There is yet another new wave of start-ups emerging in the educational technology space and like those that came before, most of this new wave neglects to address some critical issues. Every few years, a new set of companies comes out with what they refer to as, “the next wave in digital education.” However, these “new” methods and technologies are rarely actually new. Experienced educators who have followed the evolution of digital education since its inception over fifty years ago, have seen it all. The new distribution technologies offered by the new web don’t actually enable new pedagogies that haven’t been tried yet. Since the mid-1980s, there has been adequate technology and tools available to allow us to try out the entire array of pedagogical theories. Believe me, every combination of existing tools has been employed, and with a slight variance from subject to subject, very few methods used in conjunction with technology have been effective at improving educational outcomes.
For instance, numerous attempts have been to incorporate video training into Mathematical Sciences (which, since the 1980s also includes the field of Computer Programming); it has failed in both our school systems and in corporate training every decade since the 1950s. Each decade a new distribution technology for video comes to market and education companies try to seize the moment by using the new distribution model to pump video lectures into the system. In the 60s it was TV, in the 70s it was closed circuit two-way TV, in the 80s it was video tape and CD-ROM, in the 90s it was the web, and now it’s Youtube. However, these efforts continue to fail to make a positive change in education because each fails to recognize that the problem lies not the distribution and availability of lectures, but in lecturing itself. Lecturing does not propel people to learn to solve problems in subjects like mathematics, physics, and computer programming. For that reason, regardless of the way they’ve been distributed, the videos ultimately fall out of favor. And even though we’ve seen it happen before, it takes longer for the eventual admission of defeat to occur each time around, because the videos are delivered by increasingly efficient distribution models that spread them farther and wider. I think it will take ten years for people to realize that Khan Academy, the latest addition to the education video pile, isn’t making students better at mathematics (at least not at the Calculus level or higher). It’s possible that math education at the K-12 level is in such a sad state that Khan Academy may be improving that situation. And while I agree completely with Khan’s recommendation that instructors move away from the podium and into the role of coach, there are more and better ways to flip the classroom that we can demonstrate with Making Math (see: http://makingmathblog.org for updates).
I am pleasantly surprised that Codecademy has emerged as the latest online learning system to follow in the footsteps of the O’Reilly School of Technology and the Netmath program at the University of Illinois, and create an opportunity for itself to take a pedagogical stand to improve the way we teach with technology. Let me first say I like what is there so far! Currently, Codecademy is executing a what is eerily similar to an earlier version of what we started with back in the mid 1990s with Useractive, inc. and what eventually landed here at the O’Reilly School of Technology. But as is often the case, the web is abuzz about this new kid on the learning block. It is in essence a carbon copy of the first version of CodeRunner that Trish and I built in the mid-1990s; in our older version we too initially offered tutorials for free and gathered a lot of signups. Our effort back then ultimately stalled, and hopefully the web has changed enough with the addition of social networking that Codecademy can get a better foothold that we did. We learned the hard way about the way people learn, as well as the constraints of various education and training markets since then, lessons that Codecademy might have to learn as well, in order to stay afloat. This post could be of real value to them on their path toward refining their system of online education. I hope so.
Before, reviewing and analyzing Codecademy though, I’d like to address the issue of armchair journalism in the Silicon Valley and the press coverage of educational products. Armed with a blog and twitter account and then attempting to report on educational technology (a topic with which most have little to no experience either building or using), many of these “journalists” will latch onto anything “new” to them, and call it “revolutionary” without offering any perspective or context. It’s sadly similar to the way people react, for example, to new fitness products, thinking they can get rock hard abs and look like a professional athlete by using some exercise product 30 seconds a day! If you want to get the blogosphere and twitterverse raving about your online educational product, here is how:
The recipe to make your online educational product instantly popular:
- It must be free, with minimal transaction effort.
- There should be exaggerated claims of ultimate learning outcomes without evidence. Place the promise on the future, not now.
- It needs to make people feel like they’ve learned something in a few minutes by giving them a thumbs up, or badge for accomplishing something trivial.
- The copy on the site needs to tell the visitors to your site exactly what to say about it, regardless of veracity, because most reporters won’t take the time to actually go through the lessons themselves (or if they do, they won’t know enough about the subject matter itself to make an educated statement about it).
You do these things, and voila! You’ve got a popular educational technology that the twitterverse will love and the Silicon Valley will throw money at with reckless abandon!
