Well, we have shipped whidbey beta 1. This is a great release for the ado.net team, we have managed to not break away from an existing data model while adding great new features. Everything you have will continue working as is, we have fixed old problems and added features to address the biggest shortcoming of previous releases, there are a few features that I am really excited about. I would like to take the time to blog on individual features, one per week for the next few weeks, drop me some feedback on the order that you would like to see these or with suggestions for ado.net features that I might have missed.

  • Distributed Transactions with System.Transactions. Finally managed distributed transactions are easy to use, leaner meaner faster and as robust as always, this is a really good feature.
  • SqlBulkUpdate. Bulk update is nothing new, but now you can do it from managed code and it is faaast. Not able to insert 100.000rows of data into sql server with ado.net 1.1? Take one of these SqlBulkUpdates and call me in the hmm.. never mind you are probably done by now <g>
  • Adapter batch update. Nothing like a good batch to wake you up in the morning, hey seriously this is not rocket science but it is about time that adapter update worked
  • Provider SDK. So you want to be a ..managed provider writer? Do you hate to write a lot of code? Do you want pooling and performance counters to just work? boy do I have a deal for you! How much would you pay for this once in a lifetime opportunity? Well I think you know where I am going with this, we have made common classes for all of our providers and abstracted pooling and perf counters, this is not a vanilla sdk, all of our providers derive from this common model and so can you.
  • Provider factory. This is not really one of my favorite features, but writing about it allows me to bash on UDA(which is imo a bad thing). One of the biggest requests when ado.net shipped was UDA, universal data access. I did not like it then and imo I don't like it now ( I have to put in the “imo” any time I say something like that, but trust me, UDA is Eviel “imo“). This feature looks like UDA, smells like UDA but it is (thankfully) not UDA. Thanks to the common classes you can now use a provider factory to write provider independent code, at runtime you can connect to Sql or to Oracle with the same provider independent code. Performance? the exact same performance that you would get by connecting with OracleClient or SqlClient. Why is this not UDA? well this factory model breaks down any time you do _anything_ (I forgot to mention in the post below that I love _underscoring_ too) that is backend dependent. Not UDA, but an interesting feature.
  • SqlNotfications. I am bluffing I don't know anything about these. I am putting this in here hoping nobody will notice.
  • Pooling improvements. I am not kidding about this one, I know about pooling. If anybody votes for this one they are going to get hit with more information than they can shake a stick at.

Hey its friday, calling it a day.

Rambling out.

Standard disclaimer, this post (and everything in this blog) is just my opinion, information here is provided “as is” and confers no rights.