00:00:05.220
Hello and thank you for coming to this pre-recorded version of my talk. This presentation focuses on the discipline of reference librarianship and how we, as developers, can leverage their practices to help others in our community.
00:00:10.320
A little bit about me: my name is Mike Calhoun. I live in Vancouver, Washington, which is just outside of Portland, Oregon, though I'm originally from Syracuse, New York, as denoted by the T-shirt of a local bar I'm wearing in the picture. That bar no longer exists, and this picture was taken two years ago, featuring some pre-COVID hair and my now much larger son.
00:00:29.340
If you find me online, you can look for 'comelike011' in most places. I don’t tweet very much, but I would be thrilled to use that as a vector for any questions or conversations that arise from this talk. Currently, I work for Invoka, an AI-powered call tracking and conversational analytics company based in Santa Barbara, California. We are hiring for many remote positions and are a great place to work.
00:00:54.480
I've been in software engineering since a little after finishing grad school in 2006. Initially, that felt like roughly ten years, but time makes fools of us all, and now it's been 14 years, and software is getting ready to get its learner's permit. Before I started programming, I worked as an assistant archivist for St. John's University in Queens, New York, where I also earned my Master's Degree in Library Science and my State of New York Certified Librarian certificate. This certificate qualifies me to work as a public librarian in most states.
00:01:25.939
The condition of this certificate reflects many changes and moves I've made in my life, as well as some water damage incurred from a burst pipe in 2010. However, what the condition of this artifact doesn't reflect directly is the sense of pride I have in it and the work I've done to earn it. While I worked in the library system as an archivist, I was sometimes required to serve as a reference librarian, which I did in an academic library context in the university archives and as part of an internship with the New York State Supreme Court Law Library.
00:01:56.280
Before we dive deeply into the topic of this talk, I want to highlight a few terms crucial to this discipline. It may feel a little reductive to be this granular, but there are distinctions among the following terms that I believe warrant clarification, as they are often used interchangeably. First of all, data is raw, decontextualized, unorganized material.
00:02:28.920
To provide an example of something I came across while trying to find my previously shown certificate: the year '1990' is just a piece of data, or individually, a datum. We can assume it is a year, but out of context, who knows? It could represent many things, like 1989 plus one; it is, at this point, truly just decontextualized material. Information is data in context; it gives that material additional meaning from which we can learn or use as a foundation for the final part of this journey.
00:02:55.260
For instance, by putting '1990' into context, we can now glean that 1990 refers to a year in which something significant happened. I apologize for reflexively falling back on what is now revealed to be a sports example, but this is knowledge: the familiarity of people, places, events, ideas, issues, or really anything that has been gleaned from information. It's a representation of collecting data into some kind of context as information and applying other information while remembering the learned outcomes.
00:03:28.560
Using the example of 1990 again, knowledge means we knew that a lottery draft took place, and our specific knowledge is that in 1990, Derek Coleman was one of the first 14 picks of the NBA Draft. I found this card I had set aside, along with other treasured items, such as my librarian certificate. This card is from a pack my dad bought me at a gas station down the road from where I grew up. With that context, let's talk for a moment about libraries in the United States.
00:04:03.960
This is a bit of a debated topic, with the expected heavy-hitting cities of early American history contending to have the first library—be it Philadelphia, New York, Boston, etc. However, when discussing the first public library in the United States, I point to Peterborough, New Hampshire, whose library opened in 1833. This was the first public library in the United States that was completely funded by tax dollars, making it a service for the people funded by the people.
00:04:39.720
The Boston Public Library, as organized today, was founded 19 years after Peterborough in 1852. Its first location officially opened its doors to the public in 1854 in an abandoned schoolhouse on Mason Street. I have a lithograph of it here. In 1858, four years after the Mason Street location opened, the Boston Public Library relocated to a more permanent home at 55 Boylston Street. Looking at this facade, I want to draw attention to the arched windows on the second story of this building.
00:05:10.800
There are three of them, framed by ornate stonework. If this were an in-person presentation, I would point to this. This is the entryway to that building where the patrons first entered, featuring mirrored staircases that meet at the top and funnel into a singular staircase, which led to the main hall of the library. The arched windows provide light to the main hall, whose grandiosity reflects the importance of this building at the time.
00:05:55.620
Before addressing the significance of the main hall further, I'd like to talk about Samuel S. Green. Samuel Sweet Green was born in Worcester, Massachusetts, in 1837. His father was a pharmacist, which granted him the privilege and access to attend Harvard, from which he graduated in 1858. He subsequently attended Harvard Divinity School but did not graduate until 1864 because of poor health.
