Image

A Librarian Reads the News

Publications and Presentations

In my classroom...

I coded from an API for the first time, and so can you!

A couple of weeks ago, my library upped our Libguides instance to Libguides CMS, which means that a few weeks ago I came to be in charge of a system that has APIs for the first time in my life. We got the system so that we can pull more robust information from Libguides into our discovery system (so that’s our systems librarian’s domain, not mine), but there have been a couple of tiny things I’ve wanted to do that would be so much better with an API than with the built-in widgets, so I cracked my knuckles and set to learning.

Here’s the first tiny project I did, narrated so that you can do something similar if you, like me, are not an expert in web coding but have access to Libguides APIs.

Building an API-fed dropdown menu

I know, I know, there are dropdown menus of guides available in the built-in widget builder, but I don’t like how they look, and I don’t like that you need the “Go” button to make them go. So this made a perfect first toe-in-the-water project for me.

Getting data via the API

First, I needed to figure out how to make an API request. Libguides CMS “Endpoints v1.1” has a section to GET Guides, a list of “parameters,” an example request, an example return, and not much else. It took me a while to figure out that the example request didn’t work because it was asking for 2 specific guide IDs — IDs for guides that do not exist in our system. Once I deleted those two numbers I could start building requests that actually worked.

Then I figured out that you add a parameter by adding “&parameter_name=parameter_value” to that base URL. So http://lgapi-us.libapps.com/1.1/guides/?site_id=…=owner&sort_by=name would take all our guides and sort them by guide name. From there I could happily keep adding parameters, and if I wanted both “Course” and “Subject” guides but not other guides, I could put a comma between the parameter values “2” and “3” to get “&guide_types=2,3” in my request URL.

Other useful terminology I learned along the way that will help my future Googling includes:

In the end, I created an API request for published guides, sorted by name, and filtered to just those having a particular tag. Here’s an example record from the JSON output I got from that request:

Using the API-generated data to feed a dropdown menu

Then came the searching around through StackOverflow for examples of code that uses a URL to point to JSON-formatted information, examples of code that use JSON fields to populate a dropdown menu, and examples of code that use javascript to add an “event sniffer” to a dropdown menu so that when a user selects an option, the menu opens a new URL without requiring anything else (like clicking a pesky “go” button). This step took me a while… In the end, I fiddled and fiddled with example code until all of a sudden, bits and pieces started to work! So exciting! And little by little I arrived at code that works for me.

Here’s an annotated version of what I built (and an html document you can download and mess with).

(And if you are someone who actually knows what they’re doing, and you see that I made dumb mistakes/choices, please let me know! I’m eager to learn.)

Information Architecture, from information about architecture

I’ve just started work with a team of people who are going to be redoing our library website’s information architecture. Maybe some design, but mostly the structure. So of course I do that thing that I do so often and try to figure out how to think about the problem at hand before actually diving into the problem at hand. And this time I seem to be doing that by reading about architecture.

I’m not the first one to do this, by any means. There’s a great talk by Dr. Molly Wright Steenson from this month’s MinneWebCon that lays out lineage of current artificial intelligence work, including the inheritance from architecture. And her talk reminded me that this is exactly what I needed to think about again in order to think about our website structure.

So I went down into the stacks and checked out Christopher Alexander’s two most famous works, A Pattern Language and The Timeless Way of Building. These are two very very large works of architectural philosophy, so there’s no way that I’ll be able to read and apply them thoroughly to my current project, but something is better than nothing, right?

And then, on page 55 of The Timeless Way of Building, the primary claim from Chapter Four jumped out at me:

We must begin by understanding that every place is given its character by certain patterns of events that keep on happening there.

This is “the user is not broken” stated a different way. Spaces are set up, and people will behave in relatively consistent ways in those spaces depending on the affordances of the environment. If they do things we think are wrong, and especially if they do them consistently, then we have clearly set up a space that invites those events. The user was not broken; our design was broken. But Alexander’s formlation works better as an aspiration to achieve rather than a resignation to past failures. If we can know these patterns that environments evoke, we can design environments that evoke patters we want. If we set up a space (physical or digital) that invites the kinds of events that our community wants and needs (actions, reactions, experiences), and invites these events consistently, then that is a space that has Alexander’s “quality that cannot be named” that animates every successful design. (He goes into detail about definitions of that quality in chapter 2, but I won’t do the same here. You can go puzzle over that chapter if you want to, but my big short-hand for it all is “it works beautifully.”)

He then usefully qualifies his thesis:

This does not mean that space creates events, or that it causes them.

For example, in a modern town, the concrete spatial pattern of a sidewalk does not ’cause’ the kinds of human behavior that happen there.

