 Kia ora tātou. My name is Flora Faltham and I am a digital archivist at the Alexander Turnbull Library and in that job I spend my time tenderly shepherding born digital heritage collections from where they start their life, so on a hard drive or in the cloud into the National Digital Heritage Archive. And today I'm going to talk to you about the transfer of a very specific kind, a material from a very specific kind of media carrier, CD-ROM, so it's a wild ride and I'm glad you're all sitting down. So our story starts in 2016 when the incredible institution that was Cacodian Stains shut down in Wellington City. It had been there for 153 years, which is a long time for a department store. And in 2017 the Cacodian Stains Archive was donated to the Alexander Turnbull Library and it's a phenomenal collection. It seems to represent the entire span of the organisation, so I suggest you check it out. The thing that I was most interested in with the collection was the 70 gigabytes of business records that we received from two of their work PCs and also the 352 CD-ROMs that contained God knows what. So you can see some stats there. We had 222 unlabeled discs, stacked on four spindles, 130 labelled discs, but 88 of them are just named storage discs, one through 88. So we did a few calculations myself and my colleague Valerie Love and based on a test sample of 30 discs, it would take us six months to read and transfer all of the data from those 352 discs. And the reason for that would be that I don't know if any of you have spent time processing optical discs, CD-ROMs, but it's a very manual process. Take it out of the case, put it in the thingy, push the button, wait a bit, take it out. And also due to the variations in kind of technical troubleshooting required for the discs, some discs you could copy in a couple of minutes, some would take half an hour to two hours due to technical issues. It's a real stop-starty, fractious process and you can't really concentrate on anything else that's incredibly time consuming. So we knew that we had a really crunchy problem which was that we needed to understand what this material was before we could figure out whether or not it was in scope for the collections. But then you put six months' work into something that you can't, that you don't know if you're going to keep it. And that's this really beautiful tension with digital material often that you can't tell what it is when you first look at it. And for me, this speaks to a really big problem, well, big challenge, actually a beautiful challenge. One of the things that keeps me coming to work every day is that with the proliferation of digital collections, we have to think about how are we going to change archival practices to kind of work more effectively with these collections. And this is the question that I really want to bring back to my community of practice is what are these shovels actually look like and how do they fit into our existing workflows? And actually thinking about some of the stuff that Chris was talking about this morning, what do we miss when we treat collections as data? So it can be really helpful, but in whose interests does it actually work when we work like that? So how do we solve the problem? At the National Library, we're really lucky to have a thing called the Business Innovation Group, which supports pretty small scale research and design. So I went to them with a research proposal after a lot of talk with several awesome colleagues. And I was lucky enough to have a research proposal accepted to do some testing and see how I could process these disks. So the research project actually has, I would say actually quite modest aims. So we just want to research and test two different automated optical disk processing solutions for use by the digital archivists. And by automating optical disk processing, we just mean using a machine to automatically load optical disks to a disk drive, read the disks and transfer data to the secure storage area. So all we really want to do is get data from here to here in a really safe way that respects the content and respects the collections and do so in a way which minimizes the manual handling. Manual handling is a health and safety issue after all. And also do so in a way that minimizes the human interaction for technical troubleshooting. And we chose the two different optical disk solutions based on machines that I had identified around the world as being currently in use for this kind of work. So it is now November, my research proposal was accepted in June. What have we actually done? I've got two minutes to tell you. So we have actually bought two machines. And I could talk to you for maybe half an hour about my first time trying to navigate procurement. Thank you. Thank you, all of the people who supported me. And we have also done a really extensive literature review of research on using these machines in the digital preservation world and have pulled out a whole bunch of workflows that people have publicized from glam institutions around the world for using these machines. And from those, we've pulled out eight different workflows to test ourselves. And most importantly, we have put some disks through these machines. And let me tell you, the Akronova in the USB plus, my friends, is living up to the hype. It's is a beautiful little machine. And we've actually tested two workflows with it. One of which was just a straight copy. Okay, so it was with the giant sample size of six disks. But, you know, it's a test. And one of them, so we did a straight copy, which was just take the data off the disks, put it over here. That only had a 50% success rate. It spat out three disks, said it couldn't read it. The other workflow we tried, which was a disk imaging workflow. So take the CD, make a disk image of it, which is a big single file bundle of all of the files and all of the file system information, make a disk image, put it over here. That had an 100% success rate. And so I was very pleased. And my last minute, I will tell you the bad news, which is that the MF digital data grabber is not working nearly so well. In fact, it doesn't actually work. And I have yet to hear back from the MF digital help desk, even though they say they have one, and it's been three and a half weeks. Unfortunately, the robot arm doesn't actually pick up disks. My colleague Val took a really amazing video of the arm, just kind of sadly just sending in it. And it doesn't work. And also the software that came with it just displays an error message and doesn't turn on. And yes, I have restarted the machine. So yeah, I'm hopeful that that will work out. So the next steps are just dealing with some of the data handling implications downstream. So we've made disk images. We need to extract the files from the disk images. That box there in the middle is the crate if we need to return the MF digital data grabber. Thank you.