Skip to content

Meeting About Using Open Content Learning At UNISA

I’ve written about open content learning materials before. In that post I mentioned that I’ve been trying to talk to people in UNISA for a while now, and on Tuesday (3 June), I finally got the promised meeting.

I met with Prof. Wendy Kilfoil (Director of the Institute for Curriculum and Learning Development), and Mr Deon van der Merwe (Director of ICT) in order to tell them all about open content learning. Their reaction caught me completely off guard.

They already knew all about it.

I was told about the African Digital Library project, one of several digital libraries available to students. They were the ones that mentioned the possibility of making all the study guides and course material open content (since everybody copies it anyway), and letting more people write through RPL. Deon’s department has already started using branded freedom toasters to distribute some of the study guides. I was shown the even newer model that they were experimenting with. Deon was also one of the first individuals to sign the Cape Town Open Education Declaration.

So if these people are so hip and up-to-date, why isn’t UNISA using open content materials yet? The short answer seems to be bureaucracy and ideological inertia. UNISA has a terribly arcane management structure. I was speaking to Directors, but the Council are the people that need to be convinced. In addition to that, many lecturers are not comfortable enough with the technology to search the Internet for open content textbooks to use in their courses.

Developing open content courses also requires funding, and they can’t use unusual models like bountysourcing, since all funding for course development must come from a single authorized source.

There does seem to be one slim wedge in the door. UNISA is trimming it’s course selection, and in some cases courses will be completely dropped. One example was an Astronomy course. It would be a terrible waste to throw away all that material, so it would be a perfect candidate for being made completely open content. Still, making it available will still require some investment, and getting that funding is still a problem.

I went there mostly to advocate and educate them about open education ideas. What I should have done was go there to learn about the environment, in order to learn how to spread these ideas inside the organization. I only realized this later, after recovering from the shock of finding out how informed these people were.

It seems that the journey has just begun.

Magnatune is not Evil

Magnatune Logo

Hey. I know where you can get some music online. Shhh!

Why am I whispering? It’s all legal. You get to listen to streaming audio online for free. If you hear something you like (or love), you can shell out to download high quality tracks, or even have a CD printed and shipped to you. It’s the radio advertising model taken online.

Magnatune is a netlabel: one of a new generation of music labels that distribute their music online. John Buckman created the site after his wife had a disastrous experience with a traditional record label. Now he hosts DRM-free music under an Open Music license, lets people choose how much to pay, shares 50% of the revenue with his artists, and encourages people to give away three copies of the music that they buy, and much-much-more.

This guy isn’t a flower smelling infohippy. He’s just someone that listened to the plight of the artists, and actually did something about it.

Open Content Learning and UNISA

MIT’s OpenCourseWare project is giving free access to all their course notes, exams, recorded lectures and other course materials. All the material is open access and open content. MIT does not confer degrees for working through the OCW materials, but you are free to educate yourself.

I’ve tried using some of the OCW material, and it’s far from perfect. It was designed to be used at MIT, and it is therefore not in a format appropriate for self-learning. The distance learning format is a better way to present materials for self-learning.

UNISA started out life as an examining body. They administered exams and conferred qualifications, but they did not provide tuition. Later on they began providing tuition and distance learning services, while continuing to act as an examining body. Today UNISA is South Africa’s leading distance learning institution.

I think UNISA should start using and publishing open content materials, and give everyone open access to it. Most of their study guides are already available as PDF files - If they started using some of the open content textbooks available online, they could distribute the coursework for entire degrees on DVD.

I’ve been trying to tell people inside UNISA about open content and open access learning materials, but I haven’t had much luck so far. It’s difficult to find out who I should speak to, and the meetings I’ve been promised never happened.

How about this for the future of higher education: If you want to learn, you buy a cheap subnotebook, get your courseware DVD burned at a Freedom Toaster, and you educate yourself. If you want a qualification, you write UNISA’s exams.

Right now schools pay hundreds of thousands of rands for textbooks. If that money instead went towards developing open content textbooks, then each kid would simply need a subnotebook to read his textbooks on. UNISA has been developing textbooks and course materials for many years. They are the ideal agent for producing open content textbooks.