00:06:38.139
In 1867, he was appointed as the Director of the Worcester Free Public Library, likely through nepotism, as his uncle was a lead funder for the library. In this position, he focused on the organization and cataloging of information. In 1876, Green wrote an article titled "Personal Relations between Librarians and Readers" for the American Library Journal. In it, he stated that reference librarians should have four goals: teaching patrons about the library's functions and resources as well as how to utilize them, answering patrons' questions, helping patrons select good reading material, and promoting the library to the greater community.
00:07:15.840
He wrote that a librarian should be as unwilling to allow an inquirer to leave the library with an unanswered question as a shopkeeper is to have a customer leave without making a purchase. He promoted objectivity in serving patrons to avoid one-sided opinions or confirmation bias. When he presented this paper in 1876, one attendee remarked, 'I wish his paper could be read by every librarian and every library director in the country.' A librarian should be more than just a keeper of books; he or she should be an educator.
00:07:49.500
His work and influence earned him the nomination to the presidency of the American Library Association. Upon accepting the position, he said, 'The function of the library is to serve its users.' He is considered the father of reference library work. In 1883, it was thought by many in public libraries that to provide service directly to the public would take too much time away from the library's administration. This isn't entirely unfounded. It's a common joke among librarians that it is ironic that they work with books because they are so busy they never have time to read.
00:08:21.120
However, in 1883, Boston Public Library hired the first reference librarian explicitly to serve the information needs of Boston's citizens. While I cannot find evidence of this, I look at a picture of the main hall of the Boston Public Library on Boylston Street, with a single blurry desk in front of those arched windows, and I believe that the reference librarian sat there. They were the first person patrons saw when they ascended that main staircase, reflecting the esteem of that individual's position in the community.
00:09:05.000
Roughly a year after Green assumed the head of the American Library Association, a world away in India, a man named S. R. Ranganathan was born in Tamil Nadu. Early in his educational life, he set out to be a mathematician; he earned both a B.A. and an M.A. in mathematics from Madras Christian College. He wanted to go into education but was hindered in this pursuit by a speech impediment. In 1923, he somewhat reluctantly accepted the role of University Librarian at the University of Madras. He claims he had forgotten he even applied until he was contacted for an interview.
00:09:55.500
He was known to be a workaholic and immediately immersed himself in studying classification systems. The library he took over was painfully disorganized and chaotic, providing him ample opportunity to work extensively. In 1931, eight years into his career as a university librarian, he published what is considered one of the most significant works of the library profession: the "Five Laws of Library Science."
00:10:35.280
The five laws state: books are for use. Ranganathan observed an emphasis on the preservation and storage of books during his tenure and asserted that these activities should serve to promote the use of the material, rather than keeping it away from the public. Libraries are not meant to be shut away from their users. He further stated that every reader has his or her book, emphasizing that the function of a librarian is to serve the community and maintain a collection that suits a variety of needs without judgment.
00:11:16.920
Ranganathan believed that every book has a place in the library, and even if a small number of people choose to read them, they still belong. He asserted that all library users should be able to locate their desired material effortlessly and efficiently. Lastly, he believed that a library is a growing organism—it's a dynamic institution that should never be obstinate in its outlook. All material must be updated over time in terms of content and medium.
00:12:10.260
To this point, we have covered foundational concepts of both libraries and reference librarianship. However, these concepts were developed at a time when the brokering of information was becoming more democratized. With the popularization of ideas put forth by Ranganathan and Greene, information becomes something that is present for all, though there are still socioeconomic barriers preventing access. Many previous barriers have broken down, and as this process accelerates, the study of library science begins to shift more toward the study of information science.
00:12:47.640
The first significant work that shifts the focus of information behavior from the organization of information to its use was published in 1981 by Thomas Wilson. He was born in 1935 in Shincliffe Station, County Durham, England. After dropping out of school at age 16, he worked as a library assistant for the Durham County Library. He completed his national service in the Royal Air Force and later qualified as a public librarian. Wilson became interested in the use of new technology and information science while working as a corporate librarian for a nuclear research center.
00:13:45.120
By 1961, Wilson had begun his academic career at the University of Sheffield. In 1981, he presented his model of information-seeking behavior at a conference at the University of Maryland. The aim of this model was to outline the various areas covered by what Wilson proposed as information-seeking behavior, as an alternative to the common concept of information needs. I'll save more granular detail for the next diagram, but here's an overview of the path an individual might take in information-seeking behavior.
00:14:27.600
This process involves the emergence of an informational need that progresses to making demands on an information system and leveraging other individuals for the fulfillment of that need, alongside the transfer of accompanying information. Like many in the industry, Wilson became prone to overwork. He became so engrossed in other responsibilities and research that he left his model neglected for 18 years before returning in 1999, somewhat dismayed that no one else had furthered his study.
00:15:21.300
Undeterred, he published an updated model in 1999, this time incorporating research on information behavior from disciplines such as psychology, innovation, health communication, and consumer research. The individual or information seeker remained at the focus of the information need, not the need itself. Wilson asserted that the root of information-seeking behavior lies in the concept of information need, which he proposed is a subjective experience identifiable only by clues in an individual's behavior or through direct admission.
00:16:09.600
For instance, I think of myself wandering the grocery store with a puzzled look on my face, trying to find a product I've never purchased before, hoping a friendly store clerk will ask if they can help me until I inevitably break down and just ask where the dried mushrooms are. The need for information comes from an activating mechanism, which itself is linked to stress and coping theory in health information research.
00:16:47.560
It suggests that stress is often a product of whether the individual has maximal information or minimal information about circumstances or events and their effects. In other words, stress arises from a lack of contextual knowledge or information. Information seeking can often act as a coping mechanism to alleviate that stress. However, this stress isn't easily resolved, and there are intervening variables that can obstruct an information search before it can remedy that initial stress.
00:17:45.060
Psychological barriers represent cognitive dissonance between initial information and the context of the searcher, playing into concepts like confirmation bias. This concept borrows from consumer research theory, demonstrating that if an advertisement aligns with an individual's beliefs, they are more likely to support the arguments for that product. Conversely, if presented information contradicts their previous beliefs, they are likely to dismiss or refute the ad.
00:18:34.060
We see similar behavior in relation to new information. Demographic variables include ideas like economics; there is a direct economic cost associated with pursuing information, and time is not limitless. The effort required to pursue an informational need is rarely unbounded by time. Interpersonal variables arise when new information software requires personal interaction for access. This can lead to problems with information gatekeepers that can obstruct the free flow of information.
00:19:19.740
Environmental variables are perhaps the shortest variable to cover, but it focuses on access. Generally, access has decreased as information has proliferated; however, the issue of consistent and quality access to informational sources can be highly variable depending on the population considered. Lastly, there are source characteristics, which overlap with environmental variables, concerning the credibility and reliability of the sources available in one's environment.
00:20:09.039
The final stage of initiating an information search occurs once an individual has recognized a need and has advanced past any intervening variables. At this stage, it could be said that the information need exists enough to be considered as part of a coping mechanism. Yet, an individual may still hesitate, weighing the risk-reward ratio of this journey.
00:20:39.960
Does the outcome justify the time that may be spent? This is similar to my own thought process when I dread going running. Specifically, it may be asked whether the uncertainty of the result outweighs the cost—time or otherwise—of the search. Lastly, self-efficacy, under social learning theory, affects how long someone persists in an action and the effort invested. Positive initial experience may encourage further inquiry.
00:21:25.920
Wilson leveraged a quote from Canadian psychologist Albert Bandura in his landmark book, "Social Foundations of Thought and Action: A Social Cognitive Theory." Bandura wrote that 'the strength of people's convictions in their own effectiveness is likely to affect whether they will even try to cope with given situations.' I believe this speaks to confidence in one's abilities as an activating mechanism for believing one can figure something out.
00:22:06.840
The less confidence we have, the less likely we are to engage in information-seeking behavior to gain more knowledge. So, what does all of this mean to us at a software engineering conference? How is this relevant to anything? You might say, 'This isn’t going to make my specs faster or my code more performant.' You would be right, and although that's not a question, I'd like to return to the world of reference librarianship for a moment and reconsider its function.
00:22:45.780
The function of reference librarians is to assist in marshaling information seekers through their information behavior. This service looked significantly different in 1981, when Wilson's first model was published, compared to what it looked like in 1999 with his revisions, and now in 2021, all looking wildly different from what it was for that first reference librarian who sat at that amazing desk on 55 Boylston Street back in Boston.
00:23:28.860
In today's world, that desk is gone from that hall, and the reference librarian's desk looks more like this. With these changes, there have been some surprising statistics. Reference services experienced a broad decrease in use from roughly 1999 through 2004 as information seekers gained access to online materials. However, this experienced a resurgence beginning in 2007 amidst the Great Recession; it became the function of reference librarians to provide access to educational resources and job search information.
