I'm interested to hear others opinions on whether LightSwitch might make CSLA obsolete, particularly in the Silverlight environment.
If you haven't heard about LightSwitch, then check out this:
http://blogs.msdn.com/b/lightswitch/
Stuart
We should make a list of technologies that come out and could potentially end CSLA :-) I think EF was the latest one...
Seriously though, I dont think so. From what I understand, LightSwitch is a tool and it looks a bit "Accessy". Even if it kicks off, I would imagine there would be movement to integrate CSLA with LightSwitch.
Also the fact that it is SL only does not help. There are still a lot of folks out there using WPF, Winfworms etc with no view of switching to SL just yet.
Lightswitch looks "Accessy" because that's what it's intended to target - a .NET replacement for "low-end" Access-style development. There's been a growing discussion about the barrier to entry in .NET development, which is causing a lot of people to go to Ruby/PHP. Those developers don't usually switch to .NET when they enter "the big time". I don't know whether Lightswitch will really compete well in that space, but at least it's an acknowledgement that something needs to be done.
Given it's reliance on WCF RIA Services, I don't think CSLA could be integrated with it, even if Rocky wanted to. But that's a guess. And given what I see is the intended audience, I wouldn't see a lot of benefit to CSLA in that environment anyway.
As for the SL-only comment, I believe there are plans to expand Lightswitch to WPF - after all, that wouldn't be much of a stretch. Of course, MS has had an awful lot of plans over the years...
- Scott
As others have commented, CSLA's survived a number of "game-changing" technology onslaughts already. I think the key is in looking at what happens to apps as they grow up. There are a number of tools that'll get you out of the gates quickly, but as you start to add on all the stuff needed to make an app "real", you start to realize that you've slowly built most of CSLA...... again.
A couple weeks ago, I wrote a bit about Lightswitch with respect to some other "preview" stuff, commenting that there's just a little too much scatter in the dev tools roadmap:
http://blog.componentoriented.com/2010/08/a-schizophrenic-development-platform/
Today, after seeing a bunch of people dump on Lightswitch for a couple weeks, I added some more thoughts:
http://blog.componentoriented.com/2010/08/more-thoughts-on-microsoft-lightswitch/
I'd really love to see a tool like Lightswitch become a fast-start or prototyping tool that could throw some stuff together quickly for POC purposes, and if there's a path to grow these prototypes into real applications when needed, so much the better.
We've already got people doing code generation of CSLA classes, right? What's wrong with generating stuff based on an Access-like tool?
LS is an interesting technology, and one where I've had the opportunity to provide some input - specifically around their n-tier architectural choices.
It is very important to understand that LS has a specific target market - and it isn't "professional" or "enterprise" developers - though personally I think "professional" is a very slippery term.
If you think back to the late 80's and early 90's there was a very large cottage industry of developers (most independent consultants) that made good livings off FoxPro, Clipper, dBase and even Access. This is basically the LS target market - that kind of developer, whether a full-time employee of a company or an independent consultant, or perhaps even a consultant working for a firm (back in those days there were firms that employed this sort of developer).
CSLA is used in that market segment, and by professional and enterprise developers. So LS might replace some use of CSLA at the lower end of the spectrum - and if it is a better tool for those apps, then that's just great - I'm happy!
But for most of what I think CSLA is used for, I don't think LS competes - and its value proposition is very different at many levels.
The thing that makes me like LS though, is that it offers much of the simplicity of Access/FoxPro/dBase/etc - but with a solid n-tier architecture lurking beneath the surface. Of course only time will tell if they got all the details right so it really does scale - but it should scale very nicely - they did a good job thinking through the architectural model that is running behind the scenes.
Copyright (c) Marimer LLC