Speaker of the Month, January 2014

A whole new year. Cool.

I was at SQL Saturday DC, #233, at the beginning of December. I sat through several really good presentations. I could honestly give the award this month to any of the ones I took notes on, but I have to pick one person (although, not always, my award, my rules). So, speaker of the month for the brand new year is Konstantin Melamud (li|t). Yet another speaker without a blog. Maybe I should enforce my own rules at some at some point. <sigh> Anyway, I enjoyed Konstantin’s presentation. Let’s talk about it.

Performance Tuning – Index Optimization was an excellent presentation. Konstantin came at the topic very carefully. He started off with a knowledge level baseline, right at the start. I thought that was a pretty good idea. It let people level set from there. He then went through each index, provided a good definition of the index, then discussed information about it, tradeoffs, suggestions. It worked really well as a format for approaching the subject. He did things I wouldn’t necessarily have the guts to do, including doing a pretty good demo of full text search. Throughout the presentation he checked to see if the attendees had any questions. He even had a “I know more than you do” questioner who Konstantin handled professionally. It was a good presentation giving out good information (with an exception that I’ll point out below) that was really well presented. Konstantin’s delivery was very soft, quiet and comfortable. The way Konstantin answered questions and discussed the topic with people he demonstrated a great grasp of the topic. I’m sure people left the presentation and put some of his suggestions straight to work.

The stuff Konstantin could work on is the same stuff that many speakers (including me) need to polish. He didn’t repeat the questions. Even though it was a small room (packed to the gills, standing room only I might add), the acoustics were pretty terrible, so a question from a woman sitting directly in front of me was unintelligible. I’m sure people in the back never heard it. A few of his demos tanked because he didn’t reset after testing them earlier in the day. Understandable, but avoidable. He had one demo using clusters that basically implied the old saw that clusters are best for range data retrieval. I’d suggest changing that demo (and don’t trust me, talk to Gail about it). That’s about it. Like I said, a good presentation.

Once again, no way to determine where Konstantin is speaking next… Help me, help you… Where was I? But, I would absolutely recommend you track him down for a presentation.