Often, the last day of a conference is a real downer - the expo hall is gone, people have already started leaving. But for me and many, the conference was still in full swing. I had two must-see sessions to attend in the morning. The first was by WSE P.M. Mark Fussell on interoping Indigo and WSE 3.0 messages. The second was a 400 level session by ADO.NET P.M. Pablo Castro on Advanced Data Access Patterns in ADO.NET 2.0.
As the previous day's first session was 10am I had it in my head that the Mark's talk was at 10am, so I was really frustrated at 9:30am when I was still packing up my things in the hotel room to realize that his talk started at 8:30. Uggh. Luckily I was able to catch up with Mark later who very generously gave me a quick rundown and for the rest, I will wait for the conference DVD's.
I did manage to attend Pablo's talk. Now here is a guy who LOVES what he does. This is not a project manager who is going to disappear off to a Windows Media project any time in the near future. He lives and breathes data access. Pablo is a great presenter, even with his beautiful accent and quick pace of talking. He went through three of the more difficult things developers with high end applications have to deal with.
The first was how to update data when you have LOTS of it. Leveraging the ado.net 2.0 bulk insert is obvious for new data, but he showed how you can also use it to do updates and deletes. The basic concept is to create a new temp table and bulk insert the change rows into that, then run a stored procedure that will handle the update from that temp table.
The second was caching data in the middle tier - again, dealing with lots of data and leveraging new goo in ADO.NET 2 and SQL Server 2005 - all revolving around Query Notification.
Another of the biggies he talked about was querying data in memory - querying a dataset. Although we know that LINQ is coming for this - it is waaay down the road. Pablo gave us a quick look at some basic concepts of a query processor that he and others on his team have built that we can implement in our apps today. Although he will be p osting the code for this online because it is extremely complex, the key takeaway was that he was leveraging a lot of new innovations in the data access API to accomplish this.
And that was just the morning. I spent some time Koji Kato from the Tablet team in the track lounges and had a nice chat with Brent Rector who apparently pulled off a minor miracle in his session that morning! And, if you saw my prevous post mentioning that I would love to meet Raymond Chen, well I did. But I still forgot to tell him that I say "thank you ramond chen" every time I start up my last in production FoxPro 2.6 Windows and DOS applications on my new computer. Every windows user in the world has a lot to thank this man for. So from all of us: THANK YOU RAYMOND CHEN!
In the afternoon, I had fun hanging out with Steve Maine, Peter Provost and others (forgive me my failing memory...)
Then on to a grand never ending dinner that evening with Patrick Hynds, William Tay, and a host of people (I have to get that image off of my tablet that we all signed). It was fun because we kept picking up people throughout the meal and the waiter was grand in starting all over again with each new addition to our table.