varnish on a table stored in Grandma’s attic. I once saw a presentation from A/V expert Michael Casey from Indiana University about this issue: in the slides, he introduced a fairytale character named “Degralescence,” a fearful, two-headed dragon that embodied these threats to physical media preservation (one head being Degradation, one head being Obsolescence).[1] In his storytelling, a champion emerges to fight the dragon and save the day (metaphorically speaking, this was actually a large collective effort of the Indiana University Media Digitization and Preservation Initiative).

Fighting dragons is definitely an inspirational mindset for DAP. It’s been very satisfying to extract files for our project partners, and each time we’ve wrapped a project or finally had luck extracting files from a dusty old disk after multiple fails, I’ve found myself thinking “Not today, Degralescence!” as I fight on and save the day/data. It felt like things were going great, and that I had a solid sense of what our Lab could handle, and how quickly we could get through projects. Not to brag, but it seemed like we were chalking up a line of wins in our stats… But you know where this is going.

The Dragon’s Tail

In the summer of 2025 we had the benefit of hosting Macilee Mackenzie, our amazing (and foreshadowed) intern, to work on projects for both DAP and our neighbors in Media Preservation Services. Macilee came to us with solid experience working with film collections, and though computer media was new to her, she caught on quickly. So when archivists from Harvard University Archives (HUA) inquired about DAP’s ability to handle a collection of 75 3.5-in floppy disks, I thought this would be a perfect opportunity for Macilee to gain summer internship experience. The floppies arrived in a lantern box, looking like totally normal floppies, with slightly dirty labels and minimal wear, save one or two bent metal sliders that we noted we’d need to pop off before inserting into the disk drive. After attempting to explain floppy disks and floppy disk save icons on desktop versions of Word, then realizing that the cooler cloud/browser Word has no floppy icon (unless you harass it, see above), I was ready to train our intrepid intern on the DAP workflow for 3.5-in floppies. We started out by taking quick reference photos of each the floppies to capture the label information, along with the unique identifier HUA had assigned to each media item. After discussing goals with our HUA partners, considering their processing deadlines, and the apparent good shape of the floppies, this DAP Project had a straightforward plan: we’d load disks into a sturdy TEAC USB floppy reader and extract a disk image using FTK Imager, and then export the files for archivists to review. This was working as expected—disk images were successful, some with a few bad sectors—but files were renderable, and we began checking each disk off. Progress looked great.

And then we heard it—the normal clicks and whirs of the floppy drive (which were so strange to our intern) were occasionally a little chirpier than I was used to. Then the chirps became longer, eventually building into an unpleasant squeal, at which point I asked Macilee to eject the disk, note the squeal, and put the floppy into the “needs more help” box for me to evaluate later. The next disk sounded okay, so Macilee kept going. Checking in later, I pulled out the help box with some concern, after seeing how many had been put in this box of noise complaints. I picked a disk up, looked it over on the outside and saw nothing of note, then I slid back the metal protective shutter…and…… *THWUMP.* Looked like Degradation had gotten me.

I’d been hit, hard, and it hurt to hold the floppy up to the window and see what had happened—the squealing sound had not been a loose belt in the drive, the scenario I’d very mistakenly offered to our intern. The squeal was the sound of the binder being ripped off of the disk by the drive head, leaving gut-wrenching scrapes on the media that looked like the rings of Saturn. I immediately checked the other disks she’d tried to read and saw similar damage. But then I noted something I had not seen before—some of the disks had a raised, crackled, brown texture on the magnetic media surface. This had to be peeling binder, I reasoned. We stopped the project and I dove into research mode. 

Humbled by this, I systematically looked more closely at all of the disks we’d received. I noted that some of the hubs had slight oxidation. I suddenly remembered hearing from the folks at HUA that this collection had come to Harvard from years of storage in a tropical climate—I realized we should have asked about at the project start. I looked at the untouched disks, turning them by the hub so I could inspect the media surface on both sides, and I saw so many of these raised, opaque, crackled looking spots that ranged from the size of a lentil to covering half or more of the disk. It kind of looked like bubbled up brown paint. A memory from my past emerged: I’d once worked with magnetic tapes that had exhibited this kind of squeal during reading, and I knew this was usually caused by tape damage from storing in humid conditions.[2] But does this really happen with floppy disks? My next step: to find others with experience who could help me understand what was going on.

The Quest for Knowledge

I immediately reached out to my colleagues at Harvard Library Preservation Services (HLPS), including Elizabeth Walters, Preservation Librarian for Audiovisual Materials. We realized there are a handful of resources about care and cleaning for floppy disks, all from decades past, when these were still used as active media. As these guides discussed, any contaminant on the magnetic disk layer will interfere with reading the disk, because the drive head must make direct contact with the magnetic layer to properly read the signal encoded in the embedded bits that are suspended in the binder layer. Elizabeth and I looked up the typology of floppy disks and confirmed that it’s made with the same iron oxide as magnetic computer tape, which was made by the same manufacturers as other A/V magnetic media. A/V practitioners have been discussing known issues with binder degradation—affected tape can adhere to itself as it’s wound up, and adhere to deck heads, ripping into the tape surface and destroying playback (and damaging equipment).[2][3] We reached out to Dr. Erin Mysak, Senior Preventative Conservation Scientist at HLPS, who was able to take a sample disk (with permission from our HUA partners) for further analysis.

