With the drop in mortgage rates that started in early August , it comes as no surprise that mortgage applications are now starting to rise 
The next thought that comes to mind is whether or not home sales will follow. Sales activity seemed fairly brisk in New York, althought the NAR’s Pending Home Sales Index [Note: PDF]  showed modest declines in all regions except the south. The idea here is that contracts are the better indicator of the current state of the real estate market. However, this is more of an informal survey from their members. It is still behind the market since they have only report through mid-July.
With Katrina and higher oil prices, it will be interesting to see what happens in September. I’m thinking good thoughts.
I will probably get a little annoyed if interpretation the the next round of housing data does not consider that August is usually one of the seasonally slowest times of the year for housing sales. Hence the infamous, seasonal adjustment should be applied. See Lies, Damn Lies, And Government Statistics: Part I 
One other thought is the idea of using mortgage applications in predicting home sales. There is an interesting article published from the Dallas Fed Can Mortgage Application Help Predict Home Sales?[Note: PDF]  Its a bit dry but my sense is that the mortgage app data really lags too much to be an effective predictor of home sales without a lot of tweaking.