Future plans, thoughts on OLPC, and clearing up the Canonical misunderstanding
TS: Are you working on new models of the Pi or just focusing on the two that are already out?
I think it's very important to have a stable platform and now that we've picked our platform we're sticking with it – we're not going to make significant changes. Where we are doing tweaks is in manufacturability. Making it easy to manufacture, and reducing the defect rate. These are almost invisible tweaks and is a very common process with new technology: you start out with a version and then you do very minor revisions to it in order to improve the DFM (design for manufacturability). As we build tens of thousands of units we learn what is and what isn't working at the factory, and we just do these little tweaks just to improve them.
But actually doing functionality changes or launching a new product, we don't want to do that because we would then upset those who've already committed to the Raspberry Pi. I don't think that would really be fair.
Canonical made the decision not to support AMRv6 anymore. That's a commercial decision. It was an inconvenient decision for us because we did all of our development using Ubuntu 9.04.
TS: Out-of-the-box the Raspberry Pi comes with no software whatsoever so users currently have to load the OS on an SD card themselves. Taking this into account, when you think about intended audience, who comes to mind? Do you have any plans to improve this out-of-the-box experience?
The current audience is a very technologically savvy one. You look at some of the things people post on our forums and you'll find that some incredibly technically capable users from the Linux community using the Raspberry Pi at the moment. I think that for the current audience what we have is very suitable.
A lot of less technical users are going to want to buy an SD card with a pre-installed operating system, they're not going to want that step of doing it themselves. We are working with our partners to provide pre-installed SD cards with the Pi kit for people that are less cost sensitive, but since the focus is on cost we don't want to add another few dollars to the price if we don't have to.
But for the education market it's obviously going to need to come with an SD card. I think a lot of people are going to benefit from having a pre-installed card, and of course once you have that we can preload a load of educational software onto it so they'll get a much smoother out of the box experience.
TS: We know the Raspberry Pi can't run Ubuntu because the OS doesn't support the ARM-based processor you are using. But we were surprised to hear that not only Canonical wasn't interested in supporting your chip, they were quite vocal about trying to stop you from even mentioning Ubuntu.
We actually have a good relationship with Canonical. What happened is we had an interaction with a senior community member, not Canonical itself, maybe he didn't phrase what he wanted to say very well and probably I shouldn't have reacted to that.
Canonical made the decision not to support AMRv6 anymore. That's a commercial decision – probably a sensible and good commercial decision if you look at what Canonical is trying to do with Ubuntu.
It was an inconvenient decision for us because we did all of our development using Ubuntu 9.04. Until August last year the Pi was all an Ubuntu platform, so when we first announced the machine, we just assumed we were going to use Ubuntu and were very happy with it. Then all of a sudden I knew Ubuntu would stop working for us and we had one person tell us "please stop saying you're going to use Ubuntu." Probably because we were a little naive or inexperienced we assumed that represented the views of Canonical.
So we are very hopeful. We have no plans to make a new Raspberry Pi yet, but if we ever do a new Raspberry Pi that had a v7 core in it, I think we probably would go back and try and get Ubuntu. It is one of the more polished and professional desktop environments. We have a lot of respect for them.
TS: Have you been approached by any other big name IT companies to collaborate with you?
IBM did some really nice work with the Raspberry Pi at their Impact conference in Las Vegas recently. We also have a number of boards out with major IT companies – nothing specific to announce – just people who are interested in using it to do a range of stuff from web development to embedded applications.
The level of interest has been great. I think people do appreciate that this is a potentially revolutionary development, not necessarily the Raspberry Pi but the advent of cheap commodity ARM-based hardware. People are interested to understand what that can mean for their business.
TS: We noticed that you recently posted schematics for the Model B. Do you plan to open source these designs or in some way encourage other companies, or perhaps governments on developing countries, to imitate the schematic of the Pi and build their own inexpensive computers?
We don't have anything announced at the moment. I guess the limitation is some of the chips that are in the Pi are not currently available in general distribution. What this means is that even if we did release the hardware designs it wouldn't matter much since people wouldn't be able to get the chips to build the device.
We are thinking about ways to deal with this and hope to make some progress maybe in the next few months.
TS: We've seen the gertboard, which works as an add-in card to control and manipulate physical devices like motors or sensors. What other application have you seen working with the Pi already, and looking forward, what kind of tasks do you foresee the Pi performing?
