Looking (Suspiciously) at SP1
What's in SP1 for ASP.NET developers? At least one bug fix, a couple of enhancements and -- perhaps -- something that will change your life. But Peter is suspicious.
In my column on "Handling
Data Contention with Optimistic Concurrency
," I mentioned the bug in
handling Nulls in the ASP.NET 2.0 SqlDataSource. One of the comments for that
column was from Scott Hunter from Redmond, Wash. who noted that the bug would
be addressed in SP1 for .NET 3.5.
Well, the service pack came out on Aug. 11, and the bug is fixed.
So what else is in the service pack that's relevant to ASP.NET developers?
As usual with .NET, not much and quite a lot.
The "not much" comes from the fundamental reason that you use a framework:
Much of what you use as a Web developer is shared among all developers. The
result is that many things that you'll find useful are part of enhancements
to .NET generally. For instance, I wouldn't discuss LINQ in this column -- even
though lots of ASP.NET developers use it -- because LINQ isn't "ASP.NET-specific"
URL routing is also cool. Back in an older article on implementing REST in
RESTful ASP.NET Apps") I discussed how you could intercept a URL, decode
it and generate a matching page without actually retrieving an .ASPX file from
the disk. URL routing simplifies this process so you can provide URLs to your
users that make sense (e.g., //www.mysite.com/Products/List/Inventory) and are
independent of the actual file paths for your .ASPX files. I'll look at this
in a later column -- much later, as I describe below.
The Elephant in the Room
The big news for ASP.NET developers in SP1 is Dynamic Data (covered by Roger
Jennings in his article "Generate
Web Sites Automatically"), which is supposed to go one step beyond
DataSources and DataViews: Just tell your application about your data model
and you'll get the pages you want. Naturally, I'm suspicious, for two reasons.
First, previous attempts to develop the user interface from the data model
have been crashing failures. In application development, a data model "has
a relationship" to an object model, which "has a relationship"
with your application's user interface. But those relationships are very weak
and for good reason. The primary reason for most developers is that if you're
after reusability, the last thing you want is a data model that's tied to some
specific application's UI.
The primary reason that I don't trust having a UI tied to the data model is
that an application's UI is driven by workflow and the organic relationship
of people to each other, more so than another part of the application. There's
no equivalent, in UI design, to the kind of rigor that Codd's normalization
rules provide in database design. In my mind, an attempt to derive the UI from
the data model is, at best, a niche solution that will work in only a few cases.
But that's where the Entity Framework (EF) steps in. Acting as the Object to
Relational Mapping layer between the data model and your UI, the EF is supposed
to provide a way to generate an object layer that converts the data model into
something more useful. The question is whether you can build an effective UI
on top of EF without having to either create separate business layer objects
or write a lot of presentation layer code.
My second reason for suspicion is more prosaic: Will the tool let me do everything
I want/need to do? I had the same concerns with the DataView/DataSource model
that came with ASP.NET 3.5 but -- so far -- I'm a convert. The event model and
the options to insert/incorporate custom code have let me do what Microsoft
always promises: Let us take care of the boring, tedious grunt work while you
concentrate on writing the code that's unique to your company and delivers real
However, just because DataViews and DataSources worked out, doesn't mean that
EF and Dyanmic Data will.
So I'm going to start working through living with and creating applications
using dynamic data. I don't know the answer in advance, so stay tuned -- or
send in your own experiences.
Peter Vogel is a system architect and principal in PH&V Information Services. PH&V provides full-stack consulting from UX design through object modeling to database design. Peter tweets about his VSM columns with the hashtag #vogelarticles. His blog posts on user experience design can be found at http://blog.learningtree.com/tag/ui/.