tsJensen

A quest for software excellence...

Visual Studio LightSwitch Beyond the Basics

Before you make up your mind about LightSwitch prior to beta release, watch the new video on Channel 9. Before you decide that LightSwitch is for non-developers, watch Joe Binder in this video extend a LightSwitch application with a Silverlight user control and a custom WCF RIA Service.

The key point, I believe, made in the video is that EVERY LightSwitch application is essentially a 3-tier application. The extension of the Data Access tier with WCF RIA Services makes it possible to extend those tiers into the enterprise infrastructure easily.

lsoverview

The demo also makes a very important point. When you write a real LightSwitch application, you will be writing code. While LightSwitch might be appealing to non-professional or hobbyist or business types, whoever uses it successfully will in fact end up writing code. Will they have to be an expert developer. Not if they have support from the enterprise pros that build the infrastructure for the business.

I still want to see and play with the beta before making up my mind, but what I’ve seen so far looks very promising.

Microsoft Visual Studio LightSwitch Unfairly Judged Before Beta Released

Two days ago, Microsoft unveiled Visual Studio LightSwitch. I heard about it first on Somasegar's blog and have followed the chatter and the growing list of blog posts and comments. I watched the Channel 9 preview video and the VS Live keynote video introducing LightSwitch. And then I watched them again making notes.

I'm deeply disappointed in a number of bloggers I regularly read and their many commenters that have parroted their sentiments. I won't name them here. Their negative reactions demonstrate an ill informed prejudice and predisposition to discount all 4GL tools without careful analysis and thoughtful evaluation.

Based on my careful notes of the demo videos and "official" blog posts about the tool, these bloggers and many commenters have made statements and critical judgments of LightSwitch that are demonstrably false or grossly inaccurate. And the first early beta is not even available yet. This rush to judgment ill becomes the professional stature and experience of these otherwise well respected members of the .NET development community.

From a certain point of view, I understand the nearly autonomic reaction of these .NET development community leaders. Professional developers have long dealt with the problems that so often plague solutions created by a host of desktop 4GL tools such as Access, FoxPro, FileMakerPro, PowerBuilder and many others. These tools have been highly successful in the business world seeking to just get something working that they could afford. In enough cases to be considered a plague by many professional developers, these solutions have far exceeded their limits as the business grew and ran into some of the hard and harsh limitations of these tools.

Many professional developers have been put into that crucibal of having to tell the business owner that their pet solution will not scale, has to be completely rewritten, and the effort will cost mega dollars to get a similar solution that will scale and grow with their business. The business owner doesn't want to spend the money, so he insists that the developer just make the 4GL solution work. Happy developers then fire their client. Needy and miserable developers then swallow their pride and spend more hours in frustration dealing with a 4GL too they come to hate more often because the solution wraps around a database that was designed by someone who thinks of dinner when you say the word table.

Scott Adams owes his fame and fortune in no small part to the advent of ill conceived 4GL tools and the more horribly designed solutions created with them by the pointy haired bosses who fancy themselves tech savvy.

So I understand the rush to judgment, but the facts are not there to support the negativity and biased ignorance I've seen in so many recent blog posts and comments.

Anyone can actually take the time to watch the videos and read the official blog posts and carefully consider what they have learned. Any professional presented with a tool for solving a problem that has never been satisfactorily solved in the past who automatically dismisses the latest attempt does himself and those who respect his opinion a disservice.

Here is what I've learned after reviewing the demostrations and taking detailed notes and carefully considering the possibilities without making invalid comparisons to the host of 4GL tools on the market today.

  • LightSwitch is NOT a 4GL in the traditional sense of closed source, desktop bound, proprietary database, with few if any extension points.
  • LightSwitch is a set of .NET assemblies and Visual Studio designers and project templates that produce a model based on XAML from which a Silverlight 4 and WCF RIA Services solution is generated.
  • LightSwitch is a .NET solution and can be extended in Visual Studio Pro (and up) with standard Silverlight control projects and server side support projects.
  • LightSwitch supports three data sources: SQL Server (and SQL Azure), Sharepoint 2010, and most importantly WCF RIA Services. In fact, SQL Server access in the 3-tier solution is done via WCF RIA Services and Entity Framework. I was not able to determine exactly how Sharepoint access is supported under the covers but would assume that the same solution is generated, with WCF RIA Services providing access to the Sharepoint list libraries and other Sharepoint data on the host server.
  • There is certainly an incoherent marketing message from Redmond on the subject of the target audience but it is clear that while marketing speak would have business buyers believe that they will be able to write applications without code, the fact is that some coding skill will be required to write even the simplest of applications in LightSwitch where commonly found business rules such as validations and computed values on existing data sets are required.

Based on what I've seen so far, it seems the sweet spot for LightSwitch might be the entry level developer or what one blogger called the "productivity developer" with professional support from senior developers and architects constructing back end systems, designing databases and providing safe and simple WCF RIA Services for the less experienced developer to consume to deliver line of business applications that may never need the skilled hands of the top notch developers.

I want to learn more about LightSwitch and see it and play with it before I make up my mind about its usefulness. I want to see how easy it is to extend and expand a LightSwitch application in Visual Studio Pro as promised but not demonstrated. I want to learn more about the extension points. And I want to see and perhaps participate in providing constructive and informed feedback. More importantly, I want to see what the Microsoft team does with that feedback.

Bottom line. It's far too soon to judge. Let's get the beta and provide Microsoft with some constructive feedback instead of the diatribes I've seen so much of in the last couple of days.