I am putting this out to everyone to get some input. Since mid 2008 I had been digging into SL and lovin' it. A little background - I am the sole developer (yea, DBA, .NET Architect, etc., you name it, I do it) and was architecting an enterprise application utilizing SLx, Csla, C# x (where x is a version). I had finally gotten to a good point around mid Feb., March, and then the rug was pulled out from under all of us SL developers. So, I started digging into MVC 3, CSS, now jquery, jqGrid, and the plethora of other sh#t that it takes to make a cool looking enterprise application.
My strong point is NOT UI's but with SL it was easy to create a cool looking UI as all you needed was XAML. The business logic and business rules are a piece of cake with Csla and even Wcf Ria services with DataAnnotations, but I prefer Csla because of the consistant structure and format.
My question to you other devs out there is this: if you were developing an enterprise application (billing, entering contracts, entering futures, a myriad of other items (our database has around 250 tables), what would you use for an enterprise LOB application? Would you charge on with SL, not knowing what the future is, or would you embrace MVC 3, and all the other stuff to create an application?
Again, remember, I am the sole developer and our parent company is giving me NO support (they are running a 30 year old AS/400 written by someone who retired 4 years ago, but still comes in 2 weeks a month because nobody else can maintain the system. Yea, lots of fun:)
Thanks for your input,
It's a little bit of a gamble, right? Developing "NEW" now you want to not be on a perceived dead end a year from now.
You've said that you had gotten to a good point around Feb/March - conceivably you either have projects that are live or you're substantially down the road. For us, the same sort of revelation came upon us - most of the way towards a really robust SL application only to have open questions surface as to the technology we were using.
Starting new, presumably we're talking .NET, it's really three options at this point we're usually talking about, right - ASP.Net, WPF, or SL.
I think WPF is an easy one to lop off the list presuming your application doesn't require the full features of WPF. It seems to have the least bit of "hope" when it comes to moving to additional form factors, Metro, etc at present time; plus, when using CSLA you really aren't taking a big hit in terms of development efficiency if you're using SL instead.
When debating SL and ASP.Net - for my scenario at least - the application I'm involved with is a "next-generation" from a VB6 application which - while old - is still pretty rich and interactive. I am not knowledgable in MVC (though I've done some decent work in ASP.Net) but I just can't imagine being 30%-40% of where we are right now in our development if we'd have used ASP.Net and there is no way the richness of the application would come close to what we've done in SL.
While SL in its own right may not be the long term panacea, it certainly does seem that we have a really good possibility of retaining a great deal of our code in a translation to a WinRT app. That's pretty exciting. I think if there wasn't that avenue, things would be looking a little more bleak.
If I were starting a new LOB application today without huge reach demands and it commanded a lot of richness (depending on the application of course), I would still choose Silverlight with the idea that it's getting me a great portion of the way to having a WinRT application, and that I can really develop tremendously rich applications quickly and effectively.
The Silverlight --> WinRT path existing is a big draw; it's like a safety valve in some ways for my Silverlight application. I sure hope it ends up being as manageable as the sentiment seems to indicate.
Copyright (c) Marimer LLC