Tag Archives: software

Building for Open

I’m encouraged to see the conversation about “open” broaden to contexts beyond education recently. The general public seems to have finally become interested and concerned about the open web. Never mind that it took the threat of slowing Netflix and YouTube to a crawl, but it is encouraging nonetheless. It is important, too, to keep debating the actual meaning of open. But it is now increasingly important to begin understanding open in different contexts, the forces that shape open, and how open layers, builds, and interacts with itself.

Open is a powerful concept – a game changer with the ability to disrupt the status quo. Some say that Massive Open Online Courses threaten the core business models of universities. Others affirm that Open Educational Resources threaten the major publishers. But rather than seeing open as a threat, I would like to consider the opportunities that open creates.  To paraphrase David Wiley, you never know what you’re going to get when you choose openness – and as history informs us, while change can sometimes be painful, you’re better off to be proactive about it. To be proactive about openness not only necessitates having a pragmatic understanding of what it is and isn’t but also how it has evolved over time.

There are two main contexts within which I see open having matured in educational technology. These are first, open as from the perspective of technologists, and second, open in the eyes of content creators.

Open Software

In the first context, the opening of software has its roots in the open source software movement. It begins with providing access to the raw source code that runs a software product or system. This is akin to having the exact measurements and all the steps necessary to recreate a recipe from scratch. Over time, people realized that while having access to the source code to customize educational software was helpful, having standard, reusable hooks to plug into was a better approach, and so open APIs came into existence. Open APIs provided convenient, well-defined ways to hook into educational technology platforms, but integrations built using these APIs were often piecemeal and product-specific in nature. They were not reusable and were prone to breakage when the software was upgraded to a newer version. To overcome these weaknesses, open standards were created as a way to do things in a very uniform way across multiple software products. Standards can’t do everything APIs can do but enable all platforms to achieve a common denominator of a capability. You can think of them as “write once, run anywhere” – which is in practice often the case. While there are still some glitches, open standards have gotten us to a much better place than before. Standards also began to create efficiencies and break down barriers between products. If you wanted to plug once piece of software into many other software products, all you needed to do was write one, standards-compliant integration rather than an individual one for each other system. Because many products support standards, the same software could be plugged into many others enabling collaboration among users of systems in ways that were previously difficult at best.

Individual products then began supporting whole sets of open standards in addition to APIs. Some of these products were open source, but some were not. And once we reached this level of maturity, software products in my opinion became software platforms. They could connect in well-defined ways on the front-end, back-end, and tail-end with very-little-to-no customization. Often, only configuration of these integrations was necessary. Today, I would consider many popular LMSs that we use to be examples of platforms. But even still, not every LMS supports all standards, and sometimes the specifics of the standards support vary.

Once we can more clearly define complete sets of standards support and clearer expectations of behavior, I would expect open platforms to become part of open architectures. Open architectures will allow us not only to plug things together but also to provide clear structure and guidance as to how the whole puzzle fits together. It is often the case that “the details” make or break the success of a technology, and open architectures will allow us to define and set expectations more reliably while also defining data flows into and out of open systems taking into account privacy and security concerns. There are actually emerging examples of what open architectures might look like including ITANA’s draft Reference Architecture for Teaching and Learning. Similarly, open infrastructures will define the environments within which software is run making them more uniform over time.

In the distance, not nearly as far away in the future as it only recently was, open ecosystems will soon define how many products interact, interconnect, and behave at scale. Open ecosystems will exhibit many software products working together in tandem to create a more powerful whole. Predictions for this ecosystem might look like have already begun to materialize.

open_software
Openness in software over time

Open Content

In tandem with how software has evolved in terms of openness over time, momentum behind open content has been building as well. Open content began with the open web, which in cyclical fashion is actually making a resurgence. Content on the open web is accessible to all, directly linkable, and can be found through search engines. There are no logins or other technical constraints enforced to view it. Some of the earliest major attempts at opening educational content can be categorized as open courseware – which put entire courses worth of material on the open web for everyone to view. Soon, it became apparent that while viewing course materials in an open manner was nice, there had to be better ways to reuse and repurpose the content. Given the constraints of copyright and fair use, open licenses created the legal structure within which to enable this. Once content was put on the open web and combined with these permissive licenses, it then became possible to create reusable groupings of content called open educational resources (or OERS) which took many forms.  Sometimes an OER was a simple web page with a simple module or lesson. Other times an OER was much larger and encompassed a package of materials with graphics, video, and animation. OERs could be used in whole or part, combined with other OERs, or reworked altogether to fit a specific purpose. A framework emerged for defining openness of content in terms of ability to reuse, revise, remix, redistribute, and retain content.

