March 2013



Google's Eric Schmidt recently said that they would not be merging Android and Chrome OS together. But they are both open source, so there's nothing stopping us from having a crack a it. Let's have a go, and see what we can do... This could be quite a nice combination for a touchscreen device such as the new Chromebook Pixel, as many have speculated. Before we can begin, we will need a ChromeOS device. Then we will need to take the red pill (i.e. flip the developer mode switch), as discussed in my previous post Hacking the Chromebook for Fun and Profit.




On March 20th 2013, we held a formal launch event for HPC Midlands. It was well attended, with some 80 delegates from both research and industry. Leading software vendors were also well represented, with CD-adapco, MSC Software and ANSYS all exhibiting. For those who could not make it, or would like to take another look at the speakers' slides, we have embedded them below. For more information about HPC Midlands and the services we can provide, please see the final slide deck for contact details and next steps, fill out our contact form, or call us on 01509 223110.





















HPC Midlands is supercomputing on demand, for research and industry, with flexible licensing for commercial software. The service is brought to you by a consortium of Loughborough University and the University of Leicester, supported by funding from EPSRC/BIS under the UK e-Infrastructure Programme. We've just held our launch event, at Loughborough's award winning Design School, and this post is just to recap a few key themes from the event.

[Picture: Lionel Mazzella from E.ON presenting at the HPC Midlands launch event]

At the launch event we brought together speakers from both Universities, JANET(UK) and industrial collaborators from Rolls Royce, Tata Steel and E.ON. I was also very pleased to have delegates from large corporates like Jaguar Land Rover to small to medium sized businesses like BAST Inc, one of our Innovation Centre incubator tenants. It was also particularly gratifying to have support from key independent software vendors ANSYS, CD-adapco and MSC Software, who all exhibited at the event - and our suppliers Bull Information Systems, who are running the HPC Midlands system as a managed service for us.

We'll be sharing all the day's slides via our website in due course, but I wanted to take the opportunity to plug my own session, which was all about next steps in engaging with us. You can view a copy of the slides embedded below:


Engaging with HPC Midlands - Next Steps from Martin Hamilton

If you are using HPC in industry, we would love to talk to you about using HPC Midlands to scale your jobs up, and expedite urgent work. We also have access to a network of HPC experts that are attuned to industrial collaboration and very interested in working with you on practically focussed projects such as product design and simulation.

In the first instance we would encourage you to email Peter Strutton, our Business Development Manager, or call him on 01509 223110.





















There's been a lot of breathless talk around the launch of Google Glass. Self righteous nerds the world over are now poised to attack the first person they see wearing Glass in a public space. But what if that person were blind? What if Glass was actually the best piece of assistive technology yet for blind and partially sighted people? Let's spend a little time exploring this...

[Picture credit: Sergey Brin wearing Google Glass on the New York Subway, via Twitter user Noah Zerkin]

Regular readers may recall a post from a couple of years ago where I demonstrated how you could use the Android scripting layer to drive the camera, speech recognition and crucially interact with product data online, with the result being a trivial python script that turns your phone into an audible bar code scanner. Cue witty aside here about the Google Product Search API being closed down as part of the latest spring clean at the Googleplex...


Jam and Jelly

But this is all a bit fiddly, particularly if you are trying to navigate your way independently through the local supermarket while using a white stick or being led by a guide dog. Now picture that same supermarket scene with Glass: "OK Glass, what's in the cans in front of me?" Or perhaps... "am I in the jam or petfood section?" (jam == jelly, for my American readers ;-)

We've also heard a lot of negative stuff about facial recognition technology. Now picture it from a blind person's perspective - "Hey Alice, I can see your friend Bob". Let's imagine that we went one step further and added some opt-in location awareness a la Google Latitude, then your glasses could even help you to meet up with a blind friend. "Hey Bob, Alice is two blocks away from you. There's a locally run independent artisanal coffee shop coming up in 100 metres where we could meet - shall I send him the location and use their API to order us both a tall skinny mocaccino?"


A Grand Day Out

For some more practical examples of how Glass could help blind and partially sighted people with independent living, let's imagine that I am blind and aiming to catch a train to London. How could Glass help me out?
  • OK, Glass: When is the next train to London?
  • OK, Glass: Book me a ticket on the 1234 train to London and reserve me a seat
  • OK, Glass: Which bus route is best to get me from here to the train station?
  • OK, Glass: Directions to the nearest eastbound bus stop on the Number 7 route
  • OK, Glass: Alert me when you see the Number 7 bus coming, and direct me to the doors when it stops
  • OK, Glass: Direct me to the ticket collection kiosk
  • OK, Glass: Tell me what you see (The ticket collection kiosks in the UK are pretty much inaccessible, so in reality you would probably find yourself going to the ticket office next. If it hadn't been downsized...)
  • OK, Glass: Am I holding my train ticket the right way up for the barriers?
  • OK, Glass: Direct me to the ticket barriers
  • OK, Glass: Which platform is my train leaving from, and is it on time?
  • OK, Glass: Direct me to platform 2, and alert me when my train has pulled in and stopped
  • OK, Glass: Direct me to the nearest train door
  • OK, Glass: Tell me when I reach my reserved seat on the train
If this all seems a little fanciful, here's a more homely example:
  • OK, Glass: Tell me when the kettle is directly above the coffee cup, and alert me when the cup is three quarters full

The Android Connection


Android, the open source Linux distribution running on 750 million smartphones and tablets, is also the technology underpinning Glass. It will be interesting to see how much of the Glass software Google make open source - it's easy to picture other Glass type products appearing that use Android but aren't aligned to the Google ecosystem, a business model already proven by Amazon with the Kindle Fire, and numerous Android devices in mainland China.

