Archives and Special Collections Management Software: Why Hierarchical Description, Nested Restrictions and an Unprocessed Backlog Break Ordinary Collection Systems

The request you cannot fulfil

A researcher emails asking for correspondence between a university president and a state official in the early 1970s. The reference archivist knows the material is in a collection of 412 boxes accessioned in 1998, described at collection level only, with a one page inventory typed by a graduate student. She knows some of the presidential files carry restrictions in the donor agreement, and she knows there are personnel records somewhere in the middle that cannot be served at all. Finding the right boxes takes two hours. Determining what may be served takes longer, and involves reading a paper deed of gift in a folder in the director's office.

None of that is a failure of professionalism. It is the ordinary condition of archives, where the volume of material always exceeds the capacity to describe it, and where the rules about who may see what are recorded in legal agreements rather than in software. Every repository is running a very large logistics operation with a description system attached, and the description system is the only part anyone has bought software for.

The tools reflect that. ArchivesSpace is the sector standard for accessioning and hierarchical description and it is genuinely good at what it does. Access to Memory is a capable open source alternative with strong support for international standards. Axiell Collections comes from the museum tradition and describes objects better than it describes hierarchies. Preservica addresses digital preservation, which is a different problem from arrangement and access. What none of them fully solves is the operational layer: where the box physically is, whether this researcher may see this folder, what is in the queue for scanning, and which of your 40 unprocessed accessions should be worked next.

Description is hierarchical and inheritance is the whole game

Archival description under DACS runs from collection to series to subseries to file to item, and description at any level applies to everything beneath it. That single fact defeats most general purpose collection systems, which model records as a flat list with a parent field bolted on, and then cannot answer the question that matters: given this folder, what does the researcher need to know that was said three levels up.

Encoded Archival Description exists precisely because this structure is not optional, and any system that holds archival material has to treat the hierarchy as the primary structure rather than as a nesting convenience. Practically this means a component tree where a node inherits creator, dates, restrictions, language and conditions of use from its ancestors, with local override, and where moving a node moves everything under it, including the physical container links.

ArchivesSpace does this properly, which is the main reason to keep it and build around it rather than start over. What a custom layer adds is the operational reality the description model deliberately leaves out.

Restrictions nest, and they are not one flag

A collection may be open with three exceptions. A donor agreement may close a series for 25 years from the date of the last document in it, which requires computing a date rather than storing one. Student records in a university archive carry obligations under the federal education records statute. Medical material in an institutional collection carries its own restrictions. Files concerning living individuals may be closed for a period after their death, which means the restriction depends on a fact the archive does not hold.

Off the shelf systems give you a restriction note, which is text a human reads. What is needed is a rule: a restriction record attached at any level, with a type, a basis, a computed or fixed end date, a review requirement, and inheritance downward with local override. Then access decisions become determinable rather than remembered. A reference archivist opening a folder record should see open, closed until 2031 under deed of gift clause 4, or requires curator review, without opening a filing cabinet.

This is the single highest value thing a custom layer does in an archive, because it converts institutional memory into a rule the next reference archivist inherits. It is also the part that protects the repository, since a restricted document served in error is a donor relationship and sometimes a legal problem.

Location control is warehouse logistics nobody funded

An archive is a warehouse. Boxes live on shelves in ranges in rooms in buildings, some of them offsite in commercial storage with their own barcodes and retrieval lead times. Boxes get pulled for researchers, sent to conservation, taken to digitization, and returned somewhere else.

Most repositories track this in a spreadsheet, or in the container list inside the description system, which was never designed as an inventory and does not know that box 214 is currently on a reading room cart. The consequence is time lost searching and the occasional box that is genuinely missing for years.

Building proper location control is unglamorous and immediately valuable: barcoded containers, a scan on every move, current location distinct from home location, and offsite requests generated against your storage vendor's retrieval process with lead times the reference desk can quote to a researcher.

The backlog is a triage problem, not a processing problem

Most repositories hold accessions that have never been arranged or described, sometimes for decades. The profession's answer since the More Product, Less Process argument of the mid 2000s has been to process at a level appropriate to the material rather than to item level everything, which was a genuine shift in practice.

Software can support that decision rather than just record it. Track each accession with extent, condition, restriction risk, known research demand, donor expectations and format complexity, then rank the backlog against criteria the head of collections sets. Record the intended processing level per accession so the team is not silently doing item level work on material that warranted a box list. Report the backlog in linear feet by decade of accession, which is the number that unlocks grant funding far more reliably than a narrative does.

