The next layer: Towards open pedagogy in geospatial education

Open‐source software and open data are becoming increasingly popular in the teaching and learning of geographic information science. The cost savings that are derived from using free software over proprietary software are one driving factor, yet the move from “closed” to “open” represents much more than financial austerity—it signifies a broader shift in educational philosophy. This article documents the gradual transition of an introductory undergraduate course in geographic information systems from an entirely closed course to one that has become increasingly open. Having completely adopted the first three layers of open—software, data, and educational resources—the course is now turning toward the next layer: embracing the philosophy of open pedagogy.

learning geospatial science. Building on this foundation, we are beginning to see the next "layer" of openness: open pedagogy (OP). Open-source tools, open data, and OERs can certainly enhance the traditional lab assignment, research paper, syllabus, and weekly course structure, but they also allow us to rethink what those traditional syllabi and assignments could and should look like. How can this burgeoning toolkit of openness give us an opportunity to teach and learn in important new ways?
This article documents the steady transition of an introductory undergraduate course in geographic information systems from an entirely "closed" course to one that has become increasingly open. Having completely adopted the first three layers of open-software, data, and educational resources-the course is now turning toward embracing the philosophy of OP. As such, the course structure, assignments, and learning outcomes are all being reexamined as the benefits and challenges of moving away from a surveillance model of education to a more participatory and collaborative one are investigated.
The main objective of this article is to build upon the large and growing literature on open GIS (Coetzee et al., 2020;Osaci-Costache et al., 2017;Sui, 2014) by integrating the emerging research on the value of OP (Hegarty, 2015;Tietjen & Asino, 2021). As the benefits of transitioning to "all things open" (https://www.allth ings o pen.org/; McKiernan et al., 2016) are made increasingly clear, incorporating OP into GIS education becomes imperative. This article documents one such transition in an undergraduate course on GIS, arguing that OP is the logical next step for enhancing creativity and innovation in GIS education.

| ANATOMY OF A "CLOS ED" COUR S E
I began teaching an introductory GIS course at a small liberal arts college in 2003, modeling the course after ones I had taken in my graduate studies: obtain proprietary GIS software, assign a textbook that included a CD of tutorial data, create lab assignments with the tutorial data, write quizzes and a midterm exam, and assign a final mapping project. Since my institution is one of a handful of work colleges in the country, meaning that all students worked from 8 to 20 h a week on campus, I was also able to establish a student GIS work crew. The crew was responsible for keeping the GIS lab open, reporting any problems with hardware and software, and serving as informal teaching assistants for the students needing additional help on assignments.
Most of the students in the class dutifully completed their assignments, and some became enthusiastic enough about GIS to want to apply it to other projects. But after a couple of years, I noticed that the students who were the most excited about GIS, and the ones who were actually getting jobs in GIS after graduation, were the students on the work crew. A couple of the student crew members had not even taken the introductory course, yet they still displayed a desire to learn and saw the potential for applying geographic information science in their own academic work.
The lab itself consisted of 16 desktop computers, an instructor computer, and a shared storage drive. These machines themselves were "closed," as students could not alter the operating system or any of the installed software-they could only work on projects that were then saved to the shared drive. This setup was helpful in keeping the lab in good working order but was certainly detrimental to student learning and frustrating for the student crew. In 2005, we received a small grant that allowed us to install a separate workstation for crew use.
This workstation was set up as a dual-boot computer so that students could have access to both Windows and Linux operating systems, and they could tinker to their hearts' content. The proprietary GIS software we used in class could not be installed on Linux, but before long one of the crew members had begun installing open-source software to see if he could perform some basic GIS operations. We even began to explore the idea of "GIS on a stick," since it was becoming possible to install an entire operating system and necessary software on a small USB drive (e.g., see https://live.osgeo.org).
Seeing the enthusiasm with which my student crew went about tinkering with open source, and noting that there was a growing community of open-source geospatial software developers, I began to consider the possibility of teaching an introductory GIS course with open-source software. Little did I know that what I considered to be merely a shift from costly to free software would lead to a reconsideration of my entire pedagogical approach to GIS education.