00:24:11.760
Moreover, the discipline of reference services has uneven importance and rates of use depending on the environment of the community. Areas serving younger and more technologically savvy populations have seen decreases, while other institutions—such as public law libraries, health service libraries in academia, and public libraries serving older or lower-income populations—have seen significant upticks.
00:24:57.300
William Katz's book, "Introduction to Reference Work," is a textbook you are generally required to purchase during your first semester at any Master's in Library Science program. In this work, he outlines four basic types of reference questions, which I believe have parallels with questions we regularly answer within our engineering teams. Let’s take a look at these four types of questions.
00:25:35.220
First, we have directional questions. In libraries, these are questions like 'How do I reserve a room?' or 'Where are the bathrooms?' Most of these can be resolved once identified; for example, a sign can indicate where the bathrooms are located. For engineering teams, we have many tools at our disposal to fulfill this need and even automate it.
00:26:09.300
Well-maintained READMEs and repositories can do as much heavy lifting as a sign indicating where the bathroom is. They reduce any environmental barriers to information-seeking behavior. If you have multiple services, you can create a lexicon for them and store it in a space like Confluence or any number of free wiki products. These should be kept current to prevent any source distrust.
00:27:02.160
You can leverage a Git repository to maintain this information. A significant advantage of using a repository is that changes can require a pull request that necessitates review, reducing psychological and demographic barriers to information-seeking behavior. As time and effort invested in information journeys decrease, if you need to track down where something was introduced, such as a bug or a new feature, a clean commit history is invaluable.
00:27:49.740
If you're looking for the introduction of a feature or the conversation surrounding it, finding that information in a single commit with a well-written commit message is crucial. Again, this reinforces the authoritativeness of accessible sources of information. While Slack is excellent for many purposes, it may not be ideal for this, as conversations there are often time-bound, and correct answers can change over time.
00:28:23.880
It's not uncommon to search a company’s Slack history for an issue and find multiple answers over time that contradict each other. If decisions are reached in Slack, they should be summarized in an external document to ensure a single source of truth. This practice helps reduce source distrust and environmental barriers to information-seeking behavior.
00:29:12.900
If you have foundational documents of information as directional questions inevitably arise, you can respond in Slack with a link to a relevant artifact, whose content may change, but whose location remains constant, serving as an authoritative source of truth. Next are ready reference questions, which can generally be answered in about five minutes.
00:29:56.880
In reference, these questions tend to be pretty shallow and can involve things we take for granted. However, they may not be as easily accessible for everyone. I observed this while interning at the New York State Supreme Court Law Library. As a reference librarian, you occasionally serve in the public access library, which assists individuals without means to afford legal defense to conduct their own research. The reference librarian working there was candid; finding information on the court's website was notably difficult.
00:30:36.900
Yet she could answer many questions offhand in under five minutes. In programming, these topics often involve syntax questions, questions about logging into services, or even a demo user for your supported platform. These are typically small bugs or test failures that arise unexpectedly. Such questions can be streamlined and even automated through the adoption and maintenance of a style guide that is easily accessible. When questions involving syntax arise, the style guide should provide answers.
00:31:22.920
Further, if there are controversial points in a pull request history, they should be documented in line with the decision that led to the change. Notes should not simply state: "This is better" or "The alternative is dumb"; instead, they should inspire confidence in the source and its authority. One should automate style guide enforcement as part of code review to reduce interpersonal barriers and improve a sense of self-efficacy among coders.
00:32:01.680
It's important to ensure conversations on these topics take place in Slack. While Slack is a fantastic tool, once discussions are concluded, codifying them in a single source of truth is essential. Avoid leaving it to chance that a correct answer will be found in Slack history; rather, ensure all information is current and in its expected location.
00:32:47.760
In-depth references arise when earlier types of questions fall short, requiring research that lasts more than five minutes. For these questions, reference librarians must put in considerable work and provide verified and authoritative resources, avoiding non-credible sources. When I worked in the University archives at St. John's, one of my favorite inquiries involved gathering biographical information on the 1984-85 Final Four team and individual statistics of all players.
00:33:31.920
Wikipedia captured most of this information, down to the high schools the players attended prior to St. John's. However, it didn't contain game-by-game statistics, so I spent a day sifting through student newspapers to extract that data. It sounds tedious, but I genuinely enjoyed the research process!
00:34:14.880
In software development, this mirrors the need for digging a little deeper, particularly with dense logic bugs, unexpected test failures, alerts, or alarms that can span several files or services. These issues often take more than a single day to address, but resolution transparency helps in resolving such matters.
00:34:47.597
Some solutions to this can include maintaining up-to-date incident reports that document production or public-facing issues, detailing what the issue was, what was done in the investigation, and how it was resolved. The outcomes of root cause analyses for longer-term actions can include creating searchable error logs or documentation in a knowledge repository to facilitate future retrieval.