I continued investigating: I searched for issues with floppy disks and found some great websites and forums from retro-computer enthusiasts who discussed problems and troubleshooting methods they’d used for data salvage on old disks.[4][5] I also reached out to colleagues in digital preservation, finding extremely helpful leads. After posing a question about binder degradation on the membership listserv of the Digital Preservation Coalition, I heard back from Dr. Leontien Talbloom, from Cambridge University Library, and found out she was also encountering issues with mold growth on floppy disks. I was intrigued by Leontien’s blog post documenting these fuzzy growths on disks, and how her disks had been damaged by read attempts, and I was heartened that I was not alone. My disks, however, didn’t have the powdery fuzzy mold growth she was seeing, didn’t fluoresce in UV light, and didn’t have any of the musty smells discussion forums had warned about. I did note from her work that drive heads could clearly cause damage if they hit raised mold growth on the disk surface.

At about that same time, I heard back from our Conservation Scientist that she’d found something interesting: mold growth! It was apparently all over the disks she’d sampled. After listening to her explain her methods for confirming that it was indeed mold (and getting a report full of information about chemical tests and optical microscopy analysis), there I had it: these raised, brownish stains on the disk that just didn’t look the part were actually mold growth that had taken hold in the cracks of the degraded media binder. And with her detailed images of damaged areas, we could see scrapes cutting through these filament structures, where the drive heads had carved troughs into the media surface.

Suddenly the scene became clear: the chirping sound was indeed the dual-sided drive heads scraping against the inner magnetic disk surface as it spun in the drive, as the raised mold growth hit one drive head and accumulated. This imbalance likely knocked the media against the corresponding head on the other side, resulting in more chirps--that became a squeal--as the dirty heads scraped against both sides of the disk like two tiny bulldozers. Another helpful digital preservation colleague, Tyler Thorsted at Brigham Young University, captured and shared audio of these screeching sounds, so now we can keep an ear out for it.

Any contaminant on the media—not just mold—can accumulate against the head, causing disk damage. And if that isn’t scary enough, it gets worse: when heads are dirty with this accumulation, inserting another disk into the drive can cause damage to that next disk, gifting the damage forward until the dirty heads are discovered and cleaned. So this meant all of the disks that we’d tried to read after the really dirty squealing one had been affected, as well. Perfect hindsight made it so obvious.

The final hit: we have moldy disks in our Lab, and mold is contagious. Mold spores are all around us, and while generally not toxic, it can be a health hazard for humans, and HLPS cautions that staff should use gloves and N95 masks when working with mold to prevent these irritants from causing reactions. Mold is a key problem for heritage collections because it can easily rebloom if humidity levels are not kept tightly controlled, allowing it to spread spores to other vulnerable collections materials. Inactive mold can also be spread through physical contact with contaminated items, like disk drive heads. You can read more about managing mold contamination in the Harvard Collections in this helpful HLPS page on Managing Mold Contamination.

So, after what had felt like a string of wins fighting the double heads of Degralescence, DAP had been hit twice: Degradation had given us a surprise scorch (complete with a terrifying squeal), and then we’d been blind-sighted with the dragon’s cursed, poisonous, fuzzy tail. What now? To be continued in an upcoming Part 2 post!

Notes:

[1] Michael Casey, “Why Media Preservation Can’t Wait: The Gathering Storm,” 2017 Texas Conference on Digital Libraries (TCDL), http://hdl.handle.net/2249.1/82129.   

[2] “Magnetic Tape Binder Breakdown,” Preservation Self-Assessment Program (PSAP), University of Illinois at Urbana-Champaign, Accessed September 29, 2025, https://psap.library.illinois.edu/collection-id-guide/softbindersyn.

[3] “Soft Binder Syndrome,” A/V Artifact Atlas (AVAA), Accessed October 9, 2025, https://www.avartifactatlas.com/artifacts/soft_binder_syndrome.html.

[4] Herb Johnson, “Dirty diskettes and how to clean them and their drives,” Retro Technology, Accessed September 29, 2025, https://www.retrotechnology.com/herbs_stuff/clean_disks.html. 

[5] “Delicate 8” and 5¼” double-sided floppy disks – how to avoid a ‘bulldozer’ effect,” Vintage Computer Federation Forums, Accessed October 6, 2025, https://forum.vcfed.org/index.php?threads/delicate-8%E2%80%9D-and-5%C2%BC%E2%80%9D-double-sided-floppy-disks-%E2%80%93-how-to-avoid-a-%E2%80%98bulldozer%E2%80%99-effect.1239735/.