I stumbled upon this excellent article written by Scott Haselman: Creating an OData API for StackOverflow including XML and JSON in 30 minutes.

I, like many developers, spend so much time "handcrafting" my solutions that reading about how easy is to create an OData service using the new versions of .NET and Visual Studio was really refreshing and pleasantly surprising.

For the last two weeks I've had the opportunity to work on a project which uses Entity Framework (EF) ver. 1.0. For more than a year I've been reading complains, reviews and comparisons regarding EF but finally after using it I have my own opinion:

• It's not a mature solution.
• Definitely falls short when compared to other products in the same category, like NHibernate for example.
• It's not as bad as some people want to make it appear.
• Can help some teams deliver small projects on time, especially if the developers have no experience on the use of O/R mapping tools or if like in the project I was working on, most of the business logic was implemented on the database side.

There are plenty of good Dependency Injection Container implementations out there. The selection of an IoC container is not only defined by the developer's preference, which can swing the vote big time since most developers swear by their favorite container, but also of other factors like standards in the organization, the usage of another product which already supports, or favors, a given IoC implementation, etc.

The usage of an IoC container should always be abstracted, making the process of switching between implementations fairly simple, or at least, the refactoring needed to accomplish this task will be minimal compared to the one needed if you were using references to a specific container all over your codebase.

I'm not a fan of meta posting but this information was so valuable to me that I wanted to keep track of it on my blog just in case something happens to the original article.

If you are one of the unlucky few (I would like to think there are no many of us) still using Visual Source Safe in 2010, then I bet that you've needed to get a version of your project at a given point in time. It happened to me today again and instead of setting with the "Sorry, SS doesn't support that" I set up to find a solution.

I found an old post by Brennan Stehling with the exact instructions to get a point in time snapshot using Visual Source Safe or using Brennan words: Get by date.

This post from Uncle Bob generated lots of talking on the developer community regarding Inversion of Control (IoC) and Dependency Injection (DI), which is always good for the sake of  learning and general awarness. Almost every developer with a blog posted about their take on this subject. Two people which I respect a great deal and suscribe to their blogs also posted their reactions, please read Ayende and Davy's blogs for a better understanding of their positions.

What I get from all this, in short, is that Uncle Bob accepts the advantages of using dependency injection but prefers to encapsulate the access to the IoC container on factories. Davy seems to accept this approach on another post...

Today, when writing a function that opens an image, adds a simple text on the top corner and streams the resulting image to the client I noticed that the original image used as template couldn't be renamed or deleted after the code ran for the first time and for as long as the web application remained running. Restarting the web application makes the problem to go away until of course the offending code runs again.

After installing and configuring CruiseControl.NET on my Windows XP development station I was getting a File not Found (HTTP 404) error when trying to see the details of a given build, i.e. selecting Latest Build from the main menu.

Since these URLs don't point to a file in specific but to a "fake" location that is in place handled by CruiseControl.NET I wasn't sure if there was something wrong with the directories I was using on my project's configuration.

First things first: If you think .pdb files are only used when debugging your application locally and are not needed when deploying production binaries, please read the following article by John Robbins: PDB Files: What Every Developer Must Know.

This is one of these things that bother you but not enough to make you take any action about them.

Ever since I started using Visual Studio I've been bother by the fact that if your solution contains multiple web projects, when debugging, an instance of "ASP.NET Development Server" is started for every one of them, even if you have selected a "Single startup project" on your Solution properties.

I'm currently multitasking: writing this post and also running the uninstaller for the Microsoft Visual Studio Team System 2008 Database Edition GDR.I was very excited with all the new features, already planning in my mind how I was going to create new database projects as part of my solutions, how I was going to stop using SQL Management Studio to write stored procedures, how I was going to start deploying my SQL changes as part of my releases until I actually installed the application and tried to use the functionality I need the most from the VS Database Edition: Schema and Data comparisons.After trying for several hours to make it work, excusing the application and blaming the “learning curve” for the lack of progress I just decided to wait until the next version of the GDR is released.Too many things just don’t work when comparing schemas; maybe too much attention was paid to the “new” features and the “old” features that...