00:35:38.840
Keeping such documents current is essential, as lack of trust in documentation about past alerts and incidents can lead to repeated discussions in RCA (Root Cause Analysis) sessions. Thus, it is paramount to consider why documentation might be out-of-date and to keep that information current.
00:36:27.210
Another approach to reduce barriers to information behavior could involve establishing an internal Stack Overflow. Often, we rely on this platform for solutions, but these can sometimes extend beyond ready reference contexts. Moreover, existing solutions from within the team might not be immediately findable. I know Stack Overflow offers a team plan that may be useful.
00:37:06.480
You can use numerous other tools that possess reasonable search capabilities to spearhead this initiative, though remember the risks of moving information over time may complicate retrieval. Lastly, consider implementing a reference development rotation for more tenured engineers. We often expend a lot of energy worrying about our most tenured developers leaving and taking their knowledge with them.
00:37:53.640
Yet we only address this knowledge retention on request or with interruptions to their calendars. Implementing regular on-call shifts could provide opportunities for these veterans to share their knowledge. This shared experience can enhance understanding and address deeper application questions that inevitably grow as companies age.
00:38:32.520
In the same vein, create a training rubric for these developers involved in rotation. This should be an approachable time that avoids interpersonal source characteristics and serves as an invitation for information behavior. It's hard enough to ask for help in our industry, so strive to make the process as welcoming as possible.
00:39:20.760
Lastly, consultations and demonstrations can take place. These could involve using services like WorldCat, which effectively replaces the card catalog for scholarly article databases. Many libraries will assist you with project research or conduct citation analysis—though you typically must schedule time for such aid.
00:40:03.900
In development, these opportunities may be less research-focused, instead centering on discovery. If you have a larger organization, cross-team working groups can inspire thought and knowledge transfer regarding new approaches to challenges, new tools, or effective strategies used by other groups.
00:40:50.040
Create a space that welcomes testing for new approaches to old problems or trying different ideas. There is, of course, a learning curve to this approach, so aim to prevent risk and reward biases from obstructing knowledge acquisition. Furthermore, the reference rotation system offers the opportunity for discussing ideas when you’ve been pondering why something is the way it is—without the apprehension of interrupting someone's daily workflow.
00:41:34.020
In conclusion, I propose a modified version of the Five Laws of Library Science relevant for engineering teams. Number one: all information is for use. Fundamentally, we are brokers of data to our end users or teammates. We prioritize putting that data into context as information for our users, trusting them to derive knowledge from it.
00:42:09.840
We have a responsibility to ensure that same service and capability for our teammates with the data we generate, the information we document, and the knowledge we possess. Secondly, every team member needs unobstructed and easy access to all knowledge. Earlier, we outlined the numerous barriers to information-seeking behavior that may exist; we should strive to reduce those barriers as much as possible.
00:42:49.440
Number three: all information is potential knowledge and should be recorded as such. Regardless of how trivial something may seem at the moment, there is always a chance it can assist someone in some capacity. Never miss an opportunity for that to occur. Number four: break down and maintain the elimination of barriers to information-seeking behavior. As previously discussed, there are numerous ways information journeys can be obstructed before they even begin.
00:43:34.560
If blockers to information-seeking behavior are present, identify and remedy them. Each time a journey is cut short, it's akin to a customer leaving a store without making a purchase. Finally, recognize that your engineering team's knowledge is ever-expanding. Everyone is learning constantly, not just in development, but also cross-functionally and through extracurricular activities.
00:44:12.960
Create space for this learning and foster it. If self-efficacy concerns hinder information-seeking journeys, encouragement and excitement about the knowledge others possess can catalyze further growth. Thank you once again to everyone that made this happen remotely—conference organizers, fellow speakers, attendees, and all. I encourage everyone to support their local library.
00:44:53.880
If you're in the United States, there's a link available to easily locate your closest branch. Normally, this is where I would take questions. However, I'm speaking to you from the past, and that's not possible. That said, there is a Discord session scheduled for Tuesday, April 13th, from 4:10 to 4:40 PM, Eastern Time. I might be eating lunch at that time—it will be my lunch hour.
00:45:37.380
And again, I encourage you to get your vaccine when it's your turn and you're able. Please feel free to reach out to me with questions on Twitter; don’t be misled by my lack of activity—I will respond! I have a few citations here for the sources used. Additionally, I have some plugs for the gems I referenced during this talk. Thank you very much again; I appreciate it.