What's the big deal about Android? Well, bear in mind that much of I have described above is already possible with existing apps and APIs on Android and those achingly fashionable open data feeds - it's just a bit impractical unless you strap your phone to your head, or something like that.

The science fiction part (for now - but perhaps you can change that, dear reader!) is the bit where Glass "directs" you in real time by processing the camera imagery, e.g. to the exact bit of the train where the door can be found, or your reserved seat. But if you think about it, optical character recognition of images has already been demo'd to great effect by Google Googles and there are some interesting pointers from the lesser known but seriously awesome The vOICe for Android.


Closing Thoughts

To get a feel for potential revolutionary applications for Google Glass, just try walking around for a while with your eyes closed! See how I managed to get this far without mentioning collision avoidance?


Now listen to the Google Glass promo video, embedded above, with your eyes closed. OK, Glass: Take a picture, and tell me what you see...






















This week I spent an enjoyable and stimulating couple of days in Birmingham at the CETIS 2013 conference. This was tempered by the knowledge that key national services like UKOLN and CETIS itself were looking ahead to new challenges, a euphemism for closing down due to withdrawl of funding. In this post I'll reflect both on what I learned from the conference, and on the impact and timing of these changes.

[This is going to be quite intense, so here's a little game for you to lighten the mood: If you are visiting Aston University, see if you can find the building and window where the inspiring quote from Einstein shown above appears!  Also, here is a YouTube link to funny cat videos in case you need to take a break to calm down from time to time.]

I have long been fascinated by JISC, which has delivered a fascinating mix of full production services (JANET being probably the most instantly recognisable), best practice advice created by community practitioners coming together, and what can only be described as "throw enough mud at the wall and some of it will stick" projects with hugely ambitious agendas. You may not recall individual projects, but be assured that when you download Open Source software, Open Access publications from institutional repositories, mash up Open Data, and share Open Educational Resources in MOOCs - JISC was there first. In many cases, JISC was farsighted enough to forsee requirements in the research and education sector that have subsequently turned into significant businesses in themselves.

But we are entering a new era, necessitated by funding reductions, changing student demographics and frankly an unwillingness to see "R&D" type activities (of which a large proportion can be expected to fail) facilitated through top sliced central funding. The most immediately obvious change here has been that we have swapped JISC for Jisc, saving a few bucks by making the Shift key last a little longer. Behind the scenes, a lot of people who have been working for JISC on its various centres and services have been having meetings with their local HR departments about redundancy and redeployment, and some have been transferred over to the new Jisc legal entity via TUPE.

These goings on are, sadly, mirroring many conversations taking place within institutions right now, as the realisation slowly dawns that the rolling back of UK Higher Education's post-92 expansion may well be permanent. The disruptive influence of the Internet (Clay Shirky's grandstanding notwithstanding) is also now starting to make its presence felt, with the development of for-credit MOOCs and partnerships with firms offering test centre facilities such as Pearson. In the UK, FutureLearn in particular now looms large over the sector - but will it be saviour or executioner? In any case there is little or no time for discussion and compromise if FutureLearn is to compete effectively with Udacity, Coursera and edX.

And here is the crowning irony - FutureLearn should have been the primary topic of CETIS 2013, instead it was the elephant in the room. So much of what so many of the participants have been doing leads inexorably to something like FutureLearn. Instead, we are breaking up the very services and facilities that are best placed to advise the sector on its transition to being "open by default", and in the urgency to get to market there is a very real danger that FutureLearn will be a pale me-too copy of the US MOOCs. Hence the JISC Diaspora, as we see former staffers with the relevant domain expertise being snapped up not just by canny UK institutions (greetings to Amber Thomas at Warwick ;-), but also by "foreigners" (hello to Nicole Harris of TERENA!) and with much more to come - notably as CETIS, UKOLN and OSS Watch are dismantled.

You might well say "fair enough, but what would you do in the same position?" I think the economic argument is actually a little more complex than it might at first appear - sometimes attack is the best form of defence, and there is undoubtedly a case to be made for strategic investments to improve our information literacy, get more people coding, increase the visibility and uptake of UK research, and expose more people to key technologies like High Performance Computing. But should Jisc try to do all these things and more? Probably not. And is it possible or desirable to maintain the status quo? Probably not.

However, what I am personally deeply sad about is the bridge burning aspect of these changes - once the key people from an organization like a CETIS or a UKOLN or an OSS Watch are gone, it's not a trivial exercise to recreate it. I hope that the members of the JISC Diaspora will still find ways to come together and collaborate, and that the moonshot projects will continue one way or another.  As Martyn Harrow says, JISC is the envy of the world - I hope we will be able to look back in years to come and say the same of Jisc.

























HPC Midlands is a new initiative from Loughborough University and the University of Leicester, co-funded by EPSRC. Use our 3,000 core supercomputer to accelerate your innovation - find out more at http://hpc-midlands.ac.uk, come to our launch event on 20th March 2013, or read our impact report to EPSRC / BIS, embedded below.






















This is the final report from the Open Course Data project at Loughborough University. We have been funded by Jisc to develop an XCRI-CAP machine readable version of our online prospectus. In parallel we wanted to explore improving the way that the prospectus is developed. The reader can find out more about the project at http://open.lboro.ac.uk.


HPC Midlands, a new supercomputing facility based at Loughborough University’s Science and Enterprise Park, provides state-of-the-art e-infrastructure for research and industry.