Hi Gregory,
As you can see, AtoM has the world's best open source program manager. There are lots of options out there, but only one comes with Dan Gillean :) You'll find the same level of support if you ping this question over to the Archivematica user forum.
I just wanted to add some additional perspective as a community member. Digitization workflows sometimes seem "lost" in conversations about digital preservation, Archivematica/AtoM integration, file normalization, DIPs, etc. So I'm hopeful that your question will prompt some good discussion!
We adopted AtoM in January 2015. And in October 2018, we started working with Artefactual Systems to configure a digital preservation system using Archivematica. The impact on our digitization workflow is an ongoing question for us.
Our digitization workflow assumes that items are partially or fully catalogued in our AtoM catalogue. In other words, we don't digitize unprocessed material. We use our box-folder "reference codes" as file prefixes. And we use those codes to help organize preservation masters and access copies in our storage environment. The reference code is pretty important to the whole process.
When we launched our AtoM site, we weren't using Archivematica. So we uploaded our access copies directly to AtoM, as Dan mentioned. But we don't upload TIFFs because we want to manage the TIFFs in a different environment. Instead, we create access copies during the digitization process, using Adobe tools (primarily Photoshop and Acrobat Pro, but we also use Bridge to do batch file renaming when necessary). Our policy is that OCR'd PDF/A files are the access copies for text and JPEGs are the access copies for graphic material. In both cases, preservation masters are TIFF files.
I mention this because, although Archivematica can certainly convert TIFFs to JPEGs, and it can also transcribe files with Tesseract, it doesn't provide an intuitive workflow for compiling a batch of TIFFs into a single PDF and transcribe the file. Sometimes, we need to create temporary mezzanine files to help reduce the size of the final access copy. With audivisual formats, we always ask our vendor to produce a preservation master and an access copy.
For us, all of this is much easier done outside of Archivematica or AtoM.
The pre-existing access files would be sent to AtoM. Which is really awesome if that works for you! The main point here is that you will need to consider when you perform file normalization. Archivematica provides some great tools, and AtoM can also render access JPEGs from TIFF files. But we are probably going to stick with our manual normalization process for a variety of reasons that make sense for us. And that is all good.
I think the bigger challenge is around metadata. Dan provided lots of info there, so I won't revisit that. But I think it is worth highlighting that Archivematica cannot currently send an access copy to an existing archival description in AtoM. AtoM maintains a 1:1 relationship between digital object and "information object" (i.e., a single archival description). When you send a DIP to AtoM, Archivematica asks you for a "slug" (part of the URL that points to an existing archival description) and then it nests the digital files underneath that URL. In other words, it creates one or more new archival descriptions in AtoM that will then require revision.
This represents a huge departure from our current workflow. Our team is accustomed to just navigating to the right archival description, uploading the digital object, and updating the metadata as necessary.
We use the digitization workflow as an opportunity to enhance our metadata, so it is no problem to update an existing record after a digital object is uploaded. But we don't need to pass that metadata to AtoM from Archivematica. And if we use Archivematica to upload an access copy to a new archival description in AtoM, we would introduce several extra steps to the workflow. We would need to reorganize the finding aid after the DIPs are received in AtoM. Or build a new finding aid and delete old archival descriptions as we go. This then brings into question how and when we regenerate the "slugs" that form part of the URLs that Archivematica used to send the DIPs to AtoM.
All this to say, we haven't really figured out how to fully bring our digitization workflow into an integrated Archivematica/AtoM environment because of how we use existing metadata in AtoM to support the digitization process. Maybe other institutions have taken a different approach. I'd love to hear!
For now, we are probably going to ingest our preservation masters and access copies (i.e., no normalization in Archivematica) and create the AIP but not send anything to AtoM. And we'll just keep uploading access copies to AtoM as we do. We will use the reference codes in our naming conventions within Archivematica so there will be "implicit" connections between the AIP and the DIP, but we won't see the DIP URL in the Access tab of Archivematica. It would be nice to have it all, but this approach is perfectly fine for us right now. It's all a big step forward in terms of digital preservation activities and it minimally disrupts the day-to-day work of our digitization team.
Hope this helps!
Creighton Barrett