I know i have promised a nice clean peer mesh solution, and i do have example projects (one with and one without a resolver service) but i just havn't had the time to actually write the blog and explanations etc themseleves...
So in hte meantime, something that we must all remeber more often, shiny technology can sometimes be the enemy.
I have recently been lucky (or unlucky depending on your outlook) to inherit my employers primary dev projects to lead, overall we use a wide variety of technologies surrounding .net, I am forunate enough to be working for a company that sees new technology as a good thing, and the benefits its can bring when used correctly. However there are always pitfalls, i have recently ripped out a Silverlight 3 application from our mvc site thats under construction (app is an unfinished port of an unstable silverlight 2 app on the site).
The app provides a small but essential function within the site. Now let me be clear, Silverlight itself isn't at fault, silverlight 3 with ria services is so col in so many ways, but for our requirements there were a few problems, one our domain model couldn't be shared with silverlight without very heavy modification and disconecting from the overall project (duplicating objects etc) this obviously isn't ideal, and goes against the design ideals of silverlight 3 and ria. In the end i ripped it out because i was able to design the same functionality in an mvc view, which once my train of though expanded will provide additional functionality on our site.
Now as i said Silverlight 3 is great, and its a very shiny piece of technology, but i fear shiny technology can occasionally be blinding, especially when its coupled with the porting existing code (wpf app that does a very similar thing), just remember, if its shiny and you think it will save you time, double check it, there will always be cases where it turns out to be a bad idea.