For the moment most of the stuff people have done with the Pi has been software stuff, because it's easy and quick to do. You have people using it as a media center, serving web pages, running graphics applications, and so on. Those are the things that I've seen so far.
But regarding possible uses in the future we've heard from a lot of people wanting to use it for robotics, maybe alongside the Arduino platform. There's a thing called the Arduino Uno, which is an Arduino with a USB interface, and we've seen people connecting the Raspberry Pi to it to drive hardware. Arduino already has a well developed set of software tools. I think we are going to see a lot of this.
TS: What are your thoughts on other similar initiatives like One-Laptop-Per Child? They struggled to meet their target prices and haven't seen the level of adoption they certainly would have hoped for. Where do you think they failed and you were able to succeed?
OLPC was very ambitious. It was trying to provide solutions to a community of people who we could never hope to provide a solution for. Their systems were meant for places often with very limited electrical power and people who didn't already have a monitor to connect the computer to.
Because of this it had to be a laptop. They had to have a battery --- they came up with solutions like a hand crack to charge them and stuff --- and the machines had to be very very robust so the display and hinge wouldn't break. Plus, at the time there weren't necessarily the powerful ARM based processors that are available today, so they had to go with x86 hardware which is more expensive.
All of this tends to drive up cost but what it also does is increase the minimum economic quantity of the devices – that is the smallest number of devices that you can make economically.
The netbook market exists because of OLPC. Sometimes a project can't meet its goals but it can still transform the world, and I would suggest OLPC probably did that.
I can make 10,000 Raspberry Pi's for the same unit price I can make 1 million Raspberry Pis. There's a curve, of course, but that curve for the Raspberry Pi flattens very quickly. You very rapidly reach a point of diminishing returns. In fact, I think the minimum economic quantity for the Raspberry Pi is probably in the 2,000 unit region. This is because we have a very simple piece of hardware. It's got no moving parts, no display, not even a case, just a circuit board with some components. What this means is that we were able to start small selling to individuals even a single Raspberry Pi and grow from there.
I'm only seeing this from the outside but I think the main difficulty that OLPC encountered is that their minimum economic quantity was very high, and so they had to and get these big government orders in order to justify building enough units so they could hit their cost targets. I don't know what the minimum economic quantity for OLPC was but it was probably hundreds of thousands, maybe millions, and they had trouble to achieve that.
I don't think it's necessarily about us being smarter than the OLPC guys. I think it's about trying to do something that's fundamentally simpler. I mean we provide devices for people who already have electricity and screens and stuff. They were trying to do something much more complicated.
TS: The concept was pretty bold back then but somehow it was actually Asus who popularized the form factor and then it became a big category that every for-profit company wanted a piece of.
That's the thing. We should give OLPC credit. They created the netbook market. Maybe netbooks are now being replaced by tablets but the fact is that netbook market exists because of OLPC. Sometimes a project cannot meet its goals but it can still transform the world, and I would suggest OLPC probably did that.
TS: To close our interview we'd like to ask a question that will become part of a feature we're starting at TechSpot with interviewees. Besides a Raspberry Pi, we presume, what else is on your desk? Tell us about the platforms you rely on day in and day out for desktop computing, mobile and other any other gadgets.
I'm a ThinkPad guy. Day to day I use a ThinkPad X40. I used to work for IBM and I had my first ThinkPad in 1996 – it had a 486SX-25 processor I think. So I've always been a ThinkPad guy.
I have one of those and a desktop machine for gaming, which was a very high-spec machine when I got it but it's not really anymore. I think it's an Asus P5B Deluxe with an E6400 processor in it but quite a nice modern graphics card. I do most of my gaming on my PS3 these days, though.
A lot of the Raspberry Pi success is down to my wife Liz. She has done all of our marketing communications for the last year. She uses a very high-end MacBook Pro and she keeps trying to tempt me across. She keeps saying "OS X is based on UNIX, just give up on Windows and come across use UNIX like real people!". I don't know, maybe I'll get a Macbook Air. It's awfully tempting but I'm not ready yet.
For mobile I use a corporate BlackBerry. I'm a really primitive mobile user. I had a Lumia 800 for a while and a Nokia N900 – they look identical from the outside but the N900 is probably a better phone, I do like Meego. But mostly I just use my BlackBerry.
Make sure you check out TechSpot's review and initial how-to guide of the Raspberry Pi. This interview is the first on a series of articles we'll be publishing in the coming months, where we talk to some of the tech industry's leaders and visionaries. We are calling this special feature THINK IT.