| TR AN S ITI ONING TO OPEN -SOURCE SOF T WARE
In 2002, Gary Sherman began developing Quantum GIS, a free and open-source GIS software platform that was officially released as version 1.0 in 2009. Licensed under the GNU General Public License version 2 (https://gnu.org/licen ses/old-licen ses/gpl-2.0.en.html) and eventually renamed as simply QGIS (https://qgis. org/en/docs/index.html), the software began to attract users due to its user-friendly interface and its extensibility. Users were free to use, modify, share, and study the software code, and over the next few years, the software's features expanded considerably. With the release of version 2.0 in 2013, the popularity of QGIS began to grow, and over the next few years, it was available for multiple operating systems. Importantly, the software also began to include software plugins and interfaces with other geospatial toolkits such as the Geographic Resources Analysis Support System (GRASS, https://grass.osgeo.org) and the System for Automated Geoscientific Analyses (SAGA, https://saga-gis.sourc eforge.io/en/index.html), both of which are also open source and freely available.
It was the integration of an interface with GRASS tools that made QGIS a viable alternative to the proprietary software being used in my undergraduate GIS course at the time. The course is primarily populated with students in the natural sciences, and, therefore, raster analysis is quite common; having access to a robust set of raster tools in GRASS made it possible to replicate my existing lab assignments that previously required access to an expensive software extension. As such, in 2015 I made the decision to test the waters by offering a course entirely focused on free and open-source software.
At the time, the GIS software in use at my institution was limited to 25 "seat licenses," meaning that no more than 25 copies of the software could be running simultaneously on campus. Furthermore, most of these licenses were to be restricted to one computing laboratory. This meant that students could only work on projects when the lab was open and available, which did not include nights and weekends. Limiting access to desktop computers in a single lab significantly inhibited the amount of work that could be assigned to students, and, therefore, limited students' opportunities to learn. Transitioning to open-source software not only eliminated the expense of purchasing software and paying annual maintenance fees, it provided the flexibility and freedoms that Richard Stallman insists are a "moral duty" of educational institutions (Stallman, n.d.). The benefits of the move to QGIS at my institution can be summarized by what I refer to as the "five E's," which are briefly described below.

| Expense
This is considered as the "low-hanging fruit" in terms of justifying a transition to open-source GIS software, as institutions are always looking for ways to reduce expenses without unduly harming their underlying educational mission. Given that open-source software had matured enough by 2015 that the introductory course learning objectives could be met just as easily as with proprietary software, the reduction in annual expenses was met with enthusiasm by the administration.

| Ethos
Students apply to my institution because they are excited to become collaborative members of a work college community. They understand that in addition to academics, they are expected to work each week as part of a work crew and also participate in community engagement activities. In addition, the campus has a working farm and garden, a blacksmith shop, a fine woodworking shop, and a 650-acre working forest. Most of the students consider themselves "makers" of sorts, preferring to create and produce resources rather than simply rely on store-bought consumer products. Open-source software meshes well with this ethos, as students prefer working with a set of tools developed by a community rather than produced by members of a corporate hierarchy. Many of our students are studying the natural sciences, where the scientific method is the norm; they understand, then, the argument that reproducible science requires that you be able to look at the software code being used in scientific research (Ince et al., 2012).

| Extensible
Students not only appreciate being able to "look under the hood" of the software they use, but they also recognize the power of extensibility. If the software being used does not have a tool readily available to take on a particular task, there is a good chance that some sort of extension or plugin exists that can be downloaded and added to the program. In the event that the plugin does not exist, a student with the desire to create it can learn the skills necessary to do so. There are currently more than 150 plugins available for download on the QGIS repository, and more features in the core software are being added with each new version release.
In addition, not being wed to one particular suite of expensive software means that students can search for solutions across multiple software tools. Many of our students often turn to the R statistical programming language, for example, which at the time of this writing has almost 20,000 available packages that can be downloaded and added to the core program.

