>
>If you put the RI into the wrong tier, then I agree, this scenario isn't reality.
I think when most people talk about changing the back end they think of replacing one data store with another and it magically works. This is what I was talking about. However, it doesn't work that way. There needs to be a change to the objects that interact with that data store. These objects are very much part of the data layer.
>
>Seriously, Craig: It is very much reality and I use it every day. VFP backend here, SQL Server there, XML over the web. No big deal. It works beautifully.
That's great. I'm glad to hear about the success people have doing this kind of stuff.
>
>And even if it wouldn't, it would be a very stupid thing to do, to design an application today that is invalid and useless by definition in a couple of years from now.
I totally agree with this. I'm currently redesigning a legacy Fox app that has had many programmers work on it over the years. It is all Fox 2.x code, a mixture of DOS and Windows. There isn't lots of data nor many users. I'm looking at using MSDE because management won't go for SQL Server at this point (I'm working on them) and it will give the junior programmers on my team some experience with C/S. It will definately be n-tier.
>
>Isn't that exactly the dilema we have ever since and finally got resolved by at least those that understand n-tier development and good object design? I find it very sad, if not irresponsible, to argue otherwise.
Yup. n-tier really makes sense. I was preaching n-tier on a different project a year ago. It didn't take long for the other team members to see why it is better than the 3-tier model they were thinking about.
>
>Markus
Craig Berntson
MCSD, Microsoft .Net MVP, Grape City Community Influencer
