Amidst all the hype around Whidbey and Yukon, one area that's been perhaps neglected is the fairly significant changes that have been made to ADO.NET. Having spent a fair amount of time working with customers on data-driven applications, there are a number of questions that I get asked almost every time, including:
None of these are impossible to achieve today, but all of them require a fair amount of work. The good news is that ADO.NET 2.0 provides simple solutions to each of these problems out of the box. There's a new UpdateBatchSize property on DataAdapters, that supports batched updates; there's a provider-agnostic class factory; a Load method on DataTable objects that takes a DataReader as a parameter; there's a new binary serialisation option for DataSets, and finally there's support for Multiple Active Result Sets (MARS) in MDAC 9.0.
Where can you go for more information on these? There's a little bit of information on the ADO.NET website, but today at least that consists primarily of slideware from the PDC. But I was blown away by a great demo from Pablo Castro (one of the ADO.NET PMs) that's just gone live on MSDN TV. Now I have a confession to make - I generally hate webcasts. For some reason, when I'm watching something online my attention span drops off and I start being distracted by email. But Pablo's session was pure demo - no marketing fluff - no explanation of the architecture - just straight into coding C# in Visual Studio 2005. The demo segment is about 20 minutes, and he demonstrates how ADO.NET 2.0 solves at least four of the problems highlighted above. A tip - download the demo file only rather than the full presentation - you can reduce the download from 58MB to 14MB and just miss the 2 minute intro.
Go on - skip browsing aimlessly around the web and take twenty minutes out to learn some really cool stuff - you won't regret it!
I know many developers who rely on Virtual PC 2004 these days, both for application compatibility testing purposes and also for beta testing releases of Visual Studio 2005 ("Whidbey") and SQL Server 2005 ("Yukon"). A few colleagues and I have been trying to track down a little issue for some time, and I wanted to share it more broadly outside of Microsoft to see if anyone else has had the same problem.
The problem is this: when you toggle a Virtual PC guest into full screen mode with [Alt Gr]+[Enter], the [Ctrl] key often gets "stuck" down. If you start typing in the guest, you often see funny things happen: the taskbar goes into multi-select mode; some characters don't appear etc. Hitting the [Ctrl] key once resets it and brings you back to normal.
A few of us raised this as a bug with the Virtual PC team, and it turns out to be a problem specific to UK keyboards (and possibly others in Europe). The [Alt Gr] key is effectively the same as pressing [Ctrl] + [Alt] together; when you release it, VPC only toggles off the [Alt] portion of the keystroke, leaving [Ctrl] held down as far as the guest is concerned.
The solution: change the host key to something else on a UK Virtual PC machine. I can recommend [Right Ctrl], since it's not used for anything else. Once you make this change, you'll never again have this problem - toggling to and from full screen works just fine. The bug has been fixed moving forwards too, which is good news.
If you're attending TechEd Europe this year, you may be interested to know that we're running a series of "birds of a feather" (BOF) sessions throughout the conference. The idea is simple - if you want to host an interactive discussion with other attendees on a topic of your choice, we have a room dedicated to this kind of activity. You simply nominate the topic you're interested in discussing; others can vote on the topic and the most popular topics become part of the TechEd programme. The BOF sessions are being run independently by INETA, an independent .NET organisation. So, if you want to discuss extreme programming, service-orientated architectures, mobile devices, user groups or even blogging - it's up to you to nominate it! Hope to see some of you there.
I don't normally comment on security issues, mostly because it's not my specialist subject. It's fair to say though that our record hasn't always been admirable in years gone by, as Slashdot and others gleefully point out. But sometimes perception starts to overtakes reality and comments that are no longer justified start to take on a life of their own.
I saw one statistic that shocked me this afternoon though. Windows Server 2003 has been released for a whole year now. How many vulnerabilities do you think we've had to patch in IIS 6.0 over that time? Check out the answer for yourself by visiting the TechNet security centre and selecting "Internet Information Services 6.0" from the drop-down list. I often hear developers say that Apache is far more secure, but is it?
Don't get me wrong - I'm not for a moment trying to suggest we've solved the problem, or that we're in any way complacent. I know we've got a huge amount of work to do before we can truly stand up with our heads high. We are deadly serious about getting this right. But perhaps we're not quite as bad as the industry perception suggests...
I want to do something which I don't think has ever happened before: invite open peer feedback on one of the TechEd 2004 Europe tracks before it's completely finalised. We've been working hard to come up with the best sessions we can create or find on SQL Server and ADO.NET, and finally we're at a list which is more or less complete. There are 35 sessions covering both SQL Server 2000 and 2005, but I want to hear your verdict: have we got the right balance, and are we missing anything that you think we should be covering at a conference like this?
Please give me your honest feedback here.