The State of the Library Website

T’was a time when the Library Website was an abomination. Those dark days have lightened significantly. But new clouds have appeared on the horizon.

Darkest Before the Dawn

In the dark ages of Library Websites, users suffered under UX regimes that were rigid, unhelpful and confusing. This was before responsive design became a standard in the library world. It was before search engine optimization started to creep into Library meetings. It was before user experience became an actual librarian job title.

We’ve come a long way since I wrote The Ugly Truth About Library Websites. Most libraries have evolved beyond the old “website as pamphlet” paradigm to one that is dynamic and focused on user tasks.

Public libraries have deployed platforms like BiblioCommons to serve responsive, task-oriented interfaces that integrate their catalogs, programming and website into a single social platform. Books, digital resources, programs and even loanable equipment are all accessible via a single search. What’s more, the critical social networking aspects of library life are also embedded along the user’s path. Celebrated examples of this integrated solution include the San Francisco Public Library and Chicago Public Library. Queens is also hard at work to develop a custom solution.

In the academic realm, libraries have turned to unified discovery layers like WorldCat Discovery and EBSCO Discovery Service to simplify (Googlize) the research process. These systems put a single-search box front and center that access resources on the shelf, but also all those electronic resources that make up the bulk of academic budgets.

And while there are still many laggards, few libraries ignore these problems outright.

The Storm Ahead

While the general state of online library interfaces has improved, the unforgiving, hyperbolic curve of change continues to press forward. And libraries cannot stay put. Indeed, we need to quicken our pace and prepare our organizations for ongoing recalibration as the tempo of change increases.

The biggest problem for library websites, is that there is little future for the library website. That’s because people will get less and less information through web browsers. Indeed, consider how often you use a web browser on your phone versus an app. Developments in AI, Augmented Reality and Virtual Reality will compound that trend.

If you’re like Chris Milk, videographer and VR evangelist, you see the writing on the wall. The modes of how we experience information are about to undergo a fundamental revolution. Milk likens the current state of VR to the old black and white silent films at the dawn of motion pictures.

I’d extend this line of thinking to the web page. Within a decade or two, I expect people will look back on web pages as a brief, transitory medium bridging print information to linked data. And as our AI, VR and AR technologies take off, they will liberate information from the old print paradigms altogether.

In short, people will interact with information in more direct ways. They will ask a computer to provide them the answer. They will virtually travel to a “space” where they can experience the information they seek.

Get Ready to Re-invent the Library…again

So where does the library fit into this virtualized and automated future?

One possibility is that the good work to transform library data into linked data will enable us to survive this revolution. In fact, it may be our best hope.

Another hope is that we continue to emphasize the library as a social space for people to come together around ideas. Whether its a virtual library space or a physical one, the library can be the place in both local and global communities where people meet their universal thirst for connecting with others. The modes of those ideas (books, ebooks, videos, games) will matter far less than the act of connecting.

In a sense, you could define the future online library as something between an MMORPG, Meetup.com and the TED conference.

So, the library website is vastly improved, but we won’t have long to rest on our laurels.

Ready Player One? Put on your VR goggles. Call up Siri. Start rethinking everything you know about the Library website.

 

 

AI First

Looking to the future, the next big step will be for the very concept of the “device” to fade away. Over time, the computer itself—whatever its form factor—will be an intelligent assistant helping you through your day. We will move from mobile first to an AI first world.

Google Founder’s Letter, April 2016

My Library recently finalized a Vision Document for our virtual library presence. Happily, our vision was aligned with the long-term direction of technology as understood by movers and shakers like Google.

As I’ve written previously, the Library Website will disappear. But this is because the Internet (as we currently understand it) will also disappear.

In its place, a new mode of information retrieval and creation will move us away from the paper-based metaphor of web pages. Information will be more ubiquitous. It will be more free-form, more adaptable, more contextualized, more interactive.

Part of this is already underway. For example, people are becoming a data set. And other apps are learning about you and changing how they work based on who you are. Your personal data set contains location data, patterns in speech and movement around the world, consumer history, keywords particular to your interests, associations based on your social networks, etc.