| Expandable
Not only does open-source geospatial software offer opportunities for extending its functionality through additional plugins and packages, but it also makes possible the expansion of the spaces and time in which the software can be used. No longer confined to the limited number of machines in the limited capacity of a single laboratory, open-source software allows for a much wider distribution of the necessary tools across-and beyond-campus. Once our course shifted to open-source GIS, we were able to add the software to computers in the library and students were able to download and install it on their own computers. This had a much greater impact than we originally thought-suddenly we found that students were working on projects in the library, in the cafeteria, and late at night in their dorm rooms. This increased accessibility made it possible to assign more complex projects that required time outside of class and lab hours, and the students readily embraced this work. While there was also some added complexity as students grappled with installing and troubleshooting software on their own operating systems, this seemed like an important educational component as well. Students were not walking into a lab that had been set up for a generic learning environment; rather, they were learning to get the tools they needed up and running on their own computers. And of course, when the class abruptly shifted online amid the COVID-19 pandemic, having access to the software on individual student computers became essential.

| Employment
Many students end up enrolling in a GIS class because they view it as providing them with an employable skill set.
Many of the organizations listing entry-level positions use a popular proprietary suite of software. Yet learning how to work with open-source tools should not be viewed as an inferior alternative to learning the proprietary software-instead, it can be argued that learning to apply open-source software to tackle geographic information science provides a more robust introduction to the field. Indeed, one of our students was hired by the City of Asheville after he graduated precisely because of his experience with open-source tools. The manager of business and public technology with the city at that time made it known that his office preferred hiring people with opensource software experience, as they often seemed to have a better knack for solving problems. Similarly, several of the students who learned the fundamentals of geographic information science using open-source software in class went on to participate in the NASA Develop program, which frequently makes use of proprietary GIS software. These students were able to successfully complete their work, as the concepts and skills they learned were applicable across different GIS platforms.
Taken together, these "five E's" demonstrate the value in emphasizing the use of free and open-source software in education. All the concepts and skills typically covered in an introductory GIS class-visualizing vector and raster data, geoprocessing, SQL, map algebra, composing map layouts, batch processing and more-can be just

| OPEN AND AFFORDAB LE EDUC ATIONAL RE SOURCE S
The many benefits of open-source software outlined above can also be found in the adoption of OERs. While there is no shortage of GIS textbooks on the market, these tend to be expensive and often bundled with (unnecessary, as we will see below) tutorial data. Furthermore, these textbooks can be viewed by students as too much like a cookbook for a specific software program (point-and-click guides to GIS operations), or as too theoretical and not a practical resource for hands-on learning. Therefore, switching from costly, copyrighted textbooks to OER is a logical next step after adopting open-source software for GIS education. software. This means that, unlike with the standard published textbook, OER users are free to make copies, adjust the resources for their needs, and combine the content with other resources. Instead of the "five E's" described above, OER can be explained through the "five R's": retain, reuse, revise, remix, and redistribute (Wiley & Hilton III, 2018).
Similar to the transition from proprietary software to open source, the most immediate benefit of a transition from traditional textbooks to OER is financial. Several studies have found that textbook price increases have steadily outpaced the rate of inflation for decades (Pollitz & Christie, 2006; see also Everard & St Pierre, 2014). In rapidly changing fields like geographic information science, used textbooks are typically not a viable option and often have little resale value. As a result, students often avoid purchasing an assigned textbook or have to work extra hours to be able to afford it (Hanson, 2021). The financial benefit to OER can be seen as a social justice issue, as it helps create more equitable access to education.
When my introductory GIS course focused primarily on proprietary software, I tried different textbooks each year as I struggled to find one that I thought was the most useful without being exorbitantly expensive.
Nevertheless, students complained about the cost and several avoided purchasing it. Those who did purchase it often had the complaints mentioned above: one book felt like a recipe book that led to students mindlessly following along with the exercises, while another felt too theoretical and did not offer much in the way of practical exercises to help students learn. Ultimately, I was not satisfied with any of the textbooks and decided to make the switch to OER.
Most college or university libraries will have at least one librarian who is well versed in OER, and there are many online resources available for accessing quality resources. The MERLOT (https://merlot.org) repository contains thousands of curated resources, as does the Openly Available Sources Integrated Search (OASIS, https://oasis. genes eo.edu). Openstax (https://opens tax.org) and the Open Textbook Library (https://open.umn.edu/opent extbooks) are two additional popular repositories for textbooks, while OER Commons (https://oerco mmons.org) includes access to other materials such as labs, datasets, and case studies. In addition to these resources, many individual schools, school systems, and states are forming online resources and networking opportunities for faculty to learn more about OER. In my state, for example, NC Live (https://nclive.org) has formed Open Education North Carolina, and the Appalachian College Association started Open Appalachia (https://www.acaweb.org/colla borat ions/oar) to promote OER adoption across its member schools.
For my introductory GIS course, I opted to use a GIS textbook from the Open Textbook Library (Campbell & Shin, 2011). While not containing practical lab assignments, I felt the text did a good job of covering many of the necessary concepts needed in the early weeks of the course. Because I was not forcing students to pay an exorbitant price for the book, I did not feel required to teach every chapter. Instead, I could pick and choose the chapters that worked best for my particular course and could assign them in an order I saw fit. I designed the lab assignments for the course, but there are plenty of tutorials online that could also be incorporated into a course to supplement labs and exercises (Gandhi, n.d., see also https://qgis.org).
Another benefit of using OER is that you are not limited to a single textbook. If another resource has a single chapter or a handful of chapters that suit the needs of a particular course, it can easily be assigned in addition to the primary text. A chapter from one book can be assigned for 1 week, chapters from another for the next week, and creative commons licensed articles and YouTube videos for the next week. This is one example of the benefit of "remixing" from the "five Rs" described earlier, and demonstrates the benefits of OER for allowing for a more diverse collection of voices in the course materials.

