Prototyping Weeknotes #38 (29/10/10)
An unofficial iteration planning session for Autumnwatch kicks off Monday. Duncan and Chris B are going to build a clickable wireframe of the end-to-end app which will demonstrate how it's all going to work and be joined up; we'll then iterate through the designs when they become available. Theo and Jo will be sorting the user testing stuff and some of the promo material. Tris meets up with some Vision people to explore some ideas about storytelling on the web, and some possible future projects. Paul, George and I start the week as we mean to go on; with a series of telco's.
The firehose work is starting to come together. We've successfully connected to the Firehose and have got the upload to Amazon S3 working. We're storing data at a rate of approximately 180GB/day and our tests show we can scale up to 20,000 messages/sec. The next step is to do some MapReduce queries against the data set. Kat's been developing ideas further for an application that uses lots of real-time data as well as running an excellent workshop with the team that threw up lots of new ideas and lines of thought.
George reveals some new projects that have come his way at our mid-week planning meeting. There's a workshop in the afternoon with some colleagues from A&M along with Kat and Jo; it's good to have production and UX resource on what has been a purely technical project so far. Kat is going to pick up the final stage the work to get an implementation of RadioVIS up and running for the BBC's national radio networks.
Thursday is team show and tells. Sam shows what he's done with the Mozilla audio API; Mark almost gives us a demo of AudioSync; Chris B demo's the second screen companion; Tris shows off his new Kindle and George talks us through a 113 page response to an EC Call - is sharing always necessary? We've been worrying about scaling for the upcoming Autumn watch testing but Chris B finally makes a bit of a breakthrough. He's managed to get some load testing running against our qjabberd XMPP BOSH server using Apache Jmeter. The initial results are promising, we just need to ascertain the upper limit and what the impact is on the client when things start to fail.
Vino from R&D drops in on Friday to see some of the second screen work we've been doing. Tris is doing some research for next week's trip to the W3C . Tony has spent the week tidying up his wiimote demo and trialling it at the Digital Radio Week in Belfast; it has been getting some very positive feedback with responses focusing on enjoying the novel interface of the wiimote & the fact it's video, rather than cgi. George is returning from his whirlwind visit to Florence where he was giving a keynote at the ACM conference. I finish up my final telco for the week before writing these, then home.