Paul Wilson's .NET Blog

I was one of the contributors to the ASP.NET Developer's Cookbook with Steve Smith and Rob Howard.  Today I finally got my free copy, only it was already opened and stolen off my front porch, or somewhere in route.  I've never had any stolen from my mail before, and this seemed like one of the least likely things to be taken.  My wife even had something come today that's probably more useful and expensive, but I guess the package didn't look as interesting.  The upside is that maybe someone around here will actually read it and start working with ASP.NET, and then later need my help to bail them out of a mess they get themselves into.  I can dream anyway.

One of my colleagues, Jerry Dennany, has finally got a blog here -- and it should be a good one.  Don't believe him for a minute when he describes himself as a "software configuration manager / release engineer".  That may be his title here at Roche, but Jerry is actually one of the smartest "developers" I know period, .NET or otherwise, even though he claims to not be a developer.  He has been one of the main guys on "my" team that answers my questions -- not the other way around!  Remoting, garbage collection, MSIL (or CIL), COM+ configuration, network admin -- this guy is your .NET and overall Microsoft internals resource guy!  I've been very fortunate to have worked with him for the past couple of years, and I know he will have some great posts here on some of the internals of .NET and other things Microsoft (or even Linux if we let him).  I see he's already correcting Microsoft's Suzanne Cook, from the CLR team, in his blog -- and he's right of course.  Also, check out his article on CodeProject on "TraceListeners and Reflection" -- definitely a very good read on tracing.

I found the following breaking change in the MSDN article "Development Impacts of Security Changes in Windows Server 2003" by Michael Howard:

I finally got my free PocketPC yesterday and its vastly superior to my Handspring Treo.  Of course, my Handspring Treo actually works with the Internet over the built-in cell phone, so it will continue to be my main PDA.  I suppose this is why so many other developers are already starting to put their free PocketPCs up on EBay!  I'm keeping mine to develop on though, and its amazing how easy it really is to do with VS.NET 2003 -- Microsoft definitely got this right.  There is already a project type for PocketPC development, and using it automatically lets you know what .NET functionality is valid for the PocketPC.  Setting a reference to the System.Data.SqlServerCe namespace even automatically installs SqlCE on the PocketPC when you deploy.  I haven't really done much yet, but what I have tried has been remarkably easy, so I plan on doing more.  So if you qualify for this free PocketPC you really need to sign up and get yours before the offer is no longer valid.

I got my 60-day notice earlier this week, so if anyone needs an ASP.NET guru, or .NET in general and other Microsoft technologies for that matter, I'll be available starting August 11.  I've known they were going to be closing our Atlanta office now for about a year and a half, so this wasn't a surprise, and I actually pushed for it to occur now instead of later this year, so I'm ready.  You can see my resume at http://Paul.WilsonDotnet.com if you are so inclined, otherwise just let me know if you have a lead.  I live in Atlanta, actually the North-Northwest side, and would like to stay here, although I'm willing to consider a few other things for the right opportunity.  I'm leaning towards consulting or contracts, although once again I'm open to permanent positions if they fit me and my location well.

I attended two sessions from the Patterns and Practices group this morning.  I was very impressed, not just with what they've done, but also with their honesty and openness.  They already documented a fair number of patterns and released several application blocks, like the Data Access Application Block, and more will be coming soon.  I'll have to reevaluate their work, and the community additions they've encouraged, since I wasn't real impressed the last time I checked.

I've got two sessions this morning before I head back to Atlanta this afternoon.  I struggled through two boring sessions yesterday afternoon, while finding one that was pretty good.  It was "Building Application Back-Ends Quickly" by Kate Gregory and Kenneth Spencer.  There was good discussion of design patterns and lots of examples of code generation.  They did not, however, mention Eric Smith's CodeSmith, which would beat their tools hands down.  They also did not mention O/R mappers, my latest interest, which promises to make lots of code obsolete, generated or not.  Last night was the big attendee party -- it featured SmashMouth and the WallFlowers.  Good music, but its just hard to get a bunch of techies, mostly male, to get real excited, so it seemed a little dull.

So last night I said "Real-World Tips for Real-World Web Applications" by Paul Sheriff and Ken Getz was the best ASP.NET session.  I was also asked why in the comments and I will try to explain that in a minute.  Today I went to two of Clemens Vasters sessions, "Building Distributed .NET Applications" and "Aspect-Oriented Programming", and these were also fabulous.  I will even have to go out on a limb and say Clemens' "Building Distributed .NET Applications" is the best overall session for me so far. I'm not saying its better than Paul and Ken's talk for everyone, but that depends on whether you need to learn more ASP.NET or not, and I'm simply past that point.  Clemons spent some time first talking about the difference between layers (logical separation) and tiers (physical separation), including some real cool pictures and a comparison to something Men In Black.  He showed a demo of one of his apps that works both disconnected and connected -- can't wait to get the code for this later on his blog.  Finally, there was a lot of distinctions made between the types of tiers, that being local (same machine), near (LAN), and far (everything else).  Local should use Remoting and take advantage of properties, inheritance, byref, events, and be synchronous.  Near should use Enterprise Services and take advantage of chunky calls, interfaces, byval, and may be synchronous.  Far should use Web Services and take advantage be asynchronous message based.  There were a lot more details, but I had never heard anyone say that Remoting is only best for same machine communication and once you step out into the LAN you should be using Enterprise Services.  I've heard hints of this before, including some stuff from Ingo himself at DevConnections, but this was right in your face this is the way to do it conclusion.  It sounded good to me, but I don't have much experience with these, so let me know your thoughts on this.