| OPEN DATA
One concern that GIS educators may have when switching to OERs is that an OER text may not have an accompanying website, lab materials, and/or tutorial data. This may be of particular concern for instructors and students in an introductory GIS course, where basic, cleaned datasets are valuable for use in the early weeks as students first begin to learn the fundamentals. For those who have been teaching and working in GIS for decades, the memories of how hard (and often expensive) geospatial data was to create and obtain are quite vivid.
In the last decade or so, however, this has begun to change. Data, in parallel with open-source software and OER, have increasingly become open and accessible. Data locked behind paywalls and buried in PDFs can now be easily obtained through data portals and dashboards. The petabytes of data being collected by NASA's Earth Observing System's fleet of satellites are available to anyone without restriction.
As someone who has taught GIS for 20 years, I have witnessed this transition from closed to increasingly open data-and have seen how it has transformed teaching and research. For example, around 2005 I had a student hired to work on a funded grant project in which the goal was to create a single cadastral map for approximately 40 counties in Western North Carolina. The student contacted individual county governments and requested polygon vector data files for property boundaries if available. The responses varied tremendously. While some counties were more than happy to share, this sharing often took the form of compact discs or another storage medium, which the student then had to travel to retrieve or pay to be shipped. Other counties were more reticent, offering the data for a substantial fee or not offering it at all. One county only had centroid data available. In all, the project took the student several months and a not insignificant portion of the grant budget to complete.

