NFTs are hot. OpenGLAM is hot. People have THOUGHTS and OPINIONS when NFTs and OpenGLAM get near each other.

Before spending any time discussing if NFTs and OpenGLAM should go together, it is worth considering if NFTs and OpenGLAM even can go together. Are they even conceptually compatible?

This question of compatibility between NFTs and OpenGLAM is a marginally interesting question on its face. It is an actually interesting question when used to illuminate an ongoing debate about the relationship between open collections and GLAM finances. This post is my attempt to do the second thing.

Here’s the short version: GLAM institutions should realize that NFT discussions are really discussions about leveraging the power of an institutional brand in an environment where the objects in the institution’s collection are ubiquitous. Regardless of how (or if) they decide to engage with NFTs themselves, those institutions can capitalize on NFT-spurred discussions to think more broadly about ways to move revenue generation models away from scarcity and towards ubiquity. That’s true even if the NFT discussions are only happening internally.

NFT Preamble I: Problems

This post begins, as any post about NFTs must at this point in the conversation, by recognizing the many problems that exist with NFTs today. These problems start with their truly horrifying environmental impact involved with the creation and transacting of NFTs (documented early, and I still think best, by Everest Pipkin here. One can debate as to whether or not this is strictly speaking an NFT problem or a blockchain (or a proof of work blockchain) problem, but NFTs are on a blockchain so for the purposes of this post that distinction doesn’t really matter).

The problems continue with the fact that it seems likely that the overwhelming majority of people involved in NFTs do not appear to fully realize what rights NFTs actually convey. That includes people who think that buying NFTs gives them some sort of ownership over the thing represented by the NFT, people who think selling an NFT for something you don’t own infringes on the owner of the thing’s rights, and people who think that buying an NFT in a thing gives you control over that thing beyond what you would have if you actually purchased the physical thing.

An additional set of problems emerge because the NFT market is frothy and unregulated. That makes it home to pretty much every flavor of financial scam yet conceived by humanity (we are in the ‘fraud is so widespread that content farms write articles about NFT scams’ phase of NFT evolution). Among other things, that means that pretty much any headline valuation of an NFT should be read with a degree of skepticism, and any new NFT scheme should be evaluated for potential scamminess.

Finally, the NFT space provides a place for people to waive away or just imagine their own set of legal and cultural rules. That happens with varying degrees of effectiveness and can actually end up being kind of useful in some cases, while being terrifying in others.

These problems emerge from the current incarnation of NFTs. NFTs may someday become more environmentally responsible, better understood by people involved in them, less scammy, and more tethered to reality. But that’s true of a lot of things. There is no law of gravity that demands that be so. And, as of now, the “claim to legitimacy” to “evidence of legitimacy” ratio for NFTs is still way out of whack.

To be clear, none of this means that the NFT space is completely devoid of interesting projects. It just means that they are fairly hard to find, and ones that I have found don’t really justify a multi-billion-dollar phenomenon.

NFT Preamble II: What They Are

There are a million explainers about NFTs on the internet. If you are reading this you already know the gist, so I’ll focus on the part that is important for this discussion. To the extent that you need more background, this article by Kal Raustiala and Chris Sprigman is great. This article by Foteini Valeonti et al is also helpful, and provides additional context for the GLAM space specifically.

For the purposes of this post, the most important thing about NFTs is that they do not claim any sort of ownership in the underlying thing represented in NFT, or assert a transfer of ownership of the underlying thing between the seller of the NFT and the buyer. That’s how Brian Frye, someone with a very nuanced understanding of the legal and conceptual aspects of NFTs, felt very comfortable selling an NFT of the Brooklyn Bridge, a piece of infrastructure that he certainly does not own. That’s also why it is not copyright infringement to sell an NFT of someone else’s art.

Instead, all an NFT claim is that someone wrote down that the original buyer was associated with a thing in a blockchain ledger. While NFTs can work conceptually even with a random person and a random ledger, in practice the economics of NFTs require that someone care enough about the combination of the original recorder and the ledger to pay for it.

OpenGLAM Preamble I: The Issue

OpenGLAM (that’s Galleries, Libraries, Archives, and Museums) is a community of people, institutions, and practice working to make our common culture available to anyone, so they can use it anywhere, in a manner of their choosing. In practice this means focusing on digitizing public domain works and releasing those files under a CC0 public domain dedication.