AI Emerging

All of this information makes it possible for emerging AI systems like Siri and Cortana to better serve you. Soon, it will allow AI to control the flow of information based on your mood and other factors to help you be more productive. And like a good friend that knows you very, very well, AI will even be able to alert you to serendipitous events or inconveniences so that you can navigate life more happily.

People’s expectations are already being set for this kind of experience. Perhaps you’ve noticed yourself getting annoyed when your personal assistant just fetches a Wikipedia article when you ask it something. You’re left wanting. What we want is that kernel of gold we asked about. But what we get right now, is something too general to be useful.

But soon, that will all change. Nascent AI will soon be able to provide exactly the piece of information that you really want rather than a generalized web page. This is what Google means when they make statements like “AI First” or “the Web will die.” They’re talking about a world where information is not only presented as article-like web pages, but broken down into actual kernels of information that are both discrete and yet interconnected.

AI First in the Library

Library discussions often focus on building better web pages or navigation menus or providing responsive websites. But the conversation we need to have is about pulling our data out of siloed systems and websites and making it available to all modes like AI, apps and basic data harvesters.

You hear this conversation in bits and pieces. The ongoing linked data project is part of this long-term strategy. So too with next-gen OPACs. But on the ground, in our local strategy meetings, we need to tie every big project we do to this emerging reality where web browsers are increasingly no longer relevant.

We need to think AI First.

Your Job Has Been Robot-sourced

rosie-the-robot

“People are racing against the machine, and many of them are losing that race…Instead of racing against the machine, we need to learn to race with the machine.”

– Erik Brynjolfsson, Innovation Researcher

Libraries are busy making lots of metadata and data networks. But who are we making this for anyway? Answer: The Machines

I spent the last week catching up on what the TED Conference has to say on robots, artificial intelligence and what these portend for the future of humans…all with an eye on the impact on my own profession: librarians.

A digest of the various talks would go as follows:

    • Machine learning and AI capabilities are advancing at an exponential rate, just as forecast
    • Robots are getting smarter and more ubiquitous by the year (Roomba, Siri, Google self-driving cars, drone strikes)

Machines are replacing humans at an increasing rate and impacting unemployment rates

The experts are personally torn on the rise of the machines, noting that there are huge benefits to society, but that we are facing a future where almost every job will be at risk of being taken by a machine. Jeremy Howard used words like “wonderful” and “terrifying” in his talk about how quickly machines are getting smarter (quicker than you think!). Erik Brynjolfsson (quoted above) shared a mixed optimism about the prospects this robotification holds for us, saying that a major retooling of the workforce and even the way society shares wealth is inevitable.

Personally, I’m thinking this is going to be more disruptive than the Industrial Revolution, which stirred up some serious feelings as you may recall: Unionization, Urbanization, Anarchism, Bolshevikism…but also some nice stuff (once we got through the riots, revolutions and Pinkertons): like the majority of the world not having to shovel animal manure and live in sod houses on the prairie. But what a ride!

This got me thinking about the end game the speakers were loosely describing and how it relates to libraries. In their estimation, we will see many, many jobs disappear in our lifetimes, including lots of knowledge worker jobs. Brynjolfsson says the way we need to react is to integrate new human roles into the work of the machines. For example, having AI partners that act as consultants to human workers. In this scenario (already happening in healthcare with IBM Watson), machines scour huge datasets and then give their advice/prognosis to a human, who still gets to make the final call. That might work for some jobs, but I don’t think it’s hard to imagine that being a little redundant at some point, especially when you’re talking about machines that may even be smarter than their human partner.

But still, let’s take the typical public-facing librarian, already under threat by the likes of an ever-improving Google. As I discussed briefly in Rise of the Machines, services like Google, IBM Watson, Siri and the like are only getting better and will likely, and possibly very soon, put the reference aspect of librarianship out of business altogether. In fact, because these automated information services exist on mobile/online environments with no library required, they will likely exacerbate the library relevance issue, at least as far as traditional library models are concerned.

Of course, we’re quickly re-inventing ourselves (read how in my post Tomorrow’s Tool Library on Steroids), but one thing is clear, the library as the community’s warehouse and service center for information will be replaced by machines. In fact, a more likely model would be one where libraries pool community resources to provide access to cutting-edge AI services with access to expensive data resources, if proprietary data even exists in the future (a big if, IMO).