Over time, largely driven by market forces, open textbooks emerged as packaged sets of OERs and wholly-commissioned works to replace the proprietary, copyrighted works of the traditional content publishers. While not completely different from OERs, they represent an attempt to create a complete set of materials and learning content for the scope of an entire course.

But perhaps the most game-changing recent development in open content has been the creation of the open course. And when I say open course, I don’t necessarily mean Massive Open Online Courses (or MOOCs) but all forms of courses that allow broad participation from many learners and faculty at many universities (and even just intellectually-curious individuals). Open courses in their many forms begin to break down the paradigms of the traditional course including timeframe, structure, who can participate, and to what end goal.

In the future, we will likely see whole curricula designed around open content. We see evidence of this through efforts that attempt to clearly define learning objectives, sequences, and outcomes for whole programs of study.

open_content
Openness in content over time

Content itself over time is also evolving to be much more dynamic and responsive to the individual user.

  • Static content is straight “text on a web page”
  • Dynamic content exhibits animations and video but is otherwise just static content in motion
  • Interactive content can be manipulated by the user to display or behave or respond differently based on the nature of the interaction
  • Adaptive content is able to display differently based on rule sets. These rule sets can be pre-defined for all users, or users can be grouped into “market segments” for the content to display /respond differently based on which “bucket” a user happens to fall into. Many systems are beginning to exhibit or support some form of adaptivity.
  • We’re not quite there yet, but personalized content will focus on the individual with not only a very mature and refined understanding of his or her learning styles and prior learning but also taking into consideration longer-term learning and career goals.
content_characteristics
Content characteristics over time

Future Predictions

Because of this tendency for content itself to become more personalized, I see the evolution curves of both content and software becoming increasingly intertwined over time.  Open courses in particular begin to bring the trajectories of open content and open software together in a way for large, broad sample sizes of users to help us to statistically validate (or alternatively capture more anecdotal evidence) about how to tailor content more appropriately to individual students.

One future possibility is that many instructors may collaborate to build a single, complete “master copy” of a course which can be syndicated and customized for an individual instructor’s use, sharing in the burden of creating, suggesting, and updating course materials. When used at scale, this “master course” may collect analytical data and human feedback from hundreds of institutions and suggest to other instructors which materials work best and in which order. Instructors may also add remedial materials for students who need reinforcement of key concepts and stretch materials for students who want to explore more – something that would be a significant amount of extra work for an individual professor.

Open continues to change the dynamics of education and will continue to do so into the future.  Open should not be seen as a threat – but rather an opportunity maker.  It will undoubtedly change the ways we collaborate, partner, and differentiate. I encourage everyone to see the possibilities and think creatively about what open enables. We can shape how things will materialize because of open, and by better understanding open I believe the better things will be.

This post written by George Kroner

 

The Untimely Death of Community Source

I was disappointed to learn recently of the passing away of Community Source.  Like many efforts in higher education that have aimed to enhance and support institutional collaboration around software, the primary effort highlighted in this piece, Kuali, provides evidence that it, too, is falling short. My feelings of concern for Kuali began in April when the Chronicle celebrated Kuali’s 10-year anniversary but revealed that only 74 dues-paying members participated in the organization with only dozens actually using its software. This is simply not the momentum that is needed to sustain an effort of this type given its goals.

Kuali is not, however, the only education-related organization that has experienced challenges recently and needed to transform itself in some way. My feelings for Kuali are actually not too dissimilar from the feelings I had when JASIG and Sakai merged to form Apereo when each of the organizations was under financial duress. Jisc, previously an entity of the UK government, recently reorganized into a registered independent charity to continue and sustain its work triggering similar structural changes to CETIS and OSSWatch. Edtech start-ups and established for-profit companies, too, continue to shift strategy and pivot, transforming to survive through mergers and acquisitions, new business models, or by going public or private. Even whole universities themselves are considering shifts in structure to become more nimble.  In the ways that these organizations live on despite change, I’m not entirely convinced that Kuali’s “pivot” indicates that the spirit of community source is dead nor does it indicate that a successor to community source is not viable.

There are many reasons that Kuali in particular may not have taken off as well as had been hoped (some might even argue that it actually has taken off). Maybe the project is too large or all-encompassing. Maybe it is trying to meet too many diverse needs. Maybe institutions just don’t want to accept the level of risk associated with using new software, particularly for administrative systems, for which as a foundation they are directly responsible. After all, it’s easier to blame and penalize a vendor. Maybe investment in a switch or migration would be too complex or cost prohibitive at this moment in time. There are many considerations to ponder.

But even so, I hope that we collectively take the time to reflect on both changes in the software industry and the opportunity that collaboration around software in higher ed affords us rather than simply writing these efforts off as dead. Maybe community source just needs a nudge of its own to transform into a new and more successful model. If this is the case, what are the things that we should consider and ask when making a determination of what a successor to community source might look like?