Many of the institutions involved with, or considering getting involved in, the OpenGLAM movement had or have legacy rights and licensing departments. In a pre-OpenGLAM context, these departments sold licenses for various types of reproductions of works in their collections. That includes licenses for works that were themselves in the public domain. If you want or wanted a picture of a specific painting in your fancy art book or academic article, these are the people you pay to use the picture.

This kind of licensing model is built on exclusive control. The institution could charge for a license to use a picture of a 13th century painting because they were the only ones who had full access to it and were the only ones with control over a high-quality image of it. If everyone had access to the painting, or if the high-quality image of the painting was freely available for anyone to use, there would not be a reason to pay the museum to use it. (There actually are still reasons that you might want to pay the museum to use it, but that’s an issue for another post (or a bit later on in this post)).

This raises a question for institutions who are considering transitioning to an open model: what will going open mean for the revenue they traditionally drew from licensing?

Before connecting that question with NFTs - we’re getting there, I promise - I want to add three caveats to it:

Caveat 1: Most institutions draw relatively modest revenue from their licensing programs. When you factor in the costs of actually running these licensing departments, many of them may actually lose money.

Caveat 2: There is plenty of evidence that open strategies can be used to increase revenues.

Caveat 3: It is always reasonable to ask how institutions should think about balancing their need to pay for things with their mission (and social compact) to make their collections available to the public, and if licensing revenue should even be considered legitimate in the first place. Suffice it to say that this is a question upon which reasonable minds may disagree. More importantly for the purposes of this article, regardless of its conceptual legitimacy, it is a concern that institutions regularly express.

Enter NFTs

NFTs do have at least one interesting feature: they combine exclusivity with ubiquity. As discussed earlier, the NFT isn’t the thing it represents. The NFT is just someone writing down your name next to the name of the thing in a ledger. That means that what you are buying (being written down in the ledger) is not in inherent conflict with the thing your name is being written down next to being ubiquitous and freely available to everyone. The thing being available to everyone does not change the fact that your name was the one written down in the ledger.

The result is that, unlike selling licenses to use images, there is not an inherent conflict between selling an NFT and making the underlying work connected to the NFT freely available to anyone. OpenGLAM and NFTs are compatible, at least conceptually. Widespread, open access to the underlying thing does not directly impact the economic viability of the NFT.

Instead, in many ways, most of the value of the NFT comes from leveraging the reputation of the institution that is writing the buyer’s name down in the ledger next to the name of the work. An NFT of the Mona Lisa offered by the Louvre (that’s the Louvre writing your name down next to the Mona Lisa in the ledger) would probably be worth more than an NFT of the exact same Mona Lisa offered by me (that’s me writing your name down next to the Mona Lisa in the ledger).

Why the distinction? The NFT represents officially sanctioned affiliation with the work. That is different from ownership of the work or control over the work. Nonetheless, that officially sanctioned affiliation might be worth something to someone.

NFTs as Official Merch

Viewed this way, the NFT model is strikingly similar to an ‘official licensing’ strategy for open works. Under both models, everyone has full access and use of the open access work. Also under both models, some people can pay a fee to the institution for a more official affiliation.

For the NFT, that affiliation takes the form of a record in a ledger that documents that it was the institution that wrote your name down next to the name of the work. For the official licensing model, that might take the form of the right to use the institution’s trademarks with your product that incorporates works from the open collection, some sort of certification that the use or reproduction of the work is in line with curatorial standards, or something else.

Both of these models sell an institutional endorsement of the user’s affiliation with the work without limiting anyone else’s access to the work (although only the NFT comes with a frothy market and the ability to easily transfer that affiliation).

Is This a Good Idea?

Just because the idea of NFTs is compatible with the idea of OpenGLAM does not mean that OpenGLAM should embrace NFTs. It only means that OpenGLAM could embrace NFTs if it was so moved.

The start of the answer to the “is this a good idea?” question lies in NFT Preamble I above. There are major risks and costs that come with getting involved with NFTs as they currently exist. These include environmental costs, reputational costs, and other costs in between.

These costs are real and high. Even if the sale of some NFTs could finance the opening of an entire collection, the cost of doing so under current conditions could be very hard to defend.

Another part of this answer lies in the fact that institutions need to do real thinking about how and why they use their brand to validate affiliations. That is true in the NFT context, and in its activities more broadly.