Reading room and digitization are queues with rules

A researcher request touches registration, identity verification, restriction checking, a paging slip, a retrieval, a reading room seat, supervision rules for fragile material, and a return. Aeon from Atlas Systems is the established product for this and integrating with it is often smarter than rebuilding it. Where a custom layer earns its place is in connecting the request to the restriction rules and the location record, so a paging slip is never issued for material that cannot be served or for a box that is already on a cart.

Digitization is the same shape: a queue with priority, condition assessment, copyright and restriction clearance, a capture step, quality control, metadata, and delivery into whatever preservation and access systems you run. Repositories usually manage this in a spreadsheet, then lose the link between the digital surrogate and the archival component it represents, which is how institutions end up with thousands of scans nobody can place in a finding aid.

Legacy finding aids have to be converted, not retyped

Every established repository has finding aids in Word, in PDF, in HTML from a 2003 website, and in typescript in binders. Retyping them into a new system is a multi year clerical project that never gets funded, so the material stays invisible.

Machine assisted conversion is genuinely effective here and it is one of the clearer places a model does useful work: parsing a document to identify hierarchy levels, extents, date ranges and container references, proposing structured components with confidence scores, and routing anything ambiguous to an archivist. Expect supervision rather than automation. The realistic gain is turning a task nobody would ever start into one an archivist can review at pace, which is the difference between a collection being findable and not.

What a custom layer should include

• Restriction records at any level with type, basis, computed end dates and downward inheritance with override
• An access determination visible on every component, replacing institutional memory
• Barcoded container and location control with current location distinct from home location
• Offsite storage requests with vendor lead times quoted at the reference desk
• Accession backlog triage with intended processing level and reporting in linear feet
• Reading room integration so paging slips respect restrictions and current location
• A digitization queue that preserves the link between surrogate and archival component
• Machine assisted legacy finding aid conversion with archivist review

What this costs and how long it takes

Across the 2,000 plus projects Digital Heroes has delivered, this shapes up as a first release at $60,000 to $130,000 in 12 to 16 weeks, covering restriction rules, location and container control and reading room integration on top of ArchivesSpace. Backlog triage, digitization workflow, finding aid conversion at scale, donor agreement management and a public access front end bring the program to $160,000 to $380,000 across 6 to 12 months.

What drives cost up in archives specifically: offsite storage vendor integration, since each vendor has its own interface and lead time model. Digital preservation integration with a platform such as Preservica, which brings its own metadata expectations. Multi repository institutions where a university archive, a manuscripts library and a records management program share infrastructure but not policy. The volume of legacy finding aids, which should be priced by document count rather than estimated. And any requirement to expose material publicly, because public access raises questions about restriction correctness that internal use lets you defer.

Build versus buy, honestly

Buy ArchivesSpace or Access to Memory for description. Both are open source, both implement the standards properly, and rebuilding archival description is a bad use of anyone's money. Preservica is the right answer for digital preservation if you have digital material with long term obligations, and it is not a substitute for a collections management system. Axiell Collections suits institutions whose holdings are mostly objects rather than hierarchical archival material.

Build the layer above when restriction decisions currently depend on a specific archivist's memory, when your containers are tracked in a spreadsheet and boxes go missing, when you operate offsite storage and cannot tell a researcher when material will arrive, when your backlog is large enough that triage needs to be evidenced for funders, or when legacy finding aids keep whole collections invisible. The tipping point is not holdings size, it is whether access decisions and physical control have moved out of institutional memory and into something a new hire can use.

How to choose a developer for archives software

Ask them to whiteboard the model. A developer who has worked with archives draws accession, resource, component tree, container, location, restriction and event, and they will ask whether a container can hold components from more than one series, because the answer is yes and it complicates everything downstream. If they propose a flat records table with a parent identifier, they are about to learn archival hierarchy on your budget.

Ask how a restriction attached at series level is evaluated for a folder four levels down, and whether the end date can be computed from the latest document date rather than stored. Ask what happens to container links when a component is moved during processing. Ask how the system represents material that is described only at accession level, because most of your holdings will be in that state for years and a system that assumes full description will fight you daily.

Ask what they have integrated: ArchivesSpace APIs, a reading room platform such as Aeon, a digital preservation system such as Preservica, and an offsite storage vendor are four distinct problems. Get names rather than assurances. Finally, settle code ownership and data export in writing before kickoff, including export in Encoded Archival Description so your descriptions remain portable. At Digital Heroes the repository owns the code from the first commit, and for an institution whose entire purpose is long term custody, depending on software you cannot take with you is a contradiction worth avoiding.