From what other examples of community source can we learn?
Michael mentions Apereo OAE as one project that has seen greater success by being freed from the consortial/foundation approach. I would also argue that the Sakai CLE also continues to experience more rapid improvement now that the encumbrance of decision by committee has been removed. Are there any others?

Are there any other organizations that have similar collaboration models but have achieved greater success?
If we strip the software pieces away, the IMS Global Learning Consortium (the organization behind popular, open edtech standards such as LTI and Common Cartridge) essentially operates in a similarly-closed consortial approach to the development of open standards. Yes, though these standards end up being open, one must pay to have a seat at the table in driving these discussions. The difference is that while foundation-based, consortial-driven efforts like Sakai and Kuali have faced challenges, IMS has become more successful than ever. Why is this?

How has the perception of open/community source changed in the past 10 years?
Open source software at one point in history was pitched as a way to beat commercial software providers at their own game – as an alternative or a check to commercial software rather than a flavor of commercial software. Now, I hear open source increasingly referred to as an “insurance policy” to avoid vendor lock-in rather than a competing model for building software outside of commercial channels. Does software really need to be developed in a consortial, non-profit manner if the goal is for the code to become open sourced? If collaborative software efforts were managed more like products than projects, would this make a difference?

How has software development changed over the past 10 years?
Several environmental factors have changed since Kuali (and community source) were first envisioned. Software itself is easier than ever to develop. Infrastructure is easier than ever to acquire for purposes of developing and running software. Technical limitations related to performance and scale have become more easily addressable. Standards allow software to be delivered in a more modular manner. The barriers to develop software continue to drop. In light of these changes, should the challenges Kuali faces dissuade similar development approaches or strengthen our resolve to try and find a better way?

How do these changes impact community source software adoption?
Increasingly, institutions prefer cloud or SaaS-based software solutions to in-house ones. Data centers are expensive and so is the experience necessary to keep them running. Because of this, not only has the nature of software development changed but also that of software delivery. Should institutions find a way to operate and run software on shared infrastructure?

Have dominant designs emerged for particular categories of software used in higher ed?
Software tends to become commoditized over time forcing it to compete in different ways – for example by supporting better usability or in edtech, pedagogy. Has the maturity of the types of software that higher education institutions use – such as the LMS and the SIS – led to any dominant design that would provide a good candidate to replicate in a community-source-like fashion?

Does a decision to collaborate make financial sense?
Once a dominant design emerges for any type of software, it becomes increasingly difficult for incumbents to sustain competitive advantage. Competitors and open source options eventually encroach on the existing product’s customer base as features and functionality become easy to replicate. Michael mentioned Instructure in his post. In the case of Canvas, there was a business case to be made to build a new LMS – with better usability and a superior delivery model, but nonetheless still an LMS. Instructure managed to build their product on an incredibly short timeline and now have a software solution that rivals the best and has even surpassed the community-source-based LMS Sakai in terms of market share. They proved that it is easier than ever to launch a new, competing software product in higher education, and their product is (mostly but not completely) open source. Related – just because it would be easy to build a product, does that mean it should be built?

Is certain higher education software more amenable to community source than others?
While administrative software like Kuali is not sexy, it is exorbitantly expensive. Universities are most likely never going to “compete” based on what back-end software they use. For some systems that are expensive and handle back-office tasks, maybe a development model like community source does make sense. For front-end software that directly impacts the student experience, however, maybe institutions want to reserve some way to differentiate themselves and not choose to collaborate on a common one-size-fits-all solution. Alternatively, maybe front-end software would prove too cumbersome to support from a browser/device/end user perspective. Or maybe a common, front-end software experience would open up new possibilities for collaboration among multiple universities for which a community-source-like approach could accelerate new opportunities.

If the goal is open source, do other models besides community source perform better?
There are a couple models where open source software tends to be most successful. One is when the software is a component or framework of a solution rather than the solution itself. Another is when a small, core group (for example, a commercial entity) contributes most of the heavy lifting but freely distributes the code and accepts smaller contributions from others who have interest. Instead of relying on employees from several universities (each having separate employers) to contribute to a community source project, could it make more sense to organize the core development team under a single organization to better-align getting the work done?

Could/should institutions partner differently or in lighter-weight manners but still get the benefit of community?
Maybe it’s worth considering that institutions partner in different ways for different reasons all the time, and not all of these necessitate a new foundation, a huge software development project, or a governing board. Sometimes collaboration is as simple as a virtual handshake and an e-mail between a group of individuals – when we choose to let it be so.

At the end of the day, Kuali may be facing issues, but I would not be surprised to see a successor to the Community Source model emerge. While the original model may have its flaws and potentially have run its course, building on lessons learned from its challenges and shortcomings, we can do better.

This post written by George Kroner