A few years later, the US Environmental Protection Agency awarded a grant to the North Carolina Department
of Environment and Natural Resources to begin building the "Integrated Cadastral Data Exchange" (https://www. ncone map.gov/pages/ parcels). At this same time, a statewide data portal for North Carolina, NC OneMap (https:// ncone map.gov), was created to serve as a statewide geospatial data repository. Among many other datasets, NC OneMap provides access to a seamless parcel layer for all of North Carolina (https://www.ncone map.giv/pages/ parcels). In short, what took a student several months in the early 2000s now takes a student mere minutes. This is but one example of the benefits of open data. It underscores the rationale of organizations like Code for America, a nonprofit organization whose vision is that "government can work for the people, by the people, in the digital age" (https://codef orame rica.org/about -us/visio n-and-value s/). As governments and nonprofits increasingly began to see the value in open data, portals and repositories began springing up in dozens of cities, states, and at the federal level. My students, for example, now have access to data at the municipal and county level (via the Asheville Open Data Portal, https://data-avl.opend ata.arcgis.com/), state level (http://ncone map. gov) and federal level (https://data.gov). The open data movement has unquestionably ushered in an unprecedented amount of available data to anyone-including students in their first GIS class.
Another benefit of open data in GIS education, in addition to it being freely available and untethered from an expensive textbook, is that it gives students useful skills in identifying and downloading useful datasets. In my introductory GIS course, I provide students with a handful of basic vector layers to get them started, but by the second week of the course, I have students searching open data portals for their own spatial data. This typically leads to an interesting discussion as students run into various hurdles (e.g., "I only wanted the streets for my hometown but had to download the whole county," or "I found two layers for my state but they don't overlap") that can then be used as lessons later in class (in these two cases, how to filter data to a desired geographic extent, and how to work with data in different projections). It can even be a gentle (or sometimes not so gentle) introduction to data cleaning, depending on the datasets uncovered by the students.
Some may feel that having students download geospatial data instead of relying on ready-made textbook data so early on in a GIS course might lead to frustration and inhibit students' desire to learn. I would argue the opposite: that showing students the diversity of data out there sparks their imaginations and gets them doing the important work of asking spatial questions. It also prepares students to be problem solvers rather than recipe followers. Having students face common issues such as differing projections or the need for data cleaning during the second week of an introductory course sets them up to anticipate problems and to consider the possible solutions.
It also sets them up to want to experiment, even if they fail at first, rather than to just simply follow directions from a textbook. This is an important first step in having students understand that they are not expected to simply "consume" knowledge but also to co-create it and to contribute to the learning of others. It is, in other words, a gateway to applying the principles of OP to GIS education.

| THE NE X T L AYER : TOWARD OPEN PEDAG OGY
Using the common metaphor of "layers" used in GIS to describe multiple datasets in the same geographic extent, we can consider the approaches to openness described above (open-source software, OERs, and open data) to be layers that build on one another in GIS education. Open-source software provides the "basemap" for open GIS education, as it democratizes access to geospatial software and makes it freely and easily available to anyone who wishes to use it. OERs make up the next layer, as they similarly provide anyone access to materials that may have previously only been available behind an expensive paywall or through the purchase of a copyrighted textbook. The third layer, open data, opens up the possibilities for asking spatial questions by making an incredible diversity of data freely and easily available for anyone with internet access. These layers most certainly overlap, as questions about cost, ownership, equality, accessibility, and knowledge acquisition are embedded in each. Taken together, these layers build a solid foundation for adding the next layer: OP.
What exactly "OP" means is still being debated, and no set definition has been agreed upon. There are, however, some common elements across multiple definitions, as described by Tietjen and Asino (2021) in their research seeking commonalities across these multiple definitions. They developed a "five-circle framework" to identify the key components of OP, including (1) OP welcomes diverse learners as design partners, (2) OP is a participatory pedagogy, (3) open licenses are essential to foster practices such as remixing, (4) OP encourages all learners, inside and outside the school setting, to contribute to building a knowledge community, and (5) OP fosters a culture of collaboration through sharing and editing (Tietjen & Asino, 2021).
The idea of OP builds off of other pedagogical approaches to learning that encourage the movement away from the standard classroom structure of active instructors and passive students. Sometimes referred to as "studentcentered" pedagogy, "experiential learning," "active learning," and other expressions of what we might include within the larger umbrella of "inquiry-based" learning (Biswas-Diener & Jhangiani, 2017;Khalaf & Mohammed Zin, 2018;Pedaste et al., 2015), the concept of OP draws from a substantial literature on the educational value and importance of shifting away from the more "traditional" form of classroom instruction.
At its simplest, OP urges us to depart from what Freire calls the "banking model" of education (Freire, 2000), instead working toward a model in which students are active participants in their own learning. It requires questioning of what David Wiley calls the "disposable assignment," considering instead educational activities that add value to the world (Wiley, n.d.). In GIS education, it beckons the instructor to move beyond lab assignments that encourage pointing and clicking in lieu of thinking, or written assignments that get seen and evaluated only by the individual assigning them. It asks that we move away from "here's how you do it" to "what do we want to do, and why?" Yet, this next layer of open may seem more daunting than the previous three. It is one thing to switch from a DVD of prepackaged data to open data on the internet, but quite another to transform course assignments and projects to more fully embrace OP. What might this look like in an introductory GIS course? In practice, OP might take several forms-from "crowdsourcing" a syllabus to creating a podcast to having student projects focus on a "deliverable" for an outside client. The point is that students need to be viewed as co-creators, not empty receptacles, and that the work and learning being produced is done so for the production of knowledge, not simply a letter grade.
In my introductory GIS course, I have begun to look at ways to incorporate the principles of OP into the teaching and learning environment. While the course is by no means an exemplary model of what good OP looks like, it does demonstrate some of the early steps being taken to create an environment where students are producers of knowledge, not merely consumers. Below are some examples of steps being taken to foster a knowledge community around geographic information science.