What is ironic, is that technical service librarians are actually laying the groundwork for this transformation of the library profession. Every time technical service librarians work out a new metadata schema, mark up digital content with micro-data, write a line of RDF, enhance SEO of their collections or connect a record to linked data, they are really setting the stage for machines to not only index knowledge, but understand its semantic and ontological relationships. That is, they’re building the infrastructure for the robot-infused future. Funny that.

As Brynjolfsson suggests, we will have to create new roles where we work side-by-side with the machines, if we are to stay employed.

On this point, I’d add that we very well could see that human creativity still trumps machine logic. It might be that this particular aspect of humanity doesn’t translate into code all that well. So maybe the robots will be a great liberation and we all get to be artists and designers!

Or maybe we’ll all lose our jobs, unite in anguish with the rest of the unemployed 99% and decide it’s time the other 1% share the wealth so we can all, live off the work of our robots, bliss-out in virtual reality and plan our next vacations to Mars.

Or, as Ray Kurzweil would say, we’ll just merge with the machines and trump the whole question of unemployment, let alone mortality.

Or we could just outlaw AI altogether and hold back the tide permanently, like they did in Dune. Somehow that doesn’t seem likely…and the machines probably won’t allow it. LOL

Anyway, food for thought. As Yoda said: “Difficult to see. Always in motion is the future.”

Meanwhile, speaking of movies…

If this subject intrigues you, Hollywood is also jumping into this intellectual meme, pushing out several robot and AI films over the last couple years. If you’re interested, here’s my list of the ones I’ve watched, ordered by my rating (good to less good).

  1. Her: Wow! Spike Jonze gives his quirky, moody, emotion-driven interpretation of the AI question. Thought provoking and compelling in every regard.
  2. Black Mirror, S02E01 – Be Right Back: Creepy to the max and coming to a bedroom near you soon!
  3. Automata: Bleak but interesting. Be sure NOT to read the expository intro text at the beginning. I kept thinking this was unnecessary to the film and ruined the mystery of the story. But still pretty good.
  4. Transcendence: A play on Ray Kurzwell’s singularity concept, but done with explosions and Hollywood formulas.
  5. The Machine: You can skip it.

Two more are on my must watch list: Chappie and Ex Machina, both of which look like they’ll be quality films that explore human-robot relations. They may be machines, but I love when we dress them up with emotions…I guess that’s what you should expect from a human being. 🙂

Digital Author Services

The producers of information at our academic institutions are brilliant at what they do, but they need help from experts in sharing their work online. Libraries are uniquely suited for the task.

There are three important areas where we can help our authors:

  1. Copyright and Author Rights Issues
  2. Developing Readership and Recognition
  3. Helping authors overcome technical hurdles to publishing online

Copywhat?

Several libraries are now promoting copyright and author rights information services. These services provide resources (often LibGuides) to scholars who may be sold on the benefits of publishing online, but are unclear what their publishers allow. In fact, in my experience, this is one of the most common problems. Like I said, academics are busy people and focused on their area of specialization, which rarely includes reading the legalese of their publisher agreements, let alone keeping a paper trail handy. This is particularly true for authors that began their careers before the digital revolution.

At any rate, providing online information followed up with face-to-face Q&A is an invaluable service for scholars.

Lucretia McCulley of the University of Richmond and Jonathan Bull of the University of Valpraiso have put together a very concise presentation on the matter, detailing how they’ve solved these issues at their institutions.

Another service, which I’m actually developing at my institution presently, is providing copyright clearance as a service for scholars. In our case, I hope to begin archiving all faculty works in our institutional repository. The problem has been that faculty are busy and relying on individual authors to find the time to do the due diligence of checking their agreements just ain’t gonna happen. In fact, this uncertainty about their rights as authors often stops them cold.

In the service model I’m developing, we would request faculty activity reports or query some other resource on faculty output and then run the checks ourselves (using student labor) on services like SherpaRomeo. When items check out, we publish. When they don’t we post the metadata and link to the appropriate online resource (likely in an online journal).

Developing Readership & Recognition

Another area where library’s can provide critical support is assisting authors in growing their reputations and readership. Skills commonly found in libraries from search engine optimization (SEO) to cataloging play a role in this service offering.