At a minimum, without doing anything else, NFTs may act as a mechanism to allow institutions to think more creatively about what it means to capture value from their open collections. It makes it easier for people within those institutions to think more broadly about their brand association value, especially in connection with their open collections. If that happens it will be useful. Although maybe not useful enough to justify everything else.

Header image: Scene at a Fair: A Magician from the Met Open Access collection

Today the Engelberg Center released a report from the Open Hardware Distribution & Documentation Working Group that explores what is really needed to create a distributed manufacturing network for open hardware. You can find the official launch post here and the report itself here. I recommend checking it out.

This post focuses on one part of the report: the idea of “Commercialization-as-a-Service.” For me, this was the most intriguing thing to come out of the year’s worth of discussion within the Working Group (and it was a discussion full of intriguing things!).

Open source hardware is constantly working by analogy to open source software, so let’s start there. With open source software, the manufacturing and distribution layer is fairly straightforward and so lightweight that it is easy to miss. Publishing code online effectively manufactures and distributes it worldwide.

Admittedly, in reality the “it just happens” nature of open source software distribution breaks down pretty quickly. Nonetheless, “pretty quickly” is not immediately. The ease of code distribution online means that a piece of open source software can get at least moderate use and build a decent sized community without being supported by significant commitments to building out infrastructure.

That is not really the case with open source hardware. While it is true that you can easily post design files for hardware online, design files for hardware are not hardware. In order to get beyond a relatively small core of people who will assemble and build their own version of the hardware based on instructions, developers of open source hardware are likely to confront a need to begin packaging physical objects and distributing them to others. These objects might take the form of a kit that others assemble, fully constructed hardware, or something else entirely.

Regardless of what form it takes, this manufacturing and distribution layer becomes an independent obligation for the team much earlier in the hardware lifecycle than it would in the software lifecycle. Furthermore, this layer is probably further from the work of designing the hardware than the equivalent responsibility is from writing software. That means that there is no guarantee that the team that originally came together to design and build the hardware has someone (or a group of someones) who are also excited about spinning up a manufacturing and distribution network. That is especially likely to be true in the context of research lab-developed hardware (which is the focus of the paper).

Where does that leave things? Manufacturing and distribution of open hardware becomes an important stand-alone task early in its development cycle. At the same time, the team working to create that hardware may not be excited about engaging in that task.

Enter Commercialization-as-a-Service

One way to try and solve this problem is to get more people excited about building out manufacturing and distribution networks. An alternative approach would be to make open hardware projects less reliant on building their own manufacturing and distribution networks in order to succeed.

This second option may be most realistic in situations where open hardware has been developed in scientific research lab contexts supported by grant funding. In those situations, a grantor has already paid for the development of the open source hardware in the context of the specific project. Currently, in most situations that hardware either remains within the developing lab or is effectively abandoned.

Creating commercialization-as-a-service could make it much more likely that open hardware developed in one lab becomes available to the research community more broadly. That service layer could develop a standing network of international manufacturing and distribution partners, as well as expertise in productization and marketing of hardware. It would work as a pump, pulling proven hardware out of labs and pushing it to other researchers worldwide. In doing so, it would reduce the current ‘unicorn’ condition of academic open hardware success, whereby a successful open source hardware project needs a team that happens to combine the engineering skills to develop the hardware and the logistics skills to manufacture and distribute it.

Building this layer feels like the accelerant role that foundation and other grantors are well positioned to play. Assembling the expertise required to productize, manufacture, and distribute hardware can be done once and applied to a range of hardware. Doing so could greatly increase the impact and adoption of the hardware that funders are already funding. Conversely, without building this layer, these funders are making it much more likely that effective hardware remains tucked into individual labs, hidden from the rest of the research community.

Of course, this is just one of the things that’s in the new report. You can read the entire thing here.

Header image: New Inventions of Modern Times, The Invention of the Clockwork, plate 5, from the Met’s Open Access collection.

Today is is even more clear that copyright law does not prevent you from using whatever material you want in your 3D printer.¹

That is because today the Register of Copyrights submitted her recommendations on the petitions to exempt various activities from 17 U.S.C. § 1201(a)(1)(A). In practice, this document gives people permission break digital locks for specific purposes.² Whatever you call it, the recommendations include granting the request to expand the scope of the current rule that allows you to unlock your 3D printer and use the material of your choice.

