A Tale of Two Holes in the Wall

This is going to end up being a post about preservation management with an epilogue on equitable access and shared collections, but it starts with some rough building trades-work and a visit to a notable internet site’s headquarters. In planning the trip, I’d asked something about energy consumption for their data center. That led someone to mention a clever system for pouring the waste heat from their servers into the building’s air plenum for free heating.

I love this kind of stuff.

On the day of the visit, I asked to see how this worked and, upon entering the server room, was delighted to be disappointed because their system for harvesting residual heat was… wait for it….

a hole in the wall.

Specifically a roughly 7′ x 10′ hole knocked through the drywall into the adjacent mechanical room, which served as a sort of plenum for the building.

No ductwork. No air handlers. No paint, trim, or gratings. Just a bright idea and a hole sledged into the wall.

Again, I love this kind of stuff. We had a hole in the wall at ReCAP for a similar reason. The layout of the sprinkler lines for the office area routed a test valve through an exterior wall in our break room. There, a combination of at least four different factors conspired to gradually chill and warm the valve over fifteen years, until one day it cracked.

That was an exciting day. Also, damp.

In the course of our disaster review, we knocked a hole in the wall. Deliberately, I hasten to add, so that the replacement valve would communicate with the warmer environment of the adjacent room instead of  the cold exterior wall surface.

From an engineering and risk management perspective, that was job done. However, if you have ever seen a hole in the wall, even a fine, deliberate example such as this was, you’ll know it can look like an accident. To some viewers, it could imply neglect or carelessness. Now, library high density storage and logistics center operators are perhaps not renowned for our exquisite aesthetic sensibilities, but we are careful and attentive. That’s the whole game.

The deeper lesson is that a hole in the wall is not just architecture (I think the genre is called brutalism): it is information, a symbol or signifier, and this hole was sending the wrong signal. In effect, there was not just a hole in the drywall, there was a fracture in our communications strategy. We tell you about “care and attention” but we show “slapdash and lazy.” This could not stand.

So we put a refrigerator in front of the hole.

I am not kidding. Our staff had grown enough that we needed a second fridge, and after considering the viable traffic patterns in the break room, we selected the en-holed wall as a suitable site and then got very excited as we realized that we could gain a triple free rider benefit: we could cover the hole (mitigating the communications problem), we could harness the waste heat from the fridge to warm the valve at no cost (you see how these two hole stories relate now), and here’s the real kicker: by selecting a dual-mode device that could act as a refrigerator or freezer, we added a layer of efficacy to our disaster response services.

When water damage occurs, the most common preservation measure is to freeze the affected materials. Running an empty freezer year-round is very expensive, though, so having a device that is normally used for refrigeration of staff lunches but can re-purpose for disaster response is a great economy.

(Obviously, we would provide a break period for staff to eat their displaced comestibles in an emergency situation; care and attention, you know, and it’s a rookie mistake to go into a crisis on an empty stomach.)

(Also, while we’re in the general zone of ReCAP’s aggressive over-thinking on efficiency, I want to note that these dual-mode units are a single contiguous food storage space, unlike the standard over-under fridge-freezer. Since most people bring refrigerated and not frozen food to work, this model performs better for the purpose of staff food storage, on an assignable square foot basis, but still gives us the maximum freezer space in a disaster.)

At this point in the narrative, you might think, “well, I hope this is almost over; surely this guy has some beans that need recounting.”

I am delighted to disappoint you, though, because there is one more lesson to learn from all of this.

(Oh, but first, there’s just this the one last thing about the fridge I want to mention, which is that it is low- to no- use on evenings and weekends, when it has very little food to store. It’s still on and producing waste heat, though. It just so happens that we also minimize heating and cooling during unoccupied hours to save power, so the really clever part is that overnight and weekends are exactly the times we most need supplementary heating for that plumbing run in the wall. So, when the refrigerator is least engaged in its primary food storage job, it is most engaged in its secondary heating function.)

Anyway, our thoughtfully hidden and functionally effective hole and refrigerator combination seems like a pretty good job of work in the operational efficiency line, but it still conveys information. In this case, the hole remains, though it sends its message indirectly rather than directly, and the method of conveying information can be informative in itself. Even though we meant for the hole to be there and we were deliberate in our reasons for placing the fridge in front of it, our hole behind the fridge could imply that we goofed up and were trying to conceal our mistake. We were back to “slapdash and lazy,” in spite of ourselves.

Our solution is to place a grating over the hole, inexpensively putting some fit and finish on the space while preserving the new functionality. In library management terms, this corrects the communication part of the problem by adding metadata.

Lest you think that we are flagging in our over-analysis, let me observe that there is another path to resolving the hole messaging problem. The hole at ReCAP was saw-cut and thus basically rectangular to start with. The whole in the data center was sledged-out, though, so that it was composed of an irregular series of overlapping circular fractures. Custom fabricating a grate to align with the contours of that hole would have achieved the desired communications goal, showing that the hole was intentional and functional. (heck, it might have won a design award), but the cost would have been substantial. Creating a rectangular hole in the data center wall (by removing additional material or in-filling) to make it match with a stock rectangular grating would probably be the less expensive path.

