Saturday, September 02, 2006

Head in the Computer

I have spent most of the last two or three weeks in an introspective mode. Not quite that. I was evaluating a lot of technology as part of a feasibility study and in part to catch up with the state of the art in Python and web development. Two weeks ago I worked on a DAS2 sprint. It started at 6pm local time with a phone conference call and I would work on it until late, because the sprint proper took place in California.

In addition to the spec proper I looked at two technologies which may affect DAS2 system. The first is content negotation, sometimes written conneg for short. I like the idea that clients can include an ordered list of formats, languages and encodings in the request. My Safari browser knows that I can read English and some Spanish and Swedish so when it includes that in the request, in the extremely rare chance that 1) the server understands the Accept-Language header and 2) a document isn't available in English but is in Swedish. Which is effectively never. The other things I looked at was Sleepycat's XML database.

But my technical writings page is at http://www.dalkescientific.com/writings/diary/ and I won't go into details here, much as I would like. I've made several updates there over the last week as I look into various Python technologies. I had heard of most of them before and had a general feel for how they were supposed to work. What I was mostly doing was getting a hands-on feel for how to use them and if they were worthwhile for the way that I develop software.

I like to think I'm quite good at that. It took a lot of intense single-minded effort though and there were a few times where I looked up to realize it was 5am. That's been going on for a few weeks and includes today when I did some work from 11:30am and seemingly a few minutes later it was 5pm and I needed to get ready for James' birthday parties.

I haven't been a complete homebody. I went to the Lehvaslaiho's last weekend. I write that with a smile on my face because I always say Heikki's (or Heikki's and Minna's) place as while I've heard that Finnish pronounciation is very phonetic using Swedish phonology, even if I the surname correctly no one would know who I'm talking about.

Today was James's birthday party. It was at an all-you-can-eat sushi place in Panorama. (Which, yes, does have a good view of the city and Table Mountain.) I wasn't much impressed with the fish or the service. Being in a highly introspective mode I talked shop. James is one of the students I taught last year at the NBN. He's at Stellenbosch now. He and his girlfriend (also one of my students) moved here from Durban so they mostly know bioinformatics students. Hence people I could talk shop with.

After that was salsa dancing at a bar near Cavendish mall, in Claremont. Not a big crowd so an open floor. It was Kim's birthday celebration, at least for her dancer friends. Her birthday is Monday. Lara and Lenine showed up. Of the people I dance with here they top the list.

Question for my dancing friends who read this: how well does a person's dance style reflect one's personality? What would be a good test for this?

My birthday was 1.5 weeks ago. I celebrated by ordering dessert after lunch and that evening I went to Rouge for tango dancing. Turned out to be a great night. Birthday circle and everything. The next night was more of a party for me. The students have a weekly social as part of the course, on Wednesday. It was Riaan's birthday that day as well so it was turned into a birthday party for him, and I crashed it. With warning so it was more like a bump than a crash. Good party.

Then that weekend was a party at Nicole's place in Kennelworth. (That's close to Claremont, in the Southern suburbs, in case you wanted to learn some about Cape Town geography.) I went there directly from Heikki's. Turns out the invitation didn't contain a map but with some mapping service help through Minna I got there.

Most of the intense work was this week. I worked on a parser generator, three different asynchronous toolkits, Javascript libraries, and some thoughts about using Amazon's S3 and EC2. A few years ago in my technical writings page I proposed EBI open up their systems so remote users can install software on local machines, to reduces the bandwidth and update problems. I didn't think it would happen as it's rather outside the EBI/Sanger mandate. Looking at the costs for those services through it should only be $1,000/year to maintain such a system. I'll be cautious and say $5,000/yr. That's still in the range of a hobbiest. Amazing.

Grr, see, I told you that my mind's still thinking too much about work. When I do these sorts of evaluations I'm not only concentrating hard but I'm thinking about the different ways to use a project, how it might improve things, does it fix something that I didn't know was broken, how it can fail, is it maintainable, etc. Like a blind person figuring out the elephant. To that I'll add "and what do I need to do to prove that I understand correctly." That for me requires embracing self-doubt, which can take a toll. I went a tango class and practica on Wednesday and ended up doing very poorly in part because my head was elsewhere and in part because an aspect of being a good lead is the certainty which comes from confidence.

I've planned out my schedule for the rest of my stay here. Looks like I'll go up to Pretoria for a couple of weeks then back here to be in Stellenbosch for a week. In clear traffic it's less than an hour from there into Cape Town, so like Santa Fe to Albuquerque.

One final thing - yesterday (Thursday) I went by SANBI to give a talk about what I've been doing. I turned it into more of a "how better technology and understanding may affect the way bioinformatics software is developed."

No comments: