Observations

Some notes I took at BarcampSydney, March 2007

Digital storytelling: branding yourself or your product

• Gavin Heaton update6/2/07: He has posted his notes here
• good stories are ones people can relate to
• a brand sits there and draws the ‘attention consumer’
• It’s about building a connection – +++ / —-
• What’s your promise, what are you delivering?
• Showed above video; lost on audience because screen was too small, which was shame. Said there is another version where you can add content to the video in real time

Discussion

• online is about reinventing yourself –> like joining a new school –> second life, world of warcraft
• you need to build trust –> reputation

—-

Bootstrapping a start-up

• I missed the chunk of this talk unfortunately, but one thing I picked up was that a brand is key for a web app – and building a brand is expensive

—-

How to scale – mapreduce

• Joel Pubar
• Powerpoint file from Presentation (obtained from Joel’s blog posting): how to scale – map reduce
• In 2003, we hit a wall in processor speed
• The problem is heating –> processors can’t get faster because of this issue –> 100 watts is the thermal limit
• Joel discussed how was building a .net implementation of Google’s map reduce
• random fact: 1.6 million blog posts/day
• map reduce (at the moment) isn’t used for on the fly stuff
• Biggest problem is ’structuring’ the web
  • data is currently unstructured on the web –> making ’sense’ of the internet
• Spam filtering, which is classic AI stuff (and therefore classic pattern recognition) is perfect for mapreduce

—-
Privacy and advertising

• Phil Morle – former CTO of Kazaa
• believes it is dangerous to start a business without a revenue model –> because then you rely on advertising
• Introduced APML (Attention Profiling Markup Language) which is effectively about storing your personal information locally on your machine
• problem with digital age, is that our information is now centralised
• someone suggested the $100 laptop programme was going to see the fall of online advertising –> all these poor African kids clicking on ads
  • point was negated because of the nature of targeted advertising – hence why this discussion was so important, because companies are actively profiling us
• Issue is about control of your data; anonymity
• very good discussion!!

—-

HTML5 v xHTML2

• none of the major browsers are suporting xHTML 2.0
• HTML5 supports xHTML 1.0
• 5.0 has better error handling
• backwards compatibility is a big issue
  • javascript can fix errors to ensure this occurs

• html5 in used in opera, which dominates the mobile market –> mobile usage may drive general internet usage later
  • Opera is counting on this as a business strategy for market penetration

—-

Funding

• Anne Bartlett made a big point about how much state funding there is, and how willing the government wants to help (“it’s almost scary!”)

—-

Replicating web stuff in the enterprise / search

• Not sure if that was what the topic was, because came in late
• Interesting discussion about the current state of search – basically, we are not there yet because the best we’ve got (Google) is no way near perfect yet
• Discussion on the state of enterprise search, which is terrible
• Page rank principle can’t work within enterprises
• Filtering of search was discussed –> examples such as drop down menus for hotel sites (rather than free text search), although it was commented, that they found this system actually restricted a search too much, which led the user to think the site sucked

—

Agile development

• 20% of features are always used
• 64% are rarely or never used (45% never)
• above facts from a 2002 study – Jim Johnson

—-

Planet-earth.org

• New Australia led open source initiative, that’s been around pre-Google maps
• $8 million funding from the national science something organisation
• reason for delay in Google’s roll-out of maps in Australia, is because Australia is one of the only countries that has such expense maps
• we pay taxes to pay for these maps, but businesses have trouble licensing them –> pay 1000 times more than overseas countries; other countries can get them for nearly free

How to start up your own business

• It takes 5 year to become an expert; 10 years to be an innovator; 15 years to be a genius
• Ari Paparo (Blink) – > got slayed by del.icio.us –> really good account of what happened
• Mike had a start-up before Atlassian, circa 1999, that was exactly like del.icio.us
  • difference between then and now is sharing. the default on all web apps is public, which is why sharing is so prominent now. Before, default was private

• your business pitch should be about value, team, and business –> you go to VC’s with a team, not just an idea
• avoid VC money – it makes you lazy and complacent (Marty); Mike throughout the talk resonated his dislike of VC money, and pride that Atlassian is “revenue funded…not venture capital funded”
• long discussion by Mike and Marty (Tangler) about company culture; the people you hire and fire
• Flickr, Del.icio.us, MyBlogLog, and YouTube are “star” startups –> read my blog entry on this

Other accounts of the day

• Christy, who initiated BarCamp Sydney, has a good write-up
• Tim, whom I met during the beer presentation afterwards and whose ex I used to work with (small world!),  has an overview
• Videos