This is actually a fairly important lesson to absorb. It is always possible to create custom components to conform to as-received systems, and it may be possible to justify those custom solutions in terms of strategic outcomes. On the other hand, it may also be possible to reach those strategic outcomes by modifying received systems to fit with standard components.

If the only questions a manager asks are “can we do it” and “will it meet our goals,” they are still open to a failure in efficiency, and I would argue, a subsequent failure in stewardship. I am strongly of the opinion that the core preservation question is not one of technical means for correcting problems, but of sustainable process for executing the institution’s functions. Doing things inefficiently and idiosyncratically is a process risk and, when a failure occurs, it means that technical methods are the only recourse, creating an effective single point of failure. Assuming that failures are inevitable, an effective preservation strategy manages the size and timing of failures so that they do not overwhelm existing technical means nor do they occur when there is no technical method for recovery available.

This split path between possible and efficient deserves attention for library management for another reason, as well. The two grates (rectangular and… idiomorphic, perhaps?) send the same core message, but they have different overtones. The rectangle says “intentional; unremarkable” while the other says “intentional; unconventional.” One releases attention and the other commands it, and this is where I see the bridge to our emerging idea of librarianship as the management of the attention economy, over and above the information economy.

Here is my other bit of opinion to offer: efficiency is not our ultimate goal in library (or archive, gallery, or museum) services, it is our penultimate goal. Ultimately, we need to be curatorial, which is to say that we need to be engaged in a dialogue with our communities and our collections. The virtue of centralized efficiency in collections management is that it can remove encumbrances from this curatorial work.

The parallels with the collective collections model of library services should be clear. Library storage and logistics facilities, and I would venture to say other collective library infrastructure, are square grate kinds of places. The proximate library, the room a reader walks into or the interface where they conduct research, is for indulgent idiosyncrasy.

In the emerging service model of the collective collection, I expect these functions to coalesce and clarify, and (begging pardon for the alliteration) to do that in counterpoint, so that efficient, unremarkable and anodyne services from library logistics centers are delivered via highly curated user interfaces and in library spaces that create an inviting, supportive, and productive context for encountering the record.


Lest you still think we are flagging in our duty of over-analysis around these parts, I draw a somewhat more personal lesson from this. I’ve worked with a lot of libraries and archives, with a lot of different challenges and resources. The inequities we have throughout our sector can be demoralizing and as often as not, they are structural outcomes enforced by factors like geography, economics, and politics. I don’t think these inequities can be changed by adopting a particular management style or ethical viewpoint, and even when they can be modulated a bit by cleverness or grit, the prevailing conditions limit the scope of what is possible. The structure has to change.

Changing the structure is the deep reason to do the sort of detailed work I described above, which is all just one (surprisingly long) example not of merely seeking efficiency, but of arranging things to provide added value in a way that is intrinsic to their primary functions. It is this combination that strikes me as crucial: arranging functions so that they are mutually self-supportive, to create a cooperative advantage over and above the individual functions, no matter how efficient they may be.

Libraries were built in particular places and to particular tastes, and those qualities can serve as barriers or enticements to learning. One and the same location and architecture can say “I am welcome here” and “not for me.” A building that worked well for one time in its place, may work poorly for a different time in that same place. The same is true for the way we describe an item, or create a user interface. We have a substantial incumbency effect embodied in our facilities and our metadata. One person’s jargon is another’s vernacular, making discover easy for the one and impossible for the other. The accessible interface for one user group may be nearly opaque to another.

I speculate that disentangling the collections from the spaces and the services gives us a way to redress some of the systemic problems we have with diverse and equitable access.

The virtue of a collective collection is that it:

  1. reshapes the lines of our budgets dedicated to certain core functions (store and protect, so to speak),
  2. makes those lines as small as possible, and
  3. shares the load for that expense across multiple parties.

That means more fiscal flexibility in the immediate sense (less money spent for the desired result) and in the strategic sense (more flexibility about when to spend money). “How much” and “when” are both important financial questions.

Upstream, this makes possible several important outcomes:

  • Enabling users to encounter materials in the most hospitable environment.
  • Enabling a more diverse readership to have more comprehensive access to the collected record.
  • Enabling multiple avenues of discovery, since the repository supplies objects that can be addressed through many different interpretive systems.

My basic conceit is pretty obvious, I hope. If we take a beat and reflect on the five laws — books are for use, every reader their book, every book its reader, save the reader’s time, and the library is a growing organism — a collective collection that supports a myriad of curatorial centers makes good sense. No one library can be perfect for every reader, but many libraries can draw on the same pool of resources to more perfectly address the needs of their particular readers.

To get there, we may have to knock a few holes in the wall.