This post originated from an RSS feed registered with .NET Buzz
by James Avery.
Original Post: The Passion of the IT
Feed Title: .Avery Blog
Feed URL: /blog/Install/BlogNotConfiguredError.aspx
Feed Description: .NET and everything nice
My friend Ben has a very good post on how we usually do a disservice to our clients by using a new and untested technology over an older and proven technology. I wonder how many software projects crash and burn? Most numbers I hear are over 50%. What if whenever we tried to build a bridge, 50% of the time it failed... would we still waste money building bridges? Do you think a structural engineer says "Hey guys, lets try out this new thing with the bridge and see what happens?" Somehow I doubt it, I think there is alot of testing and architecture put in before they every try to build that first bridge with a new technology.
I think the problem is two fold. The first problem is that whenever a new technology comes out Microsoft pushes it pretty hard, they drum up some case studies that show how it makes everything easier and talk about how companies will save millions, etc. This causes managers to push their people toward the new technology, even when that technology might not be ready, and their people are definitely not ready.
The second problem is that most of the good developers that I know are passionate about the technology they work with, and love working with new cutting edge technology. This is probably what they enjoy the most about programming, I know it is what I enjoy. There is no way I would be happy working on a COM app right now, no way.
So what happens is we throw ourselves into the new technology and end up feeling our way around on other people's money (unless you own your own company, then it is your money) Only now can I confidently say that we as an industry are coming up with some good .NET practices. Internally Avanade has alot of stuff going on in this area, and externally we are finally starting to see some of the application blocks mature but I still think we are still missing a good .NET reference application from Microsoft. I hear about .NET apps all over the place that were built over the last three years with an architecture that would make you cry.
By now you are probably screaming, "Luddite! Burn him!". It is not that I don't like new technology, I love new technology. It is the reason I am in the job I am in, and the reason that I enjoy it. But I think we need to, as an industry, be more careful with how quickly we move to a new technology. We need to remember that our main goal should be to create a successful software application, solve our clients problem, and make them more successful. This should be our number one goal.
I caught myself being bad the other day, I was going to try and use the UIP block to solve a particular windows form issue. I was excited about figuring it out, and working with something new, but I don't think it is ready for prime time yet. I found out that I would need to hack it up quite a bit, and customize it in a number of different ways, and would probably end up with a solution that was way more complex then was really needed. I had to catch myself and say that we could write a solution from scratch that was less complicated and more geared toward our specific problem. I had to suppress the passion that I have for new technology for the sake of the customer.
And I don't want this to read like I am validating the people who are using outdated technology, if you are building a new app with COBOL you are probably making a mistake. And even if you are building with COM you are probably making a mistake, because MS is not going to support you anymore. I think the issue that I am trying to get at is that MS is already putting all theirs gears toward the "Longhorn" wave, when most people don't even know how to build good .NET 1.0 apps. MS needs to focus more on making their customers successful, where are all the reference architectures? The Patterns and Practices group is awesome, but it is abstract... there needs to be implementations out there. I know of a project where they used IBuySpy as the reference architecture, yeah that sucked. Right now we have all the tools we need to solve our customers problems, we just need to focus more on learning how to use them, and people need the time to figure it out instead of having to worry about all of this being worthless in two years.
Previous Topic
