The Agileer

Hey, Doug! Where are you?

Doug Seelinger — Mon, 27 Oct 2008 16:22:48 GMT

If you've been wondering why I haven't posted lately, it's not because I've been lazy (at least not overly).

I've recently been hired by Microsoft and have been drinking from the infamous "Firehose".

I've also had the opportunity to host my blog on the MSDN site, which