As I explained in a previous post, every three years the Copyright Office gets to give people permission to break digital locks for specific purposes. In the past, the Copyright Office had included permission to break any digital lock that prevented someone from using the material of their choice in a 3D printer.

This was good, but the language could have been tightened up a bit. So, this year, I submitted a request to slightly modify the language of the rule in two ways.

First, I asked the Copyright Office to change the word they used to describe ‘what you put in your 3D printer’ from ‘feedstock’ to ‘material’. ‘Feedstock’ was a word left over from previous rulemakings and ‘material’ is much more widely adopted. To be clear, it is 100% my fault that the Copyright Office started using ‘feedstock’ so I am glad to be able to end that practice with this year’s rule.

Second, I asked the Copyright Office to remove language that only gave permission to break digital locks that were ‘microchip-reliant’. This seemed like an unnecessary restriction on the rule. If a digital lock was preventing the use of a third party material, there was not a good reason to litigate whether or not that lock was ‘microchip-reliant’.

The Copyright Office did both of those things. The new class is described as:

Computer programs that operate 3D printers that employ technological measures to limit the use of material, when circumvention is accomplished solely for the purpose of using alternative material and not for the purpose of accessing design software, design files, or proprietary data.

Both the petition to renew the original exemption and the petition to make these changes went unopposed (thank you to both the Free Software Foundation and NTIA for supporting the changes). Hopefully that means that the rule will be fairly stable going forward, and that renewing it again in three years will be a straightforward affair. Stay tuned for that, I guess?

Openness is on the march!

Actually, openness has been on the march for some time. As it marches, the fundamental nature of the most important debates around openness appear to be changing.

A number of different open communities - open software and hardware, open culture, open data, open GLAMs, etc. - seem to be entering a new phase of their development. This new phase is characterized by their successful growth into and adoption by the mainstream.

This new phase comes with new debates. In the early days of openness advocacy, the core debate involved a small group of people working to defend the very concept of openness to a skeptical world. This debate played out between openness advocates within the community and openness skeptics outside of the community. It was about the fundamental legitimacy of the concept of openness. The dynamics of this formative is-open-even-a-real-thing-? debate shaped the community in critical ways.

While debates about the fundamental legitimacy of openness continue (and will continue), the growth and success of openness has given open communities confidence that the core concept of ‘open’ is viable. This has helped shift the discussion around openness in new directions towards new questions. The key debates in openness advocacy now involve a larger group of people trying to reconcile openness with other values that they hold to be equally dear.

Today’s debates are shaped by two related factors: first, that openness is no longer fighting to prove its fundamental legitimacy; and second, that the openness community is now big enough to include people who view openness as one of many important values. The openness community is now larger than its original ‘single issue voters’.

This larger openness community is struggling to balance its interest in openness with other closely held values. Unlike in the past, where other values were seen as competing with openness, today’s debate strives to find ways for openness to coexist with these values. However, the fact that both sides of these discussions believe in openness can be a challenge to the adversarial debate styles that openness communities developed in earlier eras.

This confidence about the sustainability of openness and concern about other priorities results in debates that would have been hard to anticipate and appreciate in the early days of openness advocacy - concerns about openness and privacy, openness and control, openness and ethical uses. These debates shift the focus from an external conversation - between the community and skeptics - to an internal conversation - within the community itself. While none of these debates are unprecedented in these communities, their centrality does feel like a shift.

That creates a challenge for the openness community broadly, and for specific openness communities. If they continue to operate under the old debate models, the communities will treat these internal criticisms as external threats to the foundation of openness. In doing so they risk alienating the expanding community at what could be a moment of triumph. Alternatively, these communities can reorient themselves to engage with the new criticisms as a way to strengthen the work, make it more inclusive, and ultimately broaden the adoption of openness. This is a harder path, one that may prove to be self-destructive. However, it may also be the only way to continue to expand openness.

Openness Communities Start by Overcoming Skepticism

In most cases, openness communities were founded in environments that were skeptical towards openness.

Openly sharing software or hardware, making books or photographs parts of the commons, releasing data to the public, and inviting anyone to make use of GLAM collections were fringe ideas that needed to prove their legitimacy to the wider world. Early advocacy communities tended to be relatively small and homogeneous. While these communities were often full of internal debate, their most important arguments were with the rest of the world.