At my institution, we use Digital Commons for our repository, which we selected partly because it has powerful SEO built into it. I’ve seen this at work: where a faculty posts something to the repository and within weeks (and even days), that content is rising to the top of Google search results, beating out even Facebook and LinkedIn for searches on an author’s name.

And of course, while we don’t normally mark up the content with metadata for the authors, we do provide training on using the repository and understanding the implications for adding good keywords and disciplines (subject headings) which also help with SEO.

The final bit, is the reporting. With Digital Commons, reports come out every month via email to the authors, letting them know what their top downloads were and how many they had. This is great and I find the reports help spur word-of-mouth marketing of the repository and enthusiasm for it by authors. This is built into Digital Commons, but no matter what platform you use, I think this is just a basic requirement that helps win author’s hearts, drives growth and is a vital assessment tool.

Walking The Last Mile

MacKenzie Smith of MIT has described the Last Mile Problem (Bringing Research Data into the Library, 2009), which is essentially where technical difficulties, uncertainty about how to get started and basic time constraints keep authors from ever publishing online.

As I touched on above, I’m currently developing a program to help faculty walk the last mile, starting with gathering their CVs and then doing the copyright checks for them. The next step would be uploading the content, adding useful metadata and publishing it for them. A key step before all of this, of course, is setting up policies for how the collection will be structured. This is particularly true for non-textual objects like images, spreadsheets, data files, etc.

So, when we talk about walking the last mile with authors, there’s some significant preparatory work involved. Creating a place for authors to understand your digital publishing services is a good place to start. Some good examples of this include:

Once your policies are in place, you can provide a platform for accepting content. In our case (with Digital Commons), we get stellar customer service from Bepress which includes training users how to use their tools. At institutions where such services is not available, two things will be critical:

  1. Provide a drop-dead easy way to deposit content, which includes simple but logical web forms that guide authors in giving you the metadata and properly-formatted files you require.
  2. Provide personal assistance. If you’re not providing services for adding content, you must have staffing for handling questions. Sorry, an FAQ page is not enough.

Bottom Line

Digital publishing is just such a huge area of potential growth. In fact, as more and more academic content is born digital, preserving it for the future in sustainable and systematic ways is more important than ever.

The Library can be the go-to place on your campus for making this happen. Our buildings are brimming with experts on archives, metadata, subject specialists and web technologies, making us uniquely qualified to help authors of research overcome the challenges they face in getting their stuff out there.

Library as Digital Consultancy

As faculty and students delve into digital scholarly works, they are tripping over the kinds of challenges that libraries specialize in overcoming, such as questions regarding digital project planning, improving discovery or using quality metadata. Indeed, nobody is better suited at helping scholars with their decisions regarding how to organize and deliver their digital works than librarians.

At my institution, we have not marketed our expertise in any meaningful way (yet), but we receive regular requests for help by faculty and campus organizations who are struggling with publishing digital scholarship. For example, a few years ago a team of librarians at my library helped researchers from the University of Ireland at Galway to migrate and restructure their online collection of annotations from the Vatican Archive to a more stable home on Omeka.net. Our expertise in metadata standards, OAI harvesting, digital collection platforms and digital project planning turned out to be invaluable to saving their dying collection and giving it a stable, long-term home. You can read more in my Saved by the Cloud post.

These kinds of requests have continued since. In recognition of this growing need, we are poised to launch a digital consultancy service on our campus.

Digital Project Planning

A core component of our jobs is planning digital projects. Over the past year, in fact, we’ve developed a standard project planning template that we apply to each digital project that comes our way. This has done wonders at keeping us all up to date on what stage each project is in and who is up next in terms of the workflow.

Researchers are often experts at planning out their papers, but they don’t normally have much experience with planning a digital project. For example, because metadata and preservation are things that normally don’t come up for them, they overlook planning around these aspects. And more generally, I’ve found that just having a template to work with can help them understand how the experts do digital projects and give them a sense of the issues they need to consider when planning their own projects, whether that’s building an online exhibit or organizing their selected works in ways that will reap the biggest bang for the buck.

We intend to begin formally offering project planning help to faculty very soon.

