The Wrong Robots, Illustrated

A while back I wrote a post that ended with this graf:

Calls for efficiency in education are fine, and talk about affordability and social justice is critical. But by the time that teaching — one of the hardest jobs to automate — is significantly automated we will be at the end of the robo-revolution, not the beginning. And if we really think that’s the case, what we’re teaching students now seems a much more pressing issue than how we’ll teach them years from now. What we need above all else is not an education that is powered by automation, but an education that is a response to it.

Today, via Quartz, comes a decent visual to show what I mean. I’ve drawn up my own annotations here, but click the link to explore on your own:


The vertical axis is pay and the horizontal axis is “likelihood of being automated”.  As you can see, the likelihood of teaching being automated before computer programming is pretty low.

Now consider that we are trying to save money on preparing the future programmers and benefits managers of the world by attempting to automate teaching. What’s wrong with this plan?


We had a different system once. It worked pretty well.



Slideshare has been good to me. Outside my surprise Latvian Top 40 hit,  my “Slidecasts” (PowerPoints plus audio overlay) rank as some of my more popular productions. If the Slideshare analytics are to be believed, my combined presentations on Water 106 (I did a bunch) have been viewed collectively over 4,000 times. My presentation on Open Course Frameworks has another 1500 or so views.

But today I’m doing that all too familiar dance of “downloading my stuff before the third party service deletes it”.  Slideshare is discontinuing Slidecasts, and has tossed my stuff out on the street for me to collect, in true rom-com breakup style.


I’ve got to get them off the server by the end of the month. And by “get them off the server” I mean a seperate PPT file and audio file,  which are not the Slidecast, but the materials I uploaded to create the Slidecast before a good hour of syncing markings.

Some might just sigh “capitalism” and leave it at that. But I don’t think it’s that.

Because we had a system, once, that mostly worked. It was based on a seperation of concerns. As an example, back in 1997 or so I bought a computer from HP and I bought a program from FutureWave Corporation called “FutureSplash”. I made flash-like animations with that program and saved them to my hard disk. When Macromedia bought out FutureSplash, renamed it to Flash and radically changed the format, I didn’t get a notice that said “Your files will no longer work.”

Everything continued to work. I could open them, edit them, work with them. Macromedia didn’t own my drive, or even (really) the instance of the program its predecessor had sold me.  I didn’t have to worry that a TOS had changed and now my files could be mined for personal data or used in advertisements. And incentives were better aligned, as well upgraded to Macromedia Flash not because my stuff wouldn’t work without it, but because they had great new features. Imagine that — people paying money to upgrade not because someone had a gun to their head, but because paying that money allowed them to do new, neat things.

All that was capitalism, all that was companies, and all of it kinda-mostly worked.

For a while this changed, because we suddenly needed peresistently available web storage and browser-runnable apps. It made sense to push that all up to the server. And it made the coding easier as well, not having to write for all the different platforms. And suddenly the old system was gone — your files, hardware, and code were all controlled by the entity that handled your web display and bandwidth.

My point in previous posts about personal cyberinfrastructure is that one way out of this is everyone will have a server somewhere with open-source apps on it. And that may work. But the thing is I don’t think most people want to run their own server. I think they want to own what they buy, more or less forever. And I think they don’t want to be in thrall to a software subscription plan or shifting ToS just because they *don’t* run their own server. And the weird thing is that we had that, mostly.

The storage-neutral idea is just part of a seperation of concerns critique of this new paradigm. As we move away from browser-based software and back to apps that run at least partially on our devices, do we want to preserve the sort of vendor control that emerged in the web-as-middleware age? Or do we want to reintroduce the separations we lost?

Slideshare, ultimately, doesn’t spend any real money on my editing of these things or the storage of the files. Or even the editing software. But the bandwidth is killing their bottom line. So because the bandwidth is not separated from the service or the storage, everything must go.

That seems really broken to me, doesn’t it to you?







Google+ will die, because Google already owns the one lifestream that matters.

So one of my non-edtech tech predictions in January was that the OS-based lifestream would kill the web-based mega-service, discussed most clearly in “Revenge of the OS”, but also in the slightly later article titled “The OS-based Lifestream Will Kill the Web-based Mega-Service”.

