Open Platforms and the Anti-Fragile Library

18 April 2014

Earlier this week I gave a short talk for IFLA's New Librarians Global Connection: best practices, models and recommendations webinar series. They had asked me to talk about my ideas on 'technolust' and antifragility, but of course my thoughts have moved on a little since my first blog posts on those subjects.

A recording of the webinar will be available on the IFLA New Professionals SIG website soon. I have also published the text of my talk below, including the slides, if you'd prefer to read it. This also makes it a lot easier to click through to the links to the ideas I referenced!

The danger of techno-lust

Tech industry veteran Jon Udell wrote a blog post in January called "3D Printing isn't the digital literacy that libraries most need to teach". This is something I also wrote about early last year, for much the same reasons as Udell. He asks why public libraries are so obsessed with maker spaces and 3D printers, instead of teaching digital literacy.

Often we think we are teaching digital literacy when we are actually just showing people how to make a particular device or program function. But if you are simply showing people how to use new technologies without explaining what the technology is doing to them then you have failed as a librarian.

Digital literacy is about surfacing the invisible. It's about strategies, not tools. We need to equip our communities with the 'soft tech' of invisible assumptions and effects more than the 'hard tech' of particular interfaces and products. Rather than installing 3D printers and explaining how to do truncated searches in particular journal databases, there are far more interesting and useful things libraries can do to equip our communities to navigate the information landscape. The idea of 'anti-fragility' is one interesting model for thinking about those things.

Anti fragility

In 2012 Nassim Nicholas Taleb published a book called Antifragile. The idea of 'Antifragile' is that the typical binary opposites of fragile and robust is actually an incomplete way of considering how the world works, particularly in unstable or rapidly changing environments. Taleb posits that 'robust' actually is a midway point between the opposites of 'fragile' - where things are damaged by instability - and 'anti-fragile' - where things benefit from instability.

Taleb has written for some time on problem of using averages and historical data to predict risk profiles, particularly when considering unlikely but catastrophic events. Building anti fragility so that you are likely to benefit more in unstable times is a better long-term strategy than optimising for current or average conditions. Ultimately his idea of "anti-fragility" can be simply described as "having lots of positive options no matter what happens."

Taleb recommends what he calls the 'barbell strategy', where one avoids investing too heavily in the 'average' and instead spends time, money or effort on both things that seem to be very low risk and things that seem to be very high risk but have a potential payoff that is also very high.

An example in the financial world would be the 90/10 split Taleb recommended in his earlier book The Black Swan where you invest 90% of your money in very safe things like Treasury Bonds and gold, and then spread the remaining 10% amongst a bunch of very risky stocks or currencies that will probably fail but have the potential to make an enormous return.

Considering we are in one of the most unstable and rapidly changing information and professional environments librarians have ever faced, the idea of 'anti fragility' seems rather attractive.

So how do we do this in practice?

Antifragile collections

Wilkin profiles are a concept explained by Dan Cohen in 2012. Named after John Wilkin, a 'Wilkin Profile' essentially shows how unique a library's collection is compared to any given group of libraries.

A library with lots of titles that are only held by a few other libraries in the set will have a 'left leaning profile'. A library with mostly titles also held by many other libraries will have a 'right leaning profile'. A library that has a few unique titles and a few popular ones but mostly an average sort of collection will have a 'rounded profile'.

If we took the Taleb 'barbell' approach to collection management, the Wilkin Profile would look like a left-leaning barbell. This is something we have started doing with our hardcopy reference collection at my library service. In public libraries a hardcopy non-lending reference collection is really used for two things - quick reference where it is easier and/or faster to use hardcopy than online (e.g. language dictionaries, thesauri) and questions with complex or obscure answers that can't be easily answered with freely available online information (which manufacturer does that silver mark belong to? what will the tide line be at this location in two weeks? When was this local building constructed?).

Access as preservation

With hardcopy collections, generally speaking access and preservation outcomes are inversely related - the more easily people can access a document, the less well preserved it is likely to be due to physical wear and tear, light damage and so on.

With the technology we have now, we can have it both ways. By making digital copies of rare or fragile material we end up with a better preservation outcome and more access because we can store the original in optimum conditions and provide easy and widely distributed digital access simultaneously. On the rare occasions a scholar needs to see the actual copy, it will be much better preserved.

When it comes to born-digital material, we need to act in a different way. Ed Summers, in his talk last year on The Web as a Preservation Medium quoted his colleague David Brunton as saying "Digital Preservation is just access ...in the future." Summers went on to say that

The underlying implication here is that if you are not providing meaningful access in the present to digital content, then you are not preserving it. - Ed Summers

I would go further than Summers and Brunton and say that if libraries are not providing Open Access to digital content, then they are not preserving it. A great deal has been said over the last decade about Open Access journals and more recently Open Access monographs. One thing that is sometimes missed in this conversation is that open access isn't just a way for scholars to access information about the current state of science and humanities knowledge. The truly open access models using Creative Commons licenses or Public Domain transfers are by their very design widely distributed and thus more likely to be preserved over the long term - especially if they use web-native open standards like HTML.

Antifragile metadata

Speaking of web-native open standards brings us to the idea of anti fragile metadata or, if you prefer, modern cataloguing standards.

Linked open data projects like the Library of Congress' BIBFRAME and OCLC's Schema.org extension proposal are really useful because they connect longstanding library standards - in this case MARC - with more widely used (but less established) web standards to open up the metadata. HTML is very stable, much more so than the hardware and software that interprets it. If we are to truly open up our libraries and our records of what is in our collections, we need to be using the international language of the web - and that's not MARC or Library of Congress subject headings!

The beauty of a system like Schema.org for linked data is that the more people using it, the more useful it becomes. Because it is native to the web and integrated into the structure of HTML, aligning our catalogues with the Schema.org standard will make our collections vastly more visible through search engines and other tools, yet we don't have to create or fund a whole new infrastructure in order to use it.

I have a lot more to say about metadata, but my time today is limited. Since this webinar is for new professionals, I simply encourage you to spend as much time as you can learning how to create and disseminate high quality metadata, because that's where the interesting and important jobs are going to be. That's because libraries are increasingly going to operate as platforms.

Libraries as Platforms - an anti fragile idea

Tim Sherrat, the manager of the National Library of Australia's Trove portal wrote in November about libraries as platforms Sherrat references discussions by David Weinberger and Ed Summers, where they talked about libraries (and particularly large national and regional libraries) as platforms rather than thinking of them, in the way many of us do, as 'portals'. The platform idea sees libraries as acting more like an essential service - the water pipes or electrical wires - upon which our communities build exciting things. Instead of trying to control the experience of using our collections and services, the idea is to open them up to as many uses as possible, including ones that we have not considered or understood. This makes the library an anti-fragile organisation - the more people and projects using the library in new ways, the stronger and more useful it becomes.

If we think of ourselves as a platform then the key task is to open up our collections using high quality metadata and simple access protocols, and allow others to build discovery experiences on top of that. (I'm not just talking about software!)

In conclusion

  • Open Platforms are the future of libraries
  • "soft tech" is much more important to strong communities than "hard tech"
  • If you want to build anti-fragile communities and library services, when you consider a new technology, instead of asking "what are the features?" you should ask "what are the consequences?".