I don’t think it was planned, but the founders of Codecademy accomplished those tasks, and collected $2.5 million for it. For that they deserve a big congratulations, because now they have the chance to do something real. The purpose of this post is to lend both educational and historical context to the story that so many journalists seem to bypass, and to share my own perspective and experience. I am motivated by a desire to prevent others from the pitfalls I have already encountered, and further the field that has been my passion for the past twenty-five years.
At first glance, Codecademy’s technology and pedagogy bears a striking resemblance to ours here at the O’Reilly School of Technology, specifically to our web based code editor we call CodeRunner. There are some things I really like about what they’ve done so far (there isn’t much elearning I’ve liked over the past couple of decades). Here is a screen shot of their “web based code editor,” much ballyhooed in the blogosphere and twitterverse:
Now compare that to our CodeRunner system here at the O’Reilly School of Technology:
CLICK HERE to see the output a student see of the simple lesson being worked on in CodeRunner above.
It’s showing how to use the createElement method to add content to a web page dynamically.
Clearly, these two interfaces have some similarities, and I’m very pleased that people do recognize their respective values. I’m also glad that they enjoy using Codecademy, because so far they seem to be taking a step in our general direction, and a step away from static text and ineffective video lectures on the subject. As you can see, both learning environments contain some of the essential elements of what we call “useractive” style lessons. (Useractivity is interactivity where the end-user can create a project of their choosing, using the environment provided. An example might be to compare an ATM machine, which is interactive, to a laptop, which is useractive.) They both have places to edit code, they both have content that can be seen and followed while editing, and they both have a way to track progress. Yay!
But there are also some important differences between what we’re doing at OST when compared to Codecademy, but that’s is to be expected given that they are just starting out. First, their editor is not yet an actual tool that can be used by students to create interactive web applications themselves, and therefore it is extremely limited in scope so far. Further, their files can’t be saved and referenced, their students don’t have access to a web server, and not even their Javascript lessons call upon students to create actual webpages, which is a pity because that’s where the real power and fun of web programming begins. In addition, Codecademy students don’t type in the initial examples they’re working on themselves, which, in my experience, helps to transfer ownership of learning from the content to the student, and also allows them to experience real coding. Finally, and most importantly Codecademy offers no hand-in functionality, nor an available instructor to coach and give feedback to the students (yet!).
The most valuable lesson we learned over the years in building the O’Reilly School of Technology is that you cannot replace a human “code coach” with technology. The help, feedback, and encouragement given by our instructors are key to our learning system. Technology serves to get people started and prompts them to experiment, which leads to questions and teaching moments that are then utilized by our instructors to guide our students to solve problems and work toward greater understanding though their own reasoning power.
After a student finishes an OST course, they are invited to review it. In these reviews, students consistently rave about the OST instructors who have helped them through difficulties with their coding projects and who offered them encouragement and guidance. Our CodeRunner interface and lessons simply serve to set up these unique and constructive conversations between student and instructor. Without instructors on the other end to help students, I believe educational technologies are worthless.
The type of system we designed for OST (and now Making Math) is called a Tool Integrated Learning Environment
(TILE). A TILE consists of three parts built and integrated to work seamlessly with each other. An abstract diagram of a TILE system looks like this:
The tools are real, web-delivered production environments like our CodeRunner code editor, Eclipse, and Visual Studio, with modifications and a new online custom version of Mathematica we call Hilbert. So far we’ve created courses in these subjects: HTML/CSS, Javascript, Java, C/C++, Python, Perl, Linux, SQL, PHP, C#, XML, DBA, Algebra&Trig, Calculus, Differential Equations, Probability and Statistics, Linear Algebra, Matrix Theory, and we have the capabilities to teach many more. All courses start with the tool, then the content and Learning Content Management System (LCMS) are added to comprise a complete course. The LCMS supports human feedback, grading, and conversations, as well as the authoring and editing of course materials. The content is written specifically to drive students to type code and to learn the subject matter through code (even mathematics). A TILE is like three-legged stool; the removal on one leg will cause the stool to fall, much like the loss of any one leg of a TILE breaks the pedagogy and the learning system falls. From what I can tell, Codecademy currently has a two-legged stool that offers content and a code editor.
So, for all of my beliefs about the way online education ought to be presented, if people are learning about coding though Codecademy, why am I concerned about which features comprise that learning system? My first thought in response to that question is to consider whether people are really learning to code or do they just think they are? Who decides? Can people who go though those lessons actually open up a programming interface and create something from scratch? Can they explain the coding language to someone else? I doubt it, not yet anyway. Would you hire someone whose experience lies in having gone though some computer lessons and received a few badges for putting in a variable here and there? I wouldn’t, but perhaps that isn’t the point. And even if their current solution is only for children, why constrain that solution to text-only output and offer no instructor? I don’t think they intend to stay that way, but adding more functionality beyond client side execution of Javascript will start to raise the cost of serving it up. Finally, from a business perspective, if people going through Codecademy actually do learn some coding techniques, would that education be of enough value for someone to actually pay for it? I don’t think so. Though it’s possible that Codecademy intends to offer free online tutorials forever, and they have something really clever up their sleeves.
Most problems with new learning systems begin when companies try to monetize them. I anticipate that they plan to resolve some of their system’s current shortcomings, and will work to make it more capable and dynamic, but even if they were able to make something great by increasing its scope, or even adding human instructor feedback, they may still fail in the end because of some not so obvious realities and constraints of the education and training markets. How do I know? Because I’ve been there, and I discovered the hard way, through trial and error, that there is one specific and huge challenge associated with simply creating a great online learning system for end-users and anticipating good results (especially in Computer Programming)–there isn’t any money in it. In addition, there are some monumental differences in the dynamics of each of the vertical markets in the learning space. The only way to make money in education is to make sure your offerings meet ALL of the needs and expectations of your target market, only then will your products have value. But each vertical market has a completely different set of parameters for a learning product. There are different decision makers, different end-users, different pricing structures, different branding considerations, and different subject matter considerations.
The value of online education products is completely driven by the intrinsic and extrinsic requirements of the decision makers, who may or may not be the buyers or the end-users. In trying to meet the requirements of these markets, as well as those of various decision makers, you will be pressured to compromise your learning philosophy and pedagogy. The only way I’ve found to preserve my own teaching objectives, finance my philosophy and pedagogy, and still satisfy the intrinsic and extrinsic motivations of the decision maker, is to run a technical school behind a strong brand, like O’Reilly. When you operate a school, the decision maker is the buyer and the end-user. Because they have the extrinsic motivation of obtaining a degree or certification to improve their long-term job opportunities, they have the motivation required to work through difficulties in order to become proficient at their chosen field of study. They are also willing to meet the pricing demands needed to cover the costs of quality instruction and continued R&D. Their extrinsic motivation helps you meet the promise of satisfying their intrinsic motivation of becoming a good coder with the needed skill set to pursue their career path beyond their studies.
Another important lesson I learned in regard to online education is that most online learning products are not paid for by the end-users (students). They are paid for by schools, by companies, or financed by state and federal governments. No one, and I mean no one has figured out how to get end-users to pay for individual courses online unless they are attached to some kind of degree or certification. They will pay for books, ebooks, or they will pay for subscriptions to enormous amounts of online reference materials like Lynda.com or Safari Books Online, but even in those cases many of those end-users will be reimbursed by their employers. Simply put, end-users will not pay much (if anything) just to learn online.
Or maybe Codecademy has ambitions of creating a new vertical learning market? Online education and training markets are generally too mature and established for anyone to expect to create a brand new one. Of course, Amazon and Apple helped to spur the ebooks market by investing billions of dollars, so it’s technically possible to invent a market with that kind of investment behind you. If you don’t have billions of dollars though, then you HAVE to fit your product into an existing pipeline and you have to do it in a way that meets ALL of the expectations of the buyers. Hopefully, in the interest of quality education and success, you can discover ways to satisfy those expectations without compromising your vision.
At the heart of my current discussion of Codecademy is a good understanding of where they stand right now, where they need to be (which is a long way off,) and knowing the exact mistakes they are on course to make in the coming years. As I mentioned before, each type of learning product is ultimately tied to business models and markets, so it will be interesting to see which direction Codecademy goes. Right now, Codecademy doesn’t fit into any market, unless you count free tutorials as a market (which it could be if there were enough people in the world interested enough in programming to drive enough traffic revenue to support a large team or pay back investors, but currently that market simply isn’t big enough. ).