Well, the end of the year came early in 2014, because this is already going down. From the TechCrunch article “Google+ is Walking Dead”:

What we’re hearing from multiple sources is that Google+ will no longer be considered a product, but a platform — essentially ending its competition with other social networks like Facebook and Twitter.


As part of these staff changes, the Google Hangouts team will be moving to the Android team, and it’s likely that the photos team will follow, these people said. Basically, talent will be shifting away from the Google+ kingdom and towards Android as a platform, we’re hearing.

Why do this? It’s not because Google lost to Facebook. It’s that Google, via Android, has already won. Facebook is an app for sharing pictures and news stories. For a while it was a contender to own the thing that really mattered — your identity management architecture and your lifestream. But guess where IM and lifestream functions live now:


Your phone OS (and your Windows 8 OS, and your Xbox OS, etc. etc.) serve the platform functions that Facebook reached for and missed. It’s over.

Google understands that, which is why it’s moved so much of its Google+ team into building former mega-site features into the OS. Microsoft understands that, which is why it shipped the half-baked, but theoretically well-founded, Windows 8. Apple has probably known this a while which is why it just waited.

If we want to change the future of technology, that’s where we have to be. I don’t quite know what that means yet. But it’s time to start talking about it, and stop waging the battles that don’t matter anymore.

Federated Wiki, explanation as of April 2014

Someone emailed me and asked if I could point to posts on my blog that explain the federated wiki idea. I started to pull together some links, but realized that many explanations were out of date. I’ve been fooling around with this for several months now, and I know a bit more than I once did.

I find it easiest to start the explanation by talking about instructional technology support sites. But if you make it through that explanation, there’s a huge classroom application as well, and even huger OER implications.

[First, let me note I’m not talking here about Ward Cunningham’s Federated Wiki, which was largely the inspiration for this, but about our local attempts to make Dokuwiki work in a way that approximates federation. But I think the use cases apply to Ward’s work as well. Conceptually, I would love to use Ward & co.’s SFW as a base for this sort of stuff, but there are some barriers, so we’re left at the moment with Dokuwiki hacks.]

Ok, so here goes the instructional technology support explanation. It’s not the thing I’m most excited about, but it’s the easiest to explain.

There’s really two approaches to producing support documentation for your faculty. The first way, and the way most people choose, is to produce documentation more or less from scratch.  Here’s a recent write up of the tool Explain Everything by Judy Brophy at Keene State College.



Meanwhile, here’s a screencast about Explain Anything from Samuel Williams, who works in IT at the University of Portland:

If you look through Academic Technology sites, you can find half a dozen of these on this very limited subject. If look at something bigger, like uses of screencasting or flipped classroom implementation, you’ll find hundreds of these, all explained from scratch. That’s a lot of hours spent rewriting basic documentation.

Now, when we see something like this, usually someone will come up with the brilliant idea “We should have ONE CENTRAL SITE that WE ALL COULD WORK ON!” Imagine if Judy and Samuel worked together on the Explain Everything article. It’d be the Wikipedia of Edtech!

But there’s lots of reasons why such a thing doesn’t really work.

For one, maybe Judy has a very location-specific intro on that article (and actually she does, beginning it with a story about a Keene State professor). That’s a good thing, tying it back to the campus like that, and we don’t want to give that up for a wiki.

Second, when you send faculty to a generic wiki to solve their problem, they often look at you like you are throwing them off the boat. It took a lot for them to come to you with this question, and now you’re just spinning them off again into this generic resource. For various reasons, faculty find the presence of local support sites comforting, and see them as evidence that they are supported in their efforts.

Third, Judy and Samuel may have actual differences of opinion that make writing a single article hard. Samuel may think classroom use of clickers is the spawn of Hell. Judy may think it’s the Second Coming. When they send the faculty member to our group wiki, there’s an implicit endorsement of the views there — so who makes the call on the clicker issue? (Note: I really have no idea on what either Judy or Samuel think about clickers).

All of this stuff together makes it very difficult to come up with one version of things. So we continue on, either feeding faculty incoherent messes of links, or writing everything from scratch.


