Talking with "Google"

Yesterday the Open source program manager at Google, Chris DiBona, came to speak at the University and I had an opportunity to talk with him afterwards at the school’s pre-arranged social hour. It was the most insightful conversation I’ve had with an industry expert since coming to Irvine. Chris possessed an acute awareness of the open source community and the persona of a very self aware and good natured sarcastic asshole, which made him pretty engaging as both a speaker and conversationalist.

I have to say, however, that I feel for people who work at Google. Based on my observations of three conversations involving Googlers, their conversation partners will immediately start speaking to them as if they represent the entire company once the affiliation bubbles to the surface of the conversation. I shudder to imagine what it would be like if every word people spoke to me was filtered through the context of my employer.

That said, I have two nuggets of wisdom gleaned from my forty odd minutes of group conversation with Chris among others. The first is that city of Portland, which has several opportunities for software engineers, is struggling with a heroin problem. The second is that if I want to perform industrial research at a company like Google, Microsoft, or Yahoo!, I do not need a doctorate as many in the academic community suggested.

That pretty much scores on two counts for my girlfriend’s conceptualization of our collective destiny. Well, maybe one and a half taking into account the heroin issue, but that still beats two out of three, which ain't bad.

"Now this really pisses me off to no end!"

I generally prefer working with Microsoft's integrated software development tools rather than trying to sort out the issues that come with combining tools from multiple vendors, but every now and then they do something that makes me want to go David Lo-Pan on thier ass, and I'm talking about the ten foot tall roadblock, not the little old basket case on wheels.

For some ungodly reason, Microsoft included a regular expression engine in the Visual Studio 2005 Find utility that has colons in its syntax, which along with other divergences from standards essentially change the syntax for using regular expressions into something completely different.

This divergent syntax is why the team that developed Visual Studio 2005 shouldn't call it a regular expression engine, because people like me who use the traditional syntax spend hours staring at the screen wondering what the hell they've been doing wrong all day, when the only thing they did wrong was assume regular expressions in Visual Studio follow the standard syntax.

Bastards.

Marcia, Marcia, Marcia!

Marcia. She was the perfect little girl that everyone talked about, much to the detriment of the whiny, middle child syndrome addled Jan. In many respects, all of the user interface widgets in the bright and shiny new world of AJAX have a lot in common with Marcia. They are pretty, competent, and eminently datable.

Let’s hear it for the middle child, though. Let’s stick up for Jan. There has to be more to an AJAX enabled existence than a visual effects style approaching Flash. What about logging? What about monitoring? What about self-adaptive interfaces? These capabilities may not be as attractive as widgets but they deserve our attention, and there is a lot that AJAX enables on this front that people do not seem to be delving into very deeply.

So in conclusion to this pint sized rant, although the majority of me is in awe of the toolkit creation and technology adoption that is taking place in the DHTML community, a small part of me is worried that it is too UI centric (with a couple of notable exceptions, such as Ghost Train or the as of yet undocumented profiler and logger in Dojo).

P.S. Dude, Jan got hot.