Velocity Conference 2011…Morning Keynote Session

So either the conference network is overloaded or the VPN is falling on its knees right now. Yikes…

Theo Schlossnagle: A Career How and Why

A career is a pursuit; a willingness to mature. To be an expert…you have to be a craftsman…through experience, diligence and hard work. You have to be able to consume knowledge. You have to be disciplined. Learn from and share with your peers. Be patient as experience takes time and mistakes.

What I really like about this session is that he’s really keying into a shared philosophy that I have about Performance Engineering. You need to have pride in your work and pride with the people who you are serving. Subsequently it means that everyone needs operational mentality. You need to build systems that operate and work for your users.

Security is not a feature…it’s a state of mind…a state of being…Operations is no different. He kind of jumped all over the place. His key point is that software developers and operations engineers really need to be 1 and 1 the same (simpatico).

Rule #1: What You Build Will Break (So Think Operationally). Put more development into ops. Put more ops into dev. It’s interchangeable.

Unfortunately there is no rule #2 or #3…

Doug Crockford from Yahoo

Java – Scheme – Self = JavaScript

ECMAScript programs are expected to start execution instantly, which means there really isn’t time to optimize. Then jumped into calling Java the most flawed/failed programming language in the history. He’s really calling that out from the perspective of a UI framework. He did suggest that it was fine for server side code.

It looks like his presentation is about how JavaScript can be faster…faster. He’s finally getting to the point about how the JS engines are taking on the role of optimizing the script output by developers. He jumped into benchmarks (called them lies…damn lies…). Suggested that benchmarks can be used, but are not representative of real programs. Referenced JSMeter from last year about how none of the benchmarks had real meaning. The problem is that the engine makers are tuning to the benchmarks. This seems like a conflict of interest. THe benchmarks aren’t viable, but the engine makers are optimizing based on the benchmark data.

The real problem with the browser from a performance perspective, but it’s the DOM. Browsers spend most of their time in layout, painting, marshaling, etc…Most developers are spending their time trying to optimize their JS. The problem is they are not focusing their time correctly.

Badly behaving developers Benchmarks –> Drive Enginee Performance –> Drive Coding Practices.

He finally made his point…he’s a developer who uses good design patterns rather than bad design patters. He developed JSLint. Nope…looks like he keeps coming back to benchmarks. So what are you suggesting? OK…finally got to it. He developed a benchmark based on JSLint, measuring how long it takes to analyze jslint.js. It demonstrates real application behaviors that are not exercised in the other benchmarks, essentially prototypal and functional patterns.

Browser Time
Chrome 10 2.8
IE 9 1.15
opera 11 1.10
Safari 5 .98
Firefox 4 .95
IE 10 .56
Chrome 13 canary .54

Look at Your Data by John Rauser

Note To Self: Need to get a hold of his presentation. It was hilarious and full of great insight. Need to call-out how important histograms are over time-series plots. No monitoring software automates histograms. Yeah Patrick Kee for solving this problem in Galileo!!!

Imagine an old school dashboard (ops center…star treck…) or even something as simple as a time series chart. It was a cool joke about how dashboards are cool, but are they really understood? Do we dig into the data?

What do we lose when we look at data? He’s totally making fun of summary statistics. They are useful, but very lossy compression algorithms. How can we look more deeply at our data? Histogram is one key data. He’s provoking questions…how can we make our sites faster. Histograms are great, but so too are scatterplots. Don’t just look at a single summary statistic. It’s evil!!!

Key Point: Get on your hands and knees to look at your data.

Last Presentation by Keynote

20 Billion App Downloads…talking more about mobile apps and not just mobile browsers. What are the challenges of testing and monitoring native apps. Need a real network and a real phone. The tool KeyNote is pitching is Mobile Device Perspective 5.0. It’s a remote interactive testing, script/replay as well as the automated monitoring tool. Native applications took off again back in 2006/2007. Yes a lot of mobile applications are using browser interfaces.

Definitely should suggest to the Mobile team to investigate this application. It’s pretty impressive. It’s kind of a functional tool, while at the same time a user experience monitoring tool.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s