Doctor, Heal Thyself

What’s fascinating to me about this is that for a bunch of people that are constantly telling faculty about the glories of reusing materials we actually suck at reuse. And we suck at it for the same reasons faculty do. Coherence and reuse are orthogonal to one another. We’re not willing to sacrifice the coherence of custom materials, so we sacrifice efficiency.

What we need is a system that promotes revisable reuse in our own spaces.

So enter federated wiki. The idea is pretty simple. Let’s say we start with 20 colleges. We all install our own support wikis on our own sites. And then we “federate”.

What does this mean? It means when I make changes to my wiki, they flow to yours. When you make changes to your wiki they flow to mine. And not only to our individual wikis, but to everyone in the federation.

So I go into my wiki and write up this guide to lightly blended courses:

Now when I hit save, it saves to my wiki. But it also notifies all the other members of the federation that there is new content. The admins of those sites will have a number of choices when they see this new content.

  1. Reject it. We don’t want your stinking article on our wiki.
  2. Accept it into a special “drafts” folder, for editing and future posting.
  3. Accept it as is into an editor, make edits on the spot and post.
  4. Accept it as is, and agree that future edits the original author makes to it will automatically flow through to our wiki.

So let’s say that person accepts it as is. But then they notice that there are a number of spelling errors in it, and also that the section on the “Understanding by Design” approach is a bit lengthy. So they pull out the Understanding by Design stuff into a separate article,  link it, and fix the spelling errors. When saving it they put in the notes box that they did minor spelling edits and a pull-out of the UbD section.

So now everyone in the federation gets two notifications. First of all, the article has been updated. So they get notified of that. They also see the new article the reviser split off.

Not everything would have to flow through the federation. I might just edit the contact information on an article and decide to check the “Do not syndicate” box before I post. Entire portions of my site may not be federated. It may be that only one subdirectory or namespace is federated.

But you can start to see how this works. It’s like a wiki, but instead of editing a single copy, you “fork” one (and this, again, is the insight of Ward — on a wiki every page has an edit button; on a federated wiki, every page has a fork button). Changes flow back and forth through the network, but everyone maintains control of their own copy. And best of all, the backend would handle credit, so people who wanted to track who did what for the article could easily do that, and so that people who create broadly used pieces could get credit for their work (and warm fuzzy feelings about the amount of use it gets).


Academic & Pedagogical Uses

I actually started exploring this idea not because of support sites, but because of barriers I hit with an instructional design project called Water106. The idea seemed simple — get multiple classes from multiple institutions to interact in a common space of the web.  So if we were all looking at the issue of water from different scientific, engineering, and political angles we could produce a mega-wiki on water.

A lot of people want to do this stuff. Cross-institutional classes have been the “future of education” for a good decade now. But ultimately it’s a bit clunky to get going.  That mega-wiki has to live somewhere. So maybe I own it here at Vancouver. But then I’m setting all the permissions and talking to the faculty in these other places providing support. That’s not really what the instructional technologists at these other institutions want — because ultimately they’d like to build and support this thing without depending on coordinating with me.  We want to work together, but without the coordination hassle that could easily consume us.

Furthermore, there’s a pretty simple issue for classes. Instructors want outside influences in their classroom to promote critical thinking, but for grading purposes they want their students to have the last word on their work. They want to know when they are looking as a wiki article that it is in the state that the student believed was the “gradeable” state. They don’t want to have to sort through the issue if someone from outside the class modified it in ways that moved it away from student intention.

Federated wiki is a way around all this. If we all have academic wikis on the subject of water we federate them. Your students just do their work as if they were doing it alone, either on a wiki that your institution controls or on one they own themselves, with whatever security parameters, etc. that you/they need to have. But magically, this other work flows into the wiki that the students can approve or discard (or maybe even transclude?). And again, when students build pieces that get widely re-used throughout the federation, they can take some pride in that. To some extent, it’s like “Tumblr for wikis” — you post your stuff and see how many reblogs (“re-wikis”?) you get.



You know what? I’m not even going to go into Open Educational Resources. If you’ve been in the field a while, you can see that this sort of system could start to address the issues we’ve been trying to solve for over a decade. The problems we have reusing materials to help develop faculty are the same problems they have with reusing teaching materials. Solve the smaller problem and you’ve begun to crack into the bigger one.