What happens in much more complex. The people on the sidewalk, being culture-bound, know that the space which they are part of is a sidewalk, and, as part of their culture, they have the pattern of a sidewalk in their minds. (72, emphasis original)

And, in the same way, the patterns of events which govern life… cannot be separated from the space where they occur. (73, emphasis original)

So what are similar structural digital elements that we could switch out for “sidewalk” here? Search box? Menu? Bullet point? We have to keep in mind that in our culture, these things call on a whole network of cultural-bound events. We cannot say “well, we know that search boxes you’re familiar with when doing research are boxes that search through the full text or at least a paragraph of text, but this particular search box searches through a list of library research tools instead.” (Incidentally, this is part of what makes the MLA International Bibliography so hard to use and teach.) People will fall into patterns and perform topic searches in that search box, and they will do so over and over again. And they will wonder why our search box is broken. These patterns are patterns that we can often predict, and good intentions won’t matter if you ignore these predictions.

That said, we can help shape the patterns that happen again and again. For example, we don’t have to blindly follow our novice patrons to their best-known patterns just because many of our patrons are novices. We have to know what they expect and their habitual patterns, and also the expectations and patterns of the experts on our campus, and we have to create systems that invite useful events for all. This may mean educational events that help the novices learn the patterns that are typical of the new culture they have entered. This may involve educational events that help experts navigate evolving systems and patterns and academic cultures. But these educational events can’t happen unless we build bridges for the inhabitants of our spaces (digital and physical) from the patterns they have known to the patterns they are developing.

Alexander, Christopher. 1977. A Pattern Language: Towns, Buildings, Construction. New York: Oxford University Press.

———. 1979. The Timeless Way of Building. New York: Oxford University Press.

The Library Society of the World: Origin Story

Eleven years ago, in a browser far, far away, a dedicated and creative public librarian named Josh Neff opened his ALA membership renewal notice and sighed as he looked at the dues he would owe. Librarians just don’t make a lot of money, and memberships are costly. Meanwhile, he and his buddy Steve Lawson, library blogger extraordinaire, were joking on Twitter (because they were awake and near computers), and out of this joking came the tweet that changed the world. Josh tweeted that he and Steve were starting the Library Society of the World, a society that would never charge any dues and where people could join simply by saying they were members. Josh’s friend from library school, your humble Pegasus Librarian, and several others heeded the call, pledged allegiance, selected association titles for themselves, and thus the LSW was born.

That same day, because this was the early 2000s, we started a wiki. One simply could not have a library organization and not have a wiki, after all. And of course we started coming up with our organization structure and policies. As you can see from our Policies page, this was a very serious effort involving much committee work and wordsmithing.

Within the first month or two we had opened up a Meebo Room where most of us spent most of our time. (For those of you who don’t remember Meebo, this was a free group chat platform.) Within the first year many of us had colonized FriendFeed, which ultimately became our main home. (For those of you who don’t remember FriendFeed, I’m still too grief-stricken over its closing to go into details.)

We introduced ourselves to the library world at large during a raucous presentation at Computers In Libraries, 2008, where the Meebo Room chat was projected live during the session so that all of us could participate simply by typing into our browsers. Then in 2010 Josh and Steve were named Movers and Shakers for their work creating this society, and in response they (and we) developed the Shovers and Makers list, nominating ourselves and writing down what we appreciated about each other. (There was a blog for that, but it has since gone dark.)

Between then and the day FriendFeed shut down in April of 2015, life continued. People got married, had babies, changed jobs, experienced loss, got divorced, struggled, triumphed, and committed daily acts of superior librarianship. Several left the profession and the group. Others stayed in the profession but stopped being active participants. Still others joined us to bring their fresh voices and perspectives from all corners of the world and librarianship.

In April 2015, FriendFeed shut down and we began nearly a year of diaspora before finally landing on Mokum.place/lsw in February of 2016. And we have lived in this peaceful place ever since, sharing ideas, asking each other for opinions and advice, and being together through the ebbs and flows of our careers and lives.

We’re a good society. You should join us. There are no application procedures and no dues. You don’t even have to be a librarian if you can put up with all the librarian-speak. Just say hi and you’re in.

 

What advanced researchers need to know about using the MLA International Bibliography

I live and breathe the MLA International Bibliography. I’m liaison to literature departments, language departments, and linguistics, so it’s pretty much home base for me. And using it is one of the reasons I’m very very grateful to have taken some cataloging/indexing courses while getting my library degree. I will never have the skills necessary to be a cataloger/indexer (those people are amazing!), but a few years into teaching the MLA-IB I realized that one of those pesky rules that I hadn’t quite grasped was completely ruling my chances of success in that database. This is still very much a librarians’ database, and it’s the only database where I find it important to explain its history to my upper level students so that they can have any chance of mining its treasures.

Three pieces of history matter a lot for this database:

  • indexing rules
  • the long legacy of the print index
  • and the almost complete lack of abstracts in the index.