The rest of the world was largely skeptical of arguments for openness and questioned or rejected the conceptual frameworks that the open advocacy communities represented. As a result, openness communities often adopted a “for us or against us” posture towards debates around the merits of openness. They established bright line rules to help define ‘true openness’ to reduce the risk of bad actors gaming the system. Questioning these tenets was viewed as akin to questioning the legitimacy of openness itself.

This posture helped to fuel the growth of openness. However, it also created patterns for responding to critiques that can cause problems in a more expansive openness community.

Openness Has Won

While there is still a long way to go, in many cases openness has spread more effectively than most of its original proponents could have hoped. Even opponents of openness have retreated from categorical “openness can never work” arguments towards more nuanced “openness does not work in this case” arguments. The world is still full of open skeptics, but it is also full of success stories that prove that open approaches can work in a broad range of situations.

One result of this success is that the open community is much larger and more diverse than when many of the original open communities began. These diverse members bring a broader set of expectations and concerns to internal discussions within openness communities themselves.

This diversity of expectations united under the banner of openness means that the openness community as a whole, and individual open communities in particular, are grappling with a new round of conceptual concerns from within. Unlike the traditional external critics of openness, these concerns do not question the fundamental legitimacy of openness. Instead, they start with an assumption that openness is good, socially productive, and something that should be supported. They then attempt to incorporate other values that community members view as equally important to openness in the hopes that these values do not contain fundamental conflicts.

Community members with these concerns are not trying to eliminate openness from the outside. Instead, they are working to improve openness from within by finding ways to accommodate their additional concerns and uphold what they view to be important values.

What Does This Look Like in Practice?

It is all well and good to make grand pronouncements about cultural shifts. How does this conflict manifest itself in specific openness communities?

Ethical Open Source Software Licensing

One of the core tenants of the open source definition is “No Discrimination Against Fields of Endeavor.” In practice, this prohibition means that it violates the open source definition to prevent open source software from being used in ways that creators find unethical.

There are good reasons to adopt this kind of rule. Ethics are slippery. They can evolve over time and be different in different contexts. They can also be incredibly hard to define in a consistent way. Introducing ethical constraints to a copyright license can add instability to the entire process.

And yet… a growing community within open source believes that instability is a price worth paying. The more central software becomes to our lives, the more important ethical considerations may be to how it is used. People who understand themselves to hold openness as a core value are working to combine it with other values they hold equally dear.

For the purposes of this post, the most important feature of the ethical open source debate is that it is coming from within the open source community. Advocates for ethical open source fundamentally believe in open source as a development model. They are not questioning its core concepts. Instead, because they believe in open source so strongly, they think it is important that it harmonize with other equally important values.

Creative Commons Without Copyright

Creative Commons has been incredibly successful in building a commons-based approach to licensing. This success is built upon the Creative Commons licenses themselves, and also upon the social and cultural structures that have grown around the licenses. In many ways, the Creative Commons licenses - and the icons that the organization maintains to represent them - send a signal in support of openness that is just as much cultural as it is legal.

While an undeniable marker of Creative Commons’ success, these cultural signals can raise questions when they are detached from legal reality. For example, many members of the 3D printing and open source hardware communities attach Creative Commons licenses to objects that do not have copyrights that can be licensed. At best, this converts something that was designed to be a legally binding document - a license - into something more like a cultural signal or a not-legally-enforceable request.

This shift raises the potential for all sorts of problems or misunderstandings between creators and users. Is it a sign of a corruption of Creative Commons? Or an exciting expansion? While reasonable minds can disagree, both sides of the disagreement start from an assumption that Creative Commons - and the openness it represents - is a good and viable thing.

As with open source software, members of the Creative Commons community attempting to use the licenses outside of the scope of copyright are starting from a place of enthusiasm about what Creative Commons represents. That enthusiasm motivates them to try and bring what they understand as the core of the Creative Commons ethos to places that were never intended by the original creators of Creative Commons. They are doing it “wrong” not out of spite for the work of Creative Commons, but rather because of their affection for Creative Commons.

Who Should Control Open GLAM Collections?