I’ll just say this: most faculty have some materials/activities/assessments that work brilliantly, and a lot that are mediocre. You can start to see what is possible if 500 Psychology 101 professors start using federated wikis to build their course materials. A system where faculty are pulling the best and most effective bits of pedagogy from around the world is a system that could have amazing impact in education.

I’ve been involved with open education a good ten years now. I’ve gotten cynical about addressing the Reusability Paradox. Certain developments in the past couple years have un-cynicalfied me. One is the Lumen Learning approach to OER implementation. Another is this. Good software ideas/implementations can create the possibility of doing something; great software creates a *culture* of doing something. Federated wiki, properly designed, is culture-creating software.


When Will This Thing Be Built?

I’m working on Dokuwiki hacks I can, and Tim Owens has also given substantial time to it. I’ve played around with it enough at this point to understand that the primary barriers are not technical. Rather, the most difficult thing is to organize the process by which material gets syndicated throughout the federation in a way that people can easily conceptualize.

Meaning. with enough time to code we can get the files to do what we want. But that step of deciding what comes onto your wiki from the federation — what does that look like? It could get really noisy. It could get very difficult to know which updates of a dozen you want. It could be so time consuming to scan the updates and understand what they are that it’s easier to roll your own. Alternatively, the process could be so silent that everything flows into your drafts folder but never gets used. Things like attributing cleanly without gunking up the articles with too many lines of credits are also issues.  Tracking what you have approved and not approved may get confusing as well (Did I review this already?) unless the system handles it elegantly.

And I’m still toying with the idea of using the Smallest Federated Wiki that Ward & co. are developing instead of hacking Dokuwiki. If SFW only had an Installatron setup, we’d be halfway there (i.e. students get personally owned server space via DoOO, login in, punch the installatron button and voila!). SFW would also have to be Section 508 compliant though, to make sure it’s not difficult for the visually impaired to use it. It currently doesn’t seem to play nice with JAWS, and unfortunately that’s a hard stop for us.  So, until then, it’s Dokuwiki hacks for now. I’m a spaghetti coder on a good day, so it’s slow going.

Anyway, that’s the update. As I mentioned, I only have a fraction of my time to give to this right now — spending much more time getting Dokuwiki ready for internal use by faculty and putting together workshops for the summer. But I’m always willing to talk about this with anyone who is interested, so give me a shout-out on Twitter @holden. Or comment on this post.




Your Diploma Just Got Downgraded. But You Can Upgrade It At a 20% Discount!


From the comments on my last post —  friend of the blog (and Sloan-C Karoake instigator) Michael Berman lets us know he got some bad news about the Udacity certificate he earned in 2012.

Please note this is a real email, from “Amanda Sparr, Coach @Udacity”. This is not a parody. (Really!)

Dear Michael,

Nice work on earning a certificate for the Intro to Computer Science courseware. It speaks volumes for what you learned while completing this rigorous course. Congrats again!

Today we upgraded this course . This brings you new opportunities, for little extra work:

  • You can now practice and show off your skills, with a new hands-on project where you’ll build a social network.
  • You can also earn a Verified Certificate, since this course now has a subscription option.

Udacious Coaches like me work with you and award these Verified Certificates. We review your project code, share feedback and tips, and conduct a final video interview where we verify your identity. With these steps, employers see our Verified Certificates as an even stronger endorsement of your skills.

To get you started here’s a 20% discount*:

Oh my. I don’t know where to start. The fact that the “magic formula” involves personal tutors and graders? The line “Udacious coaches like me….”?

But of course the kicker is the business model: hey, we know you earned this certificate, but it’s kind of worthless now that we only do Verified Certificates. But, look, for a low, low fee you can earn it again. For real this time!

I never realized the answer to the college cost crisis was right in front of us all the time. We just need to downgrade everyone’s degrees, then let them know for a low, low price and a short trip to a testing center we will continue to verify their diploma. Instant cash flow! And accountability!

Oh, man. You really can’t make this crap up.

