April 2006 Archives

Woo, that was fun! I just finished my talk half an hour ago. (update: I've updated the slides to be a page per, uh, page rather than a page per "build")

I got a couple of "that was the {best,most useful} talk at the conference so far!" comments, so that alone made it worth getting up at 5.45am to fly up here!

The slides from the talk are now on the talks page of my new shiny website that Vani put together for me.

I had about 70 slides for 45 minutes (made it in 47!) so it went pretty fast, but I really think that's a lot more fun. I am sitting in the InnoDB: Designing and Configuring for Best Performance talk now and the pace is a bit too glacial for being in the last slot of the day, at least for me on too little sleep...

I'm increasingly convinced that "more content faster" is better for conference sessions.

Earlier I saw the My Second Life Runs on MySQL: War Stories from the Metaverse talk. Woah! Was there some cool stuff there or what!

The MySQL things were good (and mirroring some of my talk - a bit of confirmation is always nice), but the things about what the Second Life world can do and what people have created there were jaw-droppingly awesome. Ian, if you read this: Thank you, that was fun! :-)

Heading out for a quick dinner in a few moments and then to the airport to go back home.

Next year I'll have to make time free to come for the whole conference. Being used to OSCON it's a bit strange how everything is focused around just MySQL. I can't quite put my finger on it, but I like it this way.

I've been following the plethora of MySQL User Conference postings over at Planet MySQL. Woah, it looks like fun!

Sadly I am missing most of it this year. I don't have time to go the whole week, so I am only making it on Wednesday to give my talk Web Performance and Scalability with MySQL.

Usually I'd worry tons about my talk, but I think it'll be lots of fun. I've been tuning and fixing up my slides and I really like how it's turning out¹. Hopefully the people coming to the talk will like it too! :-)

The talk description says "fast paced", and it's not kidding. I have enough content for three talks - but then people would just have too much time to read their email and that's not what they come to the conference for, right? Right?? :-)

Anyway, I'm looking forward to it a lot, except for the part about getting up early to get to the airport. (Vani promised to take me though, yay! :-) )

¹ Of course it helps building on a popular talk I've given ~4 times before ...

The new 17" PowerBook MacBook Pro is cheaper than a similarly configured 15".

The 17" is $300 more when they both are in their default configurations, but if you upgrade the processor in the 15" to 2.16GHz ($300) then it's the same. Upgrade the drive to the same class as the 17" and the 15" is $100 more.

I was glad to see that the 17" still has Firewire 800. Maybe it'll come back to the 15" in a later revision?

As mentioned earlier I started putting together a system letting vendors use the NTP Pool.

While putting it together I stumbled upon the following quote in the NTP Faq

Time usually just advances. If you have communicating programs running on different computers, time still should even advance if you switch from one computer to another. Obviously if one system is ahead of the others, the others are behind that particular one. From the perspective of an external observer, switching between these systems would cause time to jump forward and back, a non-desirable effect.

As a consequence, isolated networks may run their own wrong time, but as soon as you connect to the Internet, effects will be visible. Just imagine some EMail message arrived five minutes before it was sent, and there even was a reply two minutes before the message was sent.

Even on a single computer some applications have trouble when the time jumps backwards. For example, database systems using transactions and crash recovery like to know the time of the last good state.

Therefore, air traffic control was one of the first applications for NTP.

Poul-Henning Kamp wrote an open letter to D-Link as a last-ditch attempt at getting them to show any interest in resolving the mess their badly configured software has caused (and is still causing).

I'm not entirely surprised, but still disappointed that they didn't learn from Netgears mistake in 2003.

I worry that the NTP Pool will get a similar problem at some point. I'd like the pool to be able to help the vendors (better than them abusing random servers!), but I need to setup a process and a system for vendors to use the pool in a way we can monitor and eventually disable (or point back to their own servers) if it ends up being a problem. One of these days ...