Platform Selection

It’s also our job to keep abreast of the various technologies available for distributing digital content, whether that is harvesting protocols, web content management systems, new plugins for WordPress or digital humanities exhibit platforms. Sometimes researchers know about some of these, but in my experience, their first choice is not necessarily the best for what they want to do.

It is fairly common for me to meet with campus partners that have an existing collection online, but which has been published in a platform that is ill-suited for what they are trying to accomplish. Currently, we have many departments moving old content based in SQL databases to plain HTML pages with no database behind them whatsoever. When I show them some of the other options, such as our Digital Commons-based institutional repository or Omeka.net, they often state they had no idea that such options existed and are very excited to work with us.

Metadata

I think people in general are becoming more aware of metadata, but there is still lots of technical considerations that your typical researcher may not be aware of. At our library, we have helped out with all aspects of metadata. We have helped them clean up their data to conform to authorized terms and standard vocabularies. We have explained Dublin Core. We have helped re-encode their data so that diacritics display online. We have done crosswalking and harvesting. It’s a deep area of knowledge and one that few people outside of libraries know on a suitably deep level.

One recommendation for any budding metadata consultants that I would share is that you really need to be the Carl Sagan of metadata. This is pretty technical stuff and most people don’t need all the details. Stick to discussing the final outcome and not the technical details and your help will be far more understood and appreciated. For example, I once presented to a room of researchers on all the technical fixes to a database that we made to enhance and standardize the metadata, but his went over terribly. People later came up to me and joked that whatever it was we did, they’re sure it was important and thanked us for being there. I guess that was a good outcome since they acknowledged our contribution. But it would have been better had they understood, the practical benefits for the collection and users of that content.

SEO

Search Engine Optimization is not hard, but it is likely that few people outside of the online marketing and web design world know what it is. I often find people can understand it very quickly if you simply define it as “helping Google understand your content so it can help people find you.” Simple SEO tricks like defining and then using keywords in your headers will do wonders for your collection’s visibility in the major search engines. But you can go deep with this stuff too, so I like to gauge my audience’s appetite for this stuff and then provide them with as much detail as I think they have an appetite for.

Discovery

It’s a sad statement on the state of libraries, but the real discovery game is in the major search engines…not in our siloed, boutique search interfaces. Most people begin their searches (whether academic or not) in Google and this is really bad news for our digital collections since by and large, library collections are indexed in the deep web, beyond the reach of the search robots.

I recently tried a search for the title of a digital image in one of our collections in Google.com and found it. Yeah! Now I tried the same search in Google Images. No dice.

More librarians are coming to terms with this discovery problem now and we need to share this with digital scholars as they begin considering their own online collections so that they don’t make the mistakes libraries made (and continue to make…sigh) with our own collections.

We had one department at my institution that was sitting on a print journal that they were considering putting online. Behind this was a desire to bring the publication back to life since they had been told by one researcher in Europe that she thought the journal had been discontinued years ago. Unfortunately, it was still being published, it just wasn’t being indexed in Google. We offered our repository as an excellent place to do so, especially because it would increase their visibility worldwide. Unfortunately, they opted for a very small, non-profit online publisher whose content we demonstrated was not surfacing in Google or Google Scholar. Well, you can lead a horse to water…

Still, I think this kind of understanding of the discovery universe does resonate with many. Going back to our somewhat invisible digital images, we will be pushing many to social media like Flickr with the expectation that this will boost visibility in the image search engines (and social networks) and drive more traffic to our digital collections.

Usability

This one is a tough one because people often come with pre-conceived notions of how they want their content organized or the site designed. For this reason, sometimes usability advice does not go over well. But for those instances when our experiences with user studies and information architecture can influence a digital scholarship project, it’s time well spent. In fact, I often hear people remark that they “never thought of it that way” and they’re willing to try some of the expert advice that we have to share.

Such advice includes things like:

  • Best practices for writing for the web
  • Principles of information architecture
  • Responsive design
  • Accessibility support
  • User Experience design

Marketing

It’s fitting to end on marketing. This is usually the final step in any digital project and one that often gets dropped. And yet, why do all the work of creating a digital collection only to let it go unnoticed. As digital project expert, librarians are familiar with the various channels available to promote and build followers with tools like social networking sites, blogs and the like.