The Sieve Manufacture Continues at Udacity

From Udacity last week, regarding the phasing out of free certificates:

“We owe it to you, our hard working students, that we do whatever we can to ensure your certificate is as valuable as possible.”


We have now heard from many students and employers alike that they would like to see more rigor in certifying actual accomplishments.

Jonathan Rees has more on this odd phrasing about what is essentially a decision to charge students for what used to be a free education:

Perhaps I’m reading too much into this here, but I think this announcement raises profound questions about what education actually is, or perhaps simply what it’s supposed to accomplish. Is higher education a good thing because of the skills it represents or is it a good thing because you have it and others don’t?

To which I’d answer, no, you’re not reading too much into it. As I said back in November about Thrun’s sudden pivot:

There’ll likely be lots of analysis on this article and change in direction. He’s my little contribution. Thrun can’t build a bucket that doesn’t leak, so he’s going to sell sieves…Udacity dithered for a bit on whether it would be accountable for student outcomes. Failures at San Jose State put an end to that. The move now is to return to the original idea: high failure rates and dropouts are features, not bugs, because they represent a way to thin pools of applicants for potential employers. Thrun is moving to an area where he is unaccountable, because accountability is hard.

If thinning pools through creating high-failure courses is good, then thinning pools through creating high-failure courses and costly identity verification is even better. It’s MOOCs as Meritocracies, and it’s just as dumb (and culturally illiterate) an idea as it was in 2012, when the Chronicle was drooling over it.

Meritocracy is not a system, but rather a myth power tells itself. It’s openness as a privilege multiplier. It’s the antithesis of open education, which must measure its success in outputs, not inputs. You can’t claim you’ve granted people access to the top shelf of goods if you don’t provide them a stepladder.

I would disagree with one small point in Jonathan’s post, however. He indicts all MOOCs for this view. There are, I think, significant differences in the approach of both edX and Coursera to these issues. Most notably, edX has Justin Reich on board, and Justin Reich’s research is in exactly this area — how do we make sure that openness closes gaps rather than widen them. But I’ve also seen Coursera publicly grapple with this question in a way Udacity has walked away from. It’s clear to me that Coursera at the very least is *committed* to gap-closing as a principle, even if they have not yet acheived that in practice.

And that’s a difference worth noting and encouraging. There will be fundamental changes coming to Coursera’s model soon, and the question is which way will they lean. Let’s hope it’s not towards the sieve market Udacity continues to pioneer.

Are Blogs the Vinyl Records of the Internet?

From Blogs are the Vinyl Records of the Internet

The quote comes from a full article in the Washington Post about the decline of blogging in Iran. A few years ago, Iran emerged as a culture filled with high traffic, powerful blogs. It was called Blogestan. But, these days, as in many other cultures around the world, personal blogging is retreating in favour of corporate social media sites such as Facebook, twitter, and tumblr.

Two things.

First, the premise of the article is vinyl records are better (like blogs) but they have no impact on the record industry anymore.

I don’t quite know how to judge that. What would it mean to have impact as a *format* rather than as an *artist*. Maybe I’m being stupid, but I can’t really grok it. It makes sense as a statement only if you equate impact with money. In which case, yes. On the other hand, the sonic textures that you will be enjoying in two years on your iPod are spinning right now on a short run vinyl release in some Athens, GA apartment. You haven’t heard the band, and you perhaps you never will. But just like so much in pop music today can be traced to aesthetics of mid-to-late 00s releases, the same thing will happen again. In fact, I would not doubt that the Future Sound of the Format Formerly Known As Rock is floating around currently on cassette. So format, schmormat. In philosophy I believe Gilbert Ryle called such comparisons “category mistakes”.

Secondly, I’m still on my tunnel-vision streak of seeing the re-share issue everywhere. Blogs have been replaced by….Tumblr? Um, Tumblr is at heart a blog with a reblog button. That’s the innovation there. The behavior (and integration of the non-writer into the community) is a result of that. And in fact, as you look at migration to online communities and away from free-standing social media, one of the big sells is the re-share (or retweet, re-pin, whatever). Worth thinking about why that is, and whether we in educational technology could learn a thing or two from that. A nascent thought. But still a thought. Maybe people who re-share but never create are more important than we think, right?