There's a lot of things I really like about WPF, but I think MS has made a lot of mistakes in its implementation, hence why it hasn't had a huge uptake in the 3+ years it's been available. I love the ability it has for a data driven UI (binding to ObservableCollections, etc), but interacting with the UI directly is just too hard - horrendous code required (such as continually needing to walk the visual tree) to do the simplist things. You'd think it would be a simple matter to do something like select a given item in a treeview, but methods haven't been included to do simple things like this that are really important. Excusing the rendering and IDE issues, this is one of my main turn-offs for WPF. The lack of a straightforward transition path from WinForms to WPF is a big mistake and is limiting its uptake. The learning curve is huge, and I have spent huge amounts of time trying to work out how to do things that should realistically be very easy. I know WPF is a different paradigm to WinForms, but I think there should be a layer/library that mimmicks a lot of the functionality available in the WinForms control set to help people transition. That, and add a lot more functionality (such as a Select method to the treeview items) to the WPF controls.

This is a major issue for WPF, and I can't believe MS haven't fixed it in the 3+ years WPF has been available. Blurry unreadable fonts will never be accepted in serious applications. Some people tell me when I point it out in my own apps that they didn't really notice it, but it really stands out for me. A related issue is when you put a raster graphic (such as a png image) in your app (such as a 16x16 toobar button image) it gets completely destroyed (WPF blurs the image), and there's nothing you can do about it.

I think mobile development is where it's at at the moment. With the AppStore and news stores for other device types we are back to the days where one person can write software and have it mass distributed. With a lot of programs costing less than $2 many people barely think twice before buying one - at less than the price of your morning cup of coffee why not? Store takes 30%, say 10,000 people buy your app (a number of programs on the AppStore have sold hundreds of thousands, but let's start low), that's $13,000. If you're lucky 100,000 people will buy it. That's $130,000. Not bad if you ask me if it only takes you a month or two to write part time. Chris Anderson

Why is it that every man and his dog feels it's necessary to write their own report writer? Along with code generators it seems to be a favourite project of software developers, but rarely comes up with anything more than an expensive experiment - satisfying the programmer who writes it but not the customer who's paying for it. Report writers are just another meta-problem like code generators and ORM frameworks - fun for the developer but don't produce bang for buck for the client. Clients aren't paying us as developers to solve the problems we find interesting, but paying us to solve their particular business problem. [Sorry, just a rant about the million or so custom reporting engines out there] In my experience users always want to be able to customise their reports - and most reporting frameworks don't allow this or charge exorbitant license fees (where the user then needs training to use the tool). More often than not the users come back to us as developers to just produce the report for them. I don't think there is a perfect reporting framework out there for any one particular project, and it really comes down to the particular requirements of the project as to what you should use. Can you fill us in on your particular requirements? I've used Crystal Reports before, and for the most popular reporting tool out there I'm yet to find someone who's used it that actually likes it or swears by it. Personally I've encountered so many issues just deploying Crystal Reports in the times I've used it (spending much time in their kb for workarounds to their bugs) that I couldn't recommend it. I haven't used the full SQL Server Reporting Services but have used the local report engine (RDLC), which for all intents and purposes is the same thing. But you are likely to hit problems once you get beyond anything but basic reports in my experience. There are a number of others out there that I haven't used but look OK (Telerik, XtraReports, etc). Ultimately the only solution I've ever found to be popular with users is to generate reports as Word or Excel documents (either through automation, a VSTO addin, or populate it on a server using the Office Open XML or WordML formats). Users can then customise them to their needs, and if you have a template system (easy to do) they can create their own. Especially easy with Office 2007 with the Office Open XML standards (though I've also done it with WordML for Office 2003 compatibility). The advantages are that almost every user will have Office on their desktop,

I had a similar problem on my old work machine doing .NET Compact Framework development, except it crashed every 15 minutes. VS2005 was fine on that machine, but doing anything in the forms designer in VS2008 on the same machine kept crashing it or taking 3 minutes to sort itself out with every change I made (eg. when I dropped a control on the form or changed a property). Since it was a .NET CF 3.5 project I was working on it was a very painful project to do. But I couldn't find other machines with the same problem or any report about it on the web so it must have been something to do with the configuration of my machine or the software I had installed. I was happy when that project was over. Chris Anderson Peer Placements My Blog