I think Transfer Summit is fast becoming my favourite conference. It's like a happy mashup of IWMWs of yore (and even had Brian Kelly attending this year - it was great to catch up with him!) and ApacheCon. The mix of presentations are always representing the cutting-edge research-led thoughts of academia with the experience-led insight and hard commercial focus of industry. It's a powerful combination, even before you add in the terrific venue and brilliant food.
I spoke last year on the economics of innovation in mobile technologies and was lucky enough to be invited back this year to provide an update, perhaps with a little more independence than was possible previously. You can see a short clip about the death of 'open' mobile from my talk, where I highlighted the phenomenal and catastrophic crash in the number of adopted open platforms.
I enjoyed the talks from James Governor, Bertrand Delacretaz, Nick Allott (twice!), Amanda Brock, Andrew Katz, Stephen Walli, Julian Harty and Marco Zehe, and Mark Taylor. It was also fun being on a panel with Stephen, Amanda, and Ross, arguing about Foundations, and Simon Phipps has a write-up on one angle in How Many Foundations Do We Need?
I was frustrated by Tariq Rashid's talk. He seems like one of those who 'gets it', but whenever someone from the government hastily says "we aren't saying we have a preference for open source" I'm annoyed. I can remember hearing this argument back in 2002 with "Open Source as a default exploitation route for government-funded research". There seems to be a perception that when it comes to software, the government is not allowed to have a bias towards openness. Clearly we've spent years with governments expressing a preference for proprietary packages through RFCs and procurement processes, so what is so bad about a government now expressing a preference for a common good?
If Tariq 'gets it' then it was clear from Tony Hey's talk that Tony doesn't get it. There were a few negative comments on twitter, and I found it somewhat disappointing that Tony's well-intentioned explanation of data-centric collaboration was marred by his pigeon-holing of open source as "something you do if there's no commercial software alternative".
Mark O'Neill's talk was enjoyable and light-hearted, but it's absurd hearing someone say inthe same sentence that the government spends £2billion on software but that the government has no money to spend on fixing software procurement. What about that £2billion? As Mark Taylor pointed out, both during O'Neill's talk and also later in his own talk, the money the government are spending on their skunkworks to try and fix the problem of software procurement and proprietary spending amounts to less than a rounding error on their IT budget: in the region of 0.002 percent of that £2billion expenditure. This doesn't look much like a commitment to change. There was also some nefarious sleight of hand from O'Neill, for example when talking about big failed commercial projects without providing comparisons with the number of government IT projects that fail or go catastrophically over budget.
I thought the addition of the open gadget playtime was an inspired one, making the tea breaks fun as well as good networking opportunities.
A sign of a good conference is when there are so many great talks that you come away with a list you wished you could have been in as well as a list you were glad you attended. I was especially disappointed to have missed Maha Shaikh, Mark Power, Eben Upton, Carlo Daffara, Andrew Back and Paul Downey, and Julian Harty. Perhaps next year!