The Open GLAM (Galleries, Libraries, Archives, and Museums) movement is an effort to encourage GLAM institutions to make their collections more accessible to the public. While this can take many forms, one of the most common ways to implement an Open GLAM agenda is for an institution to digitize the works in its collections and make those digital versions available online in an open and unrestricted way.

A core goal of the Open GLAM movement is to make more culture available to more people with fewer restrictions on what those people do with the collections. Institutions large and small have joined the Open GLAM community by opening parts of their collections.

While this goal of making more culture available to more people with fewer restrictions drives the push for Open GLAM, it can also overlook the sources of those collections and sidestep questions of who should have the authority to make culture more available.

Projects such as the TK (Traditional Knowledge) Labels are designed to give Indigenous communities a voice in how GLAM collections that include objects from their cultures are used by the wider world. Scholars have also raised questions in other contexts about who should be empowered to decide when an artifact is made available to the broader public.

Viewed one way, these efforts act against the core goals of openness by limiting how ‘open’ objects can be used. Viewed another, these are attempts to find a version of openness that includes equally important values like representation and inclusivity.

However one is inclined to view them, these types of critiques are not rooted in a rejection of the core concepts of Open GLAM. Instead, they are being raised by people within the Open GLAM community who believe that there are better ways for Open GLAM to achieve its goals. They believe that Open GLAM will be strengthened - not undermined - by incorporating concerns and stakeholders that were largely left out of the original discussions. To the extent that these new tools introduce restrictions on the use of open access works, those restrictions assume full open access is a natural starting point for the discussion.

Does This Need to be Resolved?

Ethical open source licensing, Creative Commons without copyright, and a slightly less open Open GLAM are examples of this tension, but they are not alone. You can find similar dynamics in discussions around privacy concerns in connection with open data, business models in connection with open publishing, and even questions about how to deal with dominant platforms in connection with the open internet.

In each of these communities it is easy - and tempting - to try and push these questions aside, categorizing them as beyond the community scope or undermining the push for openness. After all, the universal principles that founded these communities have gotten us this far. They are a success!

This type of doctrinaire approach tends to work well until it doesn’t. If open continues to succeed and grow, open communities need mechanisms to incorporate the broader set of people who will make them up. It is not reasonable to assume that this new universe of open enthusiasts will come with exactly the same values as the founding cohort, or to declare that the concerns that they bring to the community should be dismissed out of hand.

Instead, open communities should begin prioritizing developing tools and processes to manage internal dissent about the limits of openness in ways that feel legitimate to everyone involved. This does not mean that open communities need to accept all of the concerns or accommodate them. It does, however, mean that they need better ways to give them a fair hearing.

Header image: The leaders of two armies conversing, their respective troops and camps on either side; above, Minerva observes from her chariot; set design from ‘Il Pomo D’Oro’ from the Met Open Access collection

On November 13, 2014 I purchased the domain cleverhans.org because I decided that the world needed a digital version of Clever Hans.

At the time, I did not know how to program in any meaningful sense. I had used logo in elementary school and taken a class in BASIC in middle school. I also used markdown to write posts for the Public Knowledge blog. Nonetheless, the world needed digital Clever Hans and CleverHans.org was going to be my excuse to learn how to do it.

Almost seven years later, today I finally figured out how to put a functional CleverHans.org site on the internet. It took longer than I thought.

Along the way, I took massive detours. I learned some python (the hard way, of course). I learned some arduino, processing, p5.js, circuitpython, and even ml5.js. I bought books and jumped around the internet. Using Coding Train to learn p5.js tricked me into learning some javascript. I converted this site into Jekyll. I played around with phaser. I built other dumb things.

I learned the basics of most of these without mastering any. Every time I thought I had things figured out, I would come back to Clever Hans. And every time I came back, I discovered a new thing that I did not quite know how to do.

Finally, I started another run at it over the summer. I purchased Emanuele Feronato’s phaser book and this time it actually worked. Everything I had learned from all of my stabs at programming finally came together. As a bonus, in the years since I purchased the domain, tools like github and cloudflare pages meant that I could host the site for free.

CleverHans.org is a dumb site that does not do very much. It certainly does not look like something that took seven years to build. Fortunately for me, it is just dumb enough and does just little enough that I always felt like a working version was just around the corner. In its way, it provided just enough motivation to get me to learn a bit more programming and stick with it just a few minutes longer.

And now the world does have a digital version of Clever Hans. On the internet. What a world.