| Collaborative assignments
Students are encouraged to work together on daily assignments, which are to be posted to the class site on our learning management system (LMS) by the end of the class session. These posts are then viewable by all participants in the class, not just the instructor. The knowledge that one's peers, and not just the instructor, will be able to see the work provides an additional incentive to do well, and the fact that students are able to work collaboratively ensures that they see this exercise as more of a learning process rather than simply an assignment for a grade.
Our institution uses the open-source Moodle LMS for course management, which includes tools and resources for sharing work that is viewable by everyone in class. An online forum, where students can easily post their map layouts and comment on the work of others, becomes our shared repository for daily work. These in-class assignments focus on one or two specific GIS concepts, such as SQL, table joins, map algebra, or geoprocessing, and typically require students to submit a map layout that demonstrates the successful application of the assigned concepts. Here again, students can get immediate feedback on their work by examining the work of others and/ or soliciting feedback from their peers. These concepts and skills are then reinforced with longer lab assignments, which are turned in for instructor feedback and assessment.

| Critical GIS paper
Early on in an introductory GIS course, students often express amazement at the power of database querying or geoprocessing tools. It seems important, therefore, to have them simultaneously think critically about GIS as a technoscientific human construct. GIS is more than just software and hardware-it is also a science, a set of social practices, and a complex collection of unequal power relationships. Its history is embedded in a larger history of military strategy, surveillance, colonialism, and territoriality. In the course, I ask students to read about GIS as a set of social practices, then identify a specific artifact (map, book, news article) to critically examine. We then use these analyses as the basis for class discussion on the potential benefits and problems of geographic information science.

| Tutorial modules
There are typically several tools and techniques in GIS that get used repeatedly, yet until students have had the benefit of learning by repetition the application of these tools can be difficult to recall. Students often ask for a "refresher" on how to use a specific tool if they have not had a chance to use it in the past week or two. Given this, I asked students to produce a tutorial module as one of their written assignments. Each student took a particular tool (e.g., creating a buffer) and created a document that addressed the following: (1) what is the end result that this tool will achieve, and why would I use it?, (2) define the tool, using appropriate vocabulary, (3) describe a specific example of the tool in use, including screenshots, and (4) explain the desired outcome after applying the tool.
The collection of tutorial modules was then compiled for use in future GIS classes. Knowing that they were making contributions to a collection of tutorials that would be used by students in the future, the class worked hard to develop clear and thorough documentation.

| Individual student research project
The introductory course culminates with each student conducting a research project. The students have complete control over the research question being asked in the assignment; they are only given some basic parameters and shown some past examples of good student work. Students are strongly encouraged to identify an outside "client" for their work. This might take the form of a student doing some preliminary geographic exploration and mapping for their senior thesis project, or it might be a project specifically requested by a faculty member in another department, or perhaps even a project for an organization off campus. Since our campus has a working farm and forest, project requests often come to students in the class who are also engaged in work in those environments. The key point here is that they are producing a deliverable for a client, not just turning in an assignment for the course instructor.