Codecademy and others can learn from our experience:
I’ll finish up by sharing some background about Useractive and the O’Reilly School of Technology. At the very least it should serve as a warning to people who want to create new educational products. Most people have the same misconceptions about the world of education and learning that I had at the beginning. For instance, I used to think people went to college to learn. That’s not true. People go to college to become validated. Learning is a means to that end, not the end itself. Think back to how you chose your college. Did you ask how they taught their courses or what their learning outcomes were? In most cases, probably not. You wondered about their reputation and the rates at which their graduates got jobs or got into graduate schools. Colleges are about branding and validation, period. That’s why they can afford to post their lectures and learning content up on the internet for free; that’s not what they sell. They sell validation with education as the tool to get there. This is but one misconception most people have about training and education. There are many others.
From my own experience starting an education company, I’ve learned that anyone can put up a free tutorial and attract hundreds of thousands of users. I did it in the 1990s and early 2000s when I founded Useractive, which was initially a free and almost identical system to Codecademy. People loved it and we were praised mightily for it. Eventually, I decided to try to turn my work into a business, and like Codecademy, I also attracted investors. Over the next five years, we tried all kinds of ways to get people to pay for our offerings. We tried up-selling hosting accounts to students who could save their pages and code projects they were in the middle of completing. We tried making limited versions of lessons and up-selling them to purchase continuing lessons that offered more detail and advanced topics. We tried selling the service to schools and to businesses. It seemed everything we tried met with failure because the demands of each market weren’t really related to learning outcomes. It wasn’t until we added instructors to coach our students, grade and give them feedback on their projects, and simultaneously partnered with the University of Illinois and O’Reilly to offer professional certificates for the completion of these courses, that we started to make headway. Eventually, O’Reilly bought my company and we changed the branding of the O’Reilly Learning Lab to the O’Reilly School of Technology. Finally, it began to come together. We were meeting ALL the extrinsic and intrinsic motivations of the decision makers, buyers, and end-users who, in this case, were all the same people. We are now exploring the idea of becoming accredited and ultimately offering degrees ourselves. Universities, both private and public, demand the highest price points online, even though their current methods are poor, because they meet so many of the traditional extrinsic motivations if not as many of the intrinsic ones. We here at OST hope to up the ante in online education, by meeting both of those motivations the best ways possible.
Here is a short list of some of the lessons we’ve learned about education and training markets:
- Individuals who are in a casual or DIY learning mode, won’t pay much, if anything at all for online courses. They’ll gorge themselves on free learning resources or they will buy books and ebooks which they consider to be a handy references to keep permanently in their bag of tricks.
- Individuals and companies alike will pay for huge banks of learning and reference resources on a subscription or “all you can eat” basis, especially if it’s content they can’t get for free. (There has to be an enormous amount of materials, well into the thousands of whatever the selected format might be–lessons, books, videos–included to add value.)
- Individuals who are looking to take online courses with instructors know that learning takes a lot of work and commitment, and for that effort they want much more than simply to pick up some skills without proof of that achievement. They want some kind of validating stamp to put on their resumes from a reputable organization that companies will respect. They are willing to commit if they are convinced that not only will they gain the skills they need, but also earn a piece of paper with a strong reputable name on it. In some cases these courses are financed by title IV, or reimbursed by their company. Accreditation and a strong reputation is best for that purpose, with exam prep certification courses for the likes of Cisco, Java, and MSDN coming in a close second.
- Companies looking for online learning courses are mostly looking for a one-stop vendor of learning resources for the least amount of investment (like Skillsoft). They are usually just trying to acquire these resources as HR benefits to employees and they care more about quantity than quality. These “courses” are usually paid for on a subscription basis. When they want something really good for a specific business need, they look for custom classroom training from consultants, or custom online courses created specifically for them.
- At colleges, professors are the decision makers, but students are the end-users. In this case, products have to be made with the professors’ needs in mind first, and it must make their jobs easier (which is difficult to do given that they normally just walk into large classrooms and lecture for an hour or two three times a week). For the most part, they love systems that grade short answer and multiple-choice questions for them, so they can concentrate on research and other scholarly activities.
- At the K-12 level, schools don’t have money for “extras” like computer programming. (Unfortunately, most don’t even have money to teach trades, music, or art anymore. Personally, I believe that needs to be fixed before anything else.) They are having a difficult enough time supplying the subjects they’ve deemed most important like Math and Science (which explains why the free Khan Academy is currently so popular at the K-12 level).
Ultimately, I wish the folks at Codecademy the best of luck for their success and offer my assistance to them toward that end however I can. I sincerely want them to hit a mark with this style of presentation, because as Tim O’Reilly likes to say about business, “A rising tide lifts all boats.” .