Last day of the trip. Going for breakfast to “Raw Power Cafe” on Vulcan street off Queen Street in Auckland. Nice place, good food, freshly prepared juice. Hail storms in the morning, but the weather clears up and pretty nice and sunny later throughout the day.
10:30 Stuart and I meet with Prof. James Goodman
Jim is a bit late for the meeting – the ferry from his home in Davenport was late and he got stuck a bit in a hailstorm. Hi starts by explaining about hardware architecture and that synchronisation mechanisms lie at the bottom of the problem. Jim explains briefly the various TM models and their relationship to the hardware and cache eviction and consistency models. The programmer tends to think about the world as if it stops for a while when the sequence of instructions is being executed. There are traditionally two approaches to this problem: locks/critical section, and transactions. Transactions can be hardware-based and software-based. The two streams of research are merging now. There are many limiting factors on the hardware side, one being the problem with the length of transaction (limited by the CPU L1/L2 cache capacity). Software does not have this limitations, but the performance penalty is high. (Results and speculations …)
The discussion goes on about the general scalable system programming and education in Computer Science in general. The only compiler person retired recently from the AUT, student numbers drop, and it is generally difficult to keep up withthe core CS courses. 1 PhD student currently working on TM models. References to Mark Moir (Wellington, stays few summer months a year in NZ).
Jim says: “Multicore [hardware] architectures are being pushed onto our throats whether we like it or not, and it would be a shame not to take the advantage of it. The revolution is happening.”
Inviting Jim for the PDCAT 2008 conference PDCAT.
12:30 Lunch-time meeting with Telecom R&D people
Discussing Swicki – and other community and social networking search engines. Discussing document search and reference to Waikato: Richard Nielson, prof, data mining, search, context based, commercialised products. TODO watch Nomads documentary. Discussion follows up on smart cards, Sun-Ray, Secondary schools, grid and cloud computing.
Important: follow on IPv6, with Brett Tefler.
Next generartion services, mobile services, communities, etc. John talks about the idea of facilitating 10-20 ppl meetings in virtual space, similar to video conferencing for brainstorming and working on something. Talking with John Eyles and Brett Tefler about the idea – need to email about Wonderland and possibility of getting 10-20 people for a short period of time to dicsus a really focused discussion about something.
16:30 Cofee-time meeting with iJump
Meeting with Simon and Marie. Nice couple. Currently 2 people, 8-9 writers. On the market for 9 months. Discussion about new media types, influence of new social networking applications and technologies on the way people make business and expose themselves to customers, how to gather feedback and how to manage the information. Simon and Marie were interested in the Otago social networking group and the activities that are going on. Also, they are doing some consulting about second life and Web3D so they are interested in NVWG initiatives.