| A QUE S TI ON OF SC ALE: OBS ERVATI ON S ON THE CHALLENG E S AHE AD
While the early evidence-student course evaluations, student success in graduate programs and early career positions, my own assessment of student learning-indicates that the transition to a fully open GIS course provides positive outcomes, there remain several challenges to "scaling up" this effort more broadly. There are still some aspects of the class that are difficult to fully shift to open, and the nature of my specific institution may foster a transition from closed to open that is not easily replicable at other institutions. Each of these challenges is briefly addressed below.
For the current course, one area that remains challenging is in web-based geospatial visualization and analysis.
Students learn how to export vector data to Keyhole Markup Language (KML) for viewing on Google My Maps (https://www.google.com/maps/about/ mymaps) and Google Earth (https://earth.google.com), which are free (but not open source) web-based mapping platforms. Students also learn how to edit OpenStreetMap (https:// www.opens treet map.org), which is both free and open source but limited in terms of its utility beyond serving as a basemap. The more enterprising and curious students might explore using other open-source web tools, including R Shiny (https://shiny.posit.co), OpenLayers (https://openl ayers.org), Leaflet (https://leafl etjs.com), and OpenDataKit (https://getodk.org), but the open-source online geospatial world is constantly evolving and can prove daunting to students in an introductory course.
Another ongoing set of challenges-which is perhaps the flip side to the "E" that represents expandability-are the difficulties that arise in a transition from a traditional computer laboratory to a "bring your own device" laboratory. Again, students enjoy having the ability to install free software on their own computers and the freedom that provides for doing work outside of the limited time and space of formal class instruction. Yet it also potentially exposes a sense of inequality between those students who can afford expensive laptops and those who cannot. We can partially address this by making desktop computers available to those students who do not have an appropriate computing device, but this still puts those students at a disadvantage since they have much more limited access to the software. At our institution, we have tried to further address this by making open-source GIS software available on the computers in the library room that is open 24/7, yet there is still a question of equity that we have not fully been able to answer.
It is also clear that some dimensions of my particular educational institution-small class sizes, only one GIS instructor, a work program, and an educational curriculum that fosters a "maker mentality"-have made the transition to a fully open class much easier than it might be at institutions that look quite different. A new instructor inheriting an existing lab and program, or an instructor who rotates teaching GIS with other instructors, might find it difficult to make these same sorts of transitions. Similarly, an instructor in an upper-level geospatial course may use proprietary software and expect that students enrolled in the course be familiar with that same software from earlier introductory courses. In such cases, perhaps an instructor could engage open concepts by adopting additional readings that are OER, creating a lab assignment or teaching module on open-source geospatial software, providing students with examples of open data portals, or transitioning an assignment from one that asks students to follow a recipe to one that fosters inquiry-based learning. While it may be necessary to mainly stick with proprietary GIS in these settings, at least in the short term, simply exposing students to the possibilities of openness might spark their interest and catalyze their own efforts to learn more about open tools, resources, data, and pedagogy.

| SUMMARY AND CON CLUD ING REMARK S
While the assignments mentioned previously can be viewed as baby steps toward a truly inclusive classroom of mutual knowledge production, they have begun to change the classroom culture from "what do I have to do and when is it due" to "how might I harness the power of GIS to ask-and perhaps answer-interesting questions." In practice, OP can often feel uncomfortable, as there is a requisite shift in the power dynamic between instructor and student, as the specter of letter grades is overshadowed by the larger goal of knowledge creation, and as the walled garden of textbooks with packaged data is replaced with OER and open data. Yet Taken together, these four layers of open-open-source software, OERs, open data, and OP-can create dynamic new teaching and learning environments for GIS in higher education. Rather than teaching recipe-driven assignments using expensive textbooks and proprietary software, we should be encouraging students' curiosity and freely giving them the tools and resources they need to satisfy that curiosity. As a former student once shared with me as we reflected on that semester's class: If you want to build a ship, don't drum up people together to collect wood and don't assign them tasks and work, but rather teach them to long for the endless immensity of the sea. 1 This quote gets at one of the key aspirations of OP. If we work with students as co-creators of knowledge, if we give them tools and resources that they can tinker with and remix, if we replace disposable assignments with useful projects, and if we devalue letter grades in favor of deeper reflection and evaluation-we may help foster the intrinsic self-motivation that truly enriches learning.

ACK N OWLED G M ENTS
Thank you to the staff of the Pew Learning Center and Ellison Library at Warren Wilson College for their research support.

CO N FLI C T O F I NTE R E S T S TATE M E NT
The author declares no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
All data and resources used in the production of this manuscript can be made available.

E N D N OTE
1 Typically associated with the author of the Little Prince, Antoine de Saint-Exupéry.