Sunday, December 05, 2010

YOW Melbourne roundup
Recap of YOW 2010 in Melbourne.

I will write about 2 things here, Conference Organisation and Talks

Conference Organisation:
Organisation of the conference was pretty decent except for the food. Food was shit as (Oz slang learnt from my daughter). They served up same shit all the days. I thought may be for vegetarians it was like that and asked couple of my carnivorous friends, they could not wait for the conference to finish so that they can have decent lunch. I always tell my daughter "Human beings will kill themselves because of their greed" (I have googled, That saying is mine). They charge 3000$ per person for the workshops and conference. Even if they spend 50$/day/person, it is 200$/person. I fail to understand what stops them from doing that!!!

Talks:
Exploring NoSQL by Erick Meijer:
He talked about this NoSQL movement and renamed it as coSQL. He proved using category theory that NoSQL databases are a dual of RDBMS. Awesome flow in the talk. I dropped out of a boring session and had coffee with him, very nice guy.
Tomato Meter:Do not Miss (if you are going to Brisbane one)

Integration tests are a scam by Rainsberger:
Very nicely presented. Showed how Integration tests are evil as they are slow and brittle. Showed how to solve the issue by writing contract tests using mocks.
Tomato Meter: Attend

Rails in large:
My Expectation:
Most people doubt ruby on rails as a sustainable choice for large applications and the speaker was to prove that wrong as they had built a large application using ruby on rails.
What it was:
I knew the speaker personally and was aware of his speaking skills. He presents really well but the content was not up to the mark. It felt like he was asked to use the name of the company he works for, a million times. He started of talking about how people doubt ruby on rails, but later it was about how much time it took for their tests to run in their builds, project manager finding a service to deliver coffee so that developers didn't have to waste time, company pairs on everything blah blah blah. Not once it was mentioned, what was the scale of the application that they built. Number of concurrent users, data size something... come on very disappointing.
Tomato meter: OK (if you want to learn how to make your tests fast in ruby/rails land)

Emergent design:
My Expectation:
New spin on BDUF vs Evolutionary design.
What it was:
How writing tests will help in better design (like ron jeffries solved sudoku writing tests). Used finding perfect-numbers as an example to show how writing tests will help in design. Code was written in java and later in groovy (hahaha). I was wondering wish he chose as functional language, it would be list comprehension and a reduce call.
Tomato meter: Not worth attending

Worst
The State of the Art .NET, 12 months of things to learn:
My Expectation:
To know what's happening in Microsoft world.
What it was:
Worst talk I have attended in my life. Slides and slides of links and names to follow on twitter, that summarises the talk. In the end few slides of tools and technologies.
Tomato meter: Should be removed from the conference

Things I walked out after 5 minutes:
100K transactions with 1ms latency, Restful Applications, Building distributed applications google way, Erjang etc..

Problem with these talks were that, they did not offer anything more than what you cant find by googling or reading documentation. Some of the speakers were so fuckin boring, they read out the slides which had xml in them. I understand they are not paid to present in these conferences, but nobody forced you to talk either (I hope).

Conclusion: Catchup with speakers in the hallway and quiz them about things you are interested, that will be the best take away.