With our own digital projects, we discuss marketing at the very beginning so we are sure all the hooks, timing and planning considerations are understood by everyone. In fact, marketing strategy will impact some of the features of your exhibit, your choice of keywords used to help SEO, the ultimate deadlines that you set for completion and the staffing time you know you’ll need post launch to keep the buzz buzzing.

Most importantly, though, marketing plans can greatly influence the decision for which platform to use. For example, one of the benefits of Omeka.net (rather than self-hosted Omeka) is that any collection hosted with them becomes part of a network of other digital collections, boosting the potential for serendipitous discovery. I often urge faculty to opt for our Digital Commons repository over, say, their personal website, because anything they place in DC gets aggregated into the larger DC universe and has built-in marketing tools like email subscriptions and RSS feeds.

The bottom line here is that marketing is an area where librarians can shine. Online marketing of digital collections really pulls together all of the other forms of expertise that we can offer (our understanding of metadata, web technology and social networks) to fulfill the aim of every digital project: to reach other people and teach them something.

Rise of the Machines

As I write, the Roomba is cleaning my house. Googlebots are driving cars on California roads. Siri is learning what you want.

And, to the dread of many reference librarians, Watson is beating the pants off Jeopardy Champions in an opening AI move that will surely impact the library in the near future.

Already, robot shelvers are in place in many libraries, such as Santa Clara University’s Library. And if you saw the recent executive summary of Library Journal’s Patron Profiles, you saw that 76% of students reported turning to Google first when initiating their research. Compare that to just 24% that opted for the library.

This isn’t news, really, but when I heard economist Paul Krugman connecting the dots of automation, nagging unemployment, innovation and worker productivity and identifying it as a challenge to society, I had to agree with his thesis: robots are replacing people at an ever-increasing rate…and in parts of the economy we once considered safe.

Like I said in my previous post, sometimes the future sneaks up on you. But even if robotification is inevitable, we must ask ourselves, what are the human qualities that make us a value to other people?

Some might say that it’s about the in-person assistance that we can bring to our libraries: true. They might emphasize the smiles, encouraging words and subtle forms of non-verbal communication machines are pretty lousy with so far (until the David 8 release at least).

But we have to be very careful about convincing ourselves that retreating to our ramparts of physicality and empathy will serve us for very long. As the Library Journal survey illustrates, the cold, white Googlean box is often a superior tool than a library website…and quite possibly more approachable than our staff.

No, to be effective and valuable, we have to embrace the shifting technological realm and make it our own…and humanize it, improve it, augment it.

Until David 8, that is…

Omeka.net is Looking Good

We used Omeka.net to help partners at another University publish a database of Catholic letters (not published yet, so it’s not ready to share) and I think everyone was quite taken by the ease of using the hosted version of Omeka. Unlike the locally installed version of Omeka, Omeka.net is a freemium web service that requires little to no coding skills and no server to get started.

Essentially, Omeka.net works like a hosted blog platform such as Blogger or WordPress.com. You have limited options in terms of look and feel, but the underlying collection building tools are there.

What I like about Omeka.net:

  1. Free or very cheap. You can put together a free collection if your needs are small. If, as was our situation, you have to import your metadata via spreadsheet or have some other plugins you need (some you have to pay for an account to use), then you’ll have to pay a subscription fee. But the fees are likely quite affordable for your organization.
  2. Easy peasy. Very little if any understanding of web technologies is required. In our situation, we had some problems that required some technical thinking mostly in terms of cleaning up our spreadsheet , converting it to UTF-8 and making it work with Omeka, but this kind of expertise should be easy to find in your institution. Once the platform is set up, non-technical staff can pretty much run it and add to it without issue.
  3. It’s pretty good and just getting started. In the past few months, new themes and plugins have started to come online and I expect that this platform will be very robust in just another year or two.
  4. Your collection will become part of the Omeka.net network of collections, which should help with Search Engine Optimization as well as serendipitous discovery.
  5. It feels pretty rock-solid and reliable.

What I don’t like about it:

  1. It’s still early days and some features from the full Omeka version are not yet available.
  2. Very limited theming, meaning you have only 8 different themes and none of these let you control color. However, you can add banner images and footers to brand it.
  3. The help documentation is fairly helpful for most situations, but if you run into some advanced issues, you’ll have to hope someone on the Omeka forums can help.