Semantic Web Application (Finally) Arrives

After years of hype about the semantic web, we finally have an application.

Powerset has launched a public beta of it’s Wikipedia search engine, and it’s overflowing with semantic web technology. Get more details here.

You Too Can Be A Bioengineer

Foldit

Protein Scientist David Baker and game designer Zoran Popović have created an online game called Foldit that they hope will become the future of protein engineering. You can download and play the game for free, and you’ll be doing actual bioengineering!

Foldit attempts to predict the structure of a protein by taking advantage of humans’ puzzle-solving intuitions and having people play competitively to fold the best proteins.

- Foldit Website

Read more about it at Technology Review.

The Bitrate of Life

I originally wrote this post about a year ago, when I was still playing with Drupal as a blogging platform. It somehow got lost in the conversion to Wordpress. What was gratifying back then was that Charles Stross posted a related talk that he gave a couple of days later. It was nice to know that I was thinking about the same things as one of the best new hard SciFi authors. I’ve updated it to reflect current technology and my better understanding.

I’ve often imagined what it would be like to have a wearable exocortex that could record everything about me.

Heart rate. Blood pressure. Food intake. Exercise. Medication. Everything I see and hear. Total recall. What bitrate could I record data at?

These are today’s hard drive prices (obtained from The Prophecy Shop):

80 GB R 448 R 5.6/GB
160 GB R 544 R 3.4/GB
250 GB R 605 R 2.42/GB
320 GB R 783 R 2.446875/GB
500 GB R 1050 R 2.1/GB
750 GB R 1688 R 2.250667/GB
1000 GB R 2700 R 2.7/GB

It’s obvious that at the moment, the cheapest per Gigabyte is a 500GB HD. It costs about R1000. Roughly every 18 months, the storage capacity of hard disk drives double. So let’s assume that every 18 months, you buy the cheapest per Gigabyte drive, and that it costs about R1000. By the time you buy the next one, the storage capacity would have doubled.

You therefore have 18 months to fill the first drive:

1.5 years x 365.25 days x 24 hours x 60 minutes x 60 seconds = 47336400 seconds

And since Seagate measures a gigabyte as 1,000,000,000 bytes:

500 x 1,000,000,000 bytes x 8 bits = 4,000,000,000,000 bits

4,000,000,000,000 bits / 47336400 seconds = 84501 bit/s = 84 kbit/s = 10 KiB/s

That’s about 1.5 times the speed of a 56K modem. Since we sleep roughly a third of our lives away, the base bitrate could probably be boosted to about 140 kbit/s (16KiB/s).

When you buy a new drive, you can transfer all the data you have recorded up to that point, and half the capacity of the new drive will still be unused. You’ve paid the same amount, but you’ve got double the capacity. Half of the new capacity is used to store all previously recorded data, and the rest is available for new data storage.

So this is the progression of average bitrates as you buy new drives:

500GB at 140 kbit/s (16 KiB/s)

1TB at 140 kbit/s (16 KiB/s)

2TB at 280 kbit/s (32 KiB/s)

4TB at 560 kbit/s (64 KiB/s)

8TB at 1120 kbit/s (128 KiB/s)

Wikipedia says that the max average bitrate for handheld profile DivX stream is 24KiB/s, compared to about 488KiB/s for full quality. That means we can afford to start recording a live (low quality) DivX feed about three years from now, with plenty of bits for other telemetry.

You have enough bits for biometric data. You can present your medical aid with full details of your heart rate for the last couple of years. Some medical aids already lower your premiums based on your gym usage - they will probably further lower them based on the medical data you provide them with.

You have enough bits to record every click and every keystroke on your PC. This means that it is possible to implement Jef Raskin’s idea that every action ever taken on your PC must be fully undoable.

The interesting thing is that bitrates are rarely constant. When activity is low, bitrates typically drop as well. Audio and video recording both can be done at reduced quality, which would further stretch your recording times.

You don’t have to decide on quality immediately. A high quality recording can later be compressed to use fewer bits at a lower quality. So you record at the highest bitrates possible, and indicate later which events are worth keeping. As time passes, the least important times recorded will slowly lose bits and quality to make space for new recordings, until they fade away completely. Like memories.

So, do you want your life recorded?

A Steampunk Ghost

The Difference Engine No. 2

Charles Babbage, widely regarded as the father of computing, was a typical hacker.

He never completed the first Difference Engine, instead designing a more advanced Difference Engine, and later he produced several incomplete designs for a yet more advanced Analytic Engine.

None of his computing machines were built within his lifetime. But now, more than 150 years after it was designed, the Difference Engine No. 2 has been built, and run through a calculation. It’s on display at the Computer History Museum.

One man’s vision fulfilled after more than a century. I can’t think of anything comparable, except maybe the moment I first saw a Turing Machine implemented in Conway’s Game of Life.

RIP Arthur C. Clarke

Legendary scifi author, inventor and futurist Arthur C. Clarke died today at his home in Sri Lanka at the age of 90.

May it be filled with stars.

Go - An Ancient and Venerable Game

Go Table

Many people think of chess as the ultimate strategy game. I have to disagree.

I believe that Go is the ultimate strategy game. It has simpler rules, more elegant strategy, and is more than a thousand years older than chess. It is also played by more people worldwide than Western chess.

The rules are simple. Go is played on a 19×19 board. Two players, black and white, take turns placing stones. Black always plays first. Stones are placed on the points where lines intersect. Stones don’t move after being placed, although they are removed if completely surrounded. The goal of the game is to surround as much territory as possible. These are the rules of Go. (For an interactive demonstration of the rules, try The Interactive Way to Go.)

From these simple rules emerge elegant and complex strategies that require intuition as much as logic. This is what makes Go such a difficult and rewarding game to master.

Computationally, Go is far more complex than chess. While Grandmaster Gary Kasparov was beaten by Deep Blue in 1997, most modern Go programs can hardly play better than a talented amateur. Each turn has hundreds of possible moves. A brute force approach fails, simply because it would take too much time to analyze each possible move.

Because Go programs have their limits, many Go players look for their challenges online. There are many online Go servers, but my personal favorite is the Kiseido Go Server (KGS). The atmosphere is friendly, and there is usually someone willing to teach you. KGS also has a subscription service if you desire professional lessons.

We also have a couple of Go clubs in South Africa. The Joburg Go club meets every Tuesday evening at the Mugg & Bean in Rosebank mall. You can get more details at the JoburgGo Yahoo! Group or the SA Go Clubs website.

I hope this was enough to make you curious. For those interested, I have a collection of Go links. (Let me know if you find a link that I don’t already have.)

The sketch for this article was kindly provided by Radka Hanečková (alias Chidori). She is the author of the Empty Triangle, a hilarious webcomic about Go.

Open Source Bounties

Information wants to be free

Economically speaking, it is more efficient to pay for the production of information than to pay for information itself.

This is true for a number of reasons:
1. Information is a public good, and therefore can’t be sold
2. Copyright creates artificial monopolies, which cause market innefficiency
3. Because copyrighted information is non-rival, and the marginal cost of adding another consumer is zero, a zero price is required in order for the good to be used efficiently.

Some would argue that treating information as a public good would cause a market failure, since “no return on investment” could be obtained after the goods were produced. This makes the assumption that information should be sold. It is a public good, and therefore inherently unsalable. What should be sold is the service of producing information.

Open Source Bounties

Bounties seem to be an efficient way to reward the production of information. More and more open source projects are resorting to bounties to compensate programmers. Here are a few examples:

I want my OpenID is sponsoring ten $5000 bounties for Open Source projects to implement OpenID.

Funambol is sponsoring bounties of up to $3000 for coding server components, and $25 for completing cellphone tests.

Mozilla is sponsoring a $500 bounty for each critical security bug found.

Bounty Services

From a programmer’s perspective, it is preferable to find all bounties posted in a single place.

Of particular note is opensourcexperts.com’s Bounty List. Not only do they list bounties, but also calls for tenders and grants. What’s more is that they provide RSS feeds for each category.

By comparison, Bounty County is simply a blog that lists some bounties.

Finally, BountySource attempts to provide a SourceForge-like development platform, but integrates secure escrow services for posting and collecting bounties.

These bounty services are still very new, but they represent an important trend in Open Source development.