Remember back in indexing class when people dozed off while the professor said something about “aboutness” and about balancing that with local policies about how many indexing terms to assign to any individual item? And remember how, back in the days of the print index, every item had to be listed under the subject headings it was assigned, resulting in increased use of paper directly proportional to the number of subjects it got? I don’t know the MLA’s internal policies on numbers of subjects, but I do know that indexers have it deeply ingrained that they are to assign the fewest number of terms that most completely describe the piece. Every term assigned should describe the full work (its “aboutness”) rather than just parts of the work, and there really shouldn’t be more than a few terms from the same term hierarchy level on any item. The reasoning being that if you need, say, 5 sub-terms from the hierarchy to describe the work, then really you should just use the next broader term because clearly that term is better at expressing the work’s aboutness. Historically, adding additional terms was a very real, practical matter of the physics and economics of publishing, and even now there’s a limit to how many terms it’s feasible to attach to any given record.

And sure, these things exist for most other indexes that have a long history, but the MLA-IB exacerbates all this by mostly not including abstracts. So people can’t kind of muddle through on keyword searching as much because there are just so many fewer words involved in the first place. Not much of a safety net. Not only that, but high end modern systems (like Google) have background thesauri so that if someone types in “heart attack” it’ll say, “Oh, that’s the same is myocardial infarction, so I’ll return those results, too, and maybe this person is confused about the exact cause of the heart attack, so let’s throw in some information about arterial sclerosis.” (They hire computational linguists for a reason! Also, sometimes I hate this “help” so very much, but that’s a rant for a different day — basically I want it to know when to do it and when not to do it.) So that’s even more safety nets. But back in the MLA-IB, lacking all these safety nets, people really have to pay attention to the indexing in a very literal way.

Here’s how you can see that playing out. Search for “Woolf, Virginia” as a subject in the MLA-IB. You’ll get several thousand results because scholars love to study her work. Now, the first thing to note is that (as we all know) works that simply mention her somewhere in the argument won’t list her as a Subject. She gets listed if the full work’s aboutness depends on her. So that’s already a caveat to note for future search strategies.

But let’s say you really want works that are really ABOUT Virginia Woolf, and you would like to explore the scope of what people have written about gender and her work. That’s pretty typical, right? She’s a cornerstone of feminist/queer literary criticism for a reason. But if you search for the Subject “Woolf, Virginia” and then AND in the keyword “gender” you get 250 results. So I tell students to use their critical thinking skills and think about this result list. Is it really likely that less than 4% of literary criticism addresses Virginia Woolf’s depictions of gender? No, it is not very likely. Not likely at all. So this is where those indexing rules kick in.

If the indexer decided that the work was about gender – nothing more specific than that – then the indexer would assign that subject heading. But scholars are usually much more specific than that. They narrow their topics. They write about Virginia Woolf’s depiction of women, men, boys, girls, masculinity, femininity, sex, sexuality, gender, misogyny, ….. So only if a work of scholarship dealt significantly with several of those narrower terms would the indexer move up the hierarchy to “gender.” Relatively rare, right? So the way to get around this is to search for narrower terms ORed together (I sometimes call this “teaching the computer what I mean by gender”). And remember to OR in some narrower terms that are the opposite of the main topic, too. So if I’m interested in the depiction of femininity, articles on masculinity can be almost equally useful, or they can point me to useful literature through their bibliographies.

One more thing to note: this applies to authors and their works as well. If you’re searching for a particular poem, you can OR together the poem and the name of the collection it was originally published with (if applicable) and the name of the poet if the previous two don’t pull in enough results, or OR together specific shakespeare tragedy play names if you’re interested in Shakespearean tragedy, or specific European country names plus “Europe.” You get the idea. The MLA-IB is not a database that functions in a post-boolean, single search box world.

Which brings us to taking notes about terms associated with your topic. This is just utterly and completely necessary if you’re going to spend time in the MLA-IB. As you’re reading and searching, keep adding to your list of words associated with your topic and play around with those in your searches. And watch bibliographies for useful new vocabulary, because unlike search (which is just matching letters-in-a-row in a very, very literal way), bibliographies are compiled by humans who know the field and can tell if a work is related by more than just pure letter-matching.

For really advanced students starting long-term research, this is where I teach them how to save searches and do advanced searches that link previous searches. So say you spent a really long time developing an extensive ORed together “definition” of what you mean by “gender.” Save that search. Then you can go into your search history and AND it together with other topics (searches for Mrs. Dalloway, maybe, or To the Lighthouse). That way you can pretty efficiently explore how your broad topic looks when applied to various literary works or genres or themes or whatever.

Three final mini-tips, especially for foreign language literature:

  1. The MLA-IB doesn’t parse accent marks well (they were entered in various encodings over the years, if entered at all), so you pretty much always get better results if you leave them out.
  2. Translated works are indexed under their translated name and their original name. Find that original name and do your subsequent searches using that original name as a subject. If things seem inconsistent, try ORing together both names.
  3. Don’t want just things that are ABOUT Virginia Woolf, but also anything that mentions her? That’s when you skip out of the MLA-IB into a full text search system like JSTOR or Project Muse. No one tool can do it all.

So there. That’s pretty much my one area of “expertise” as a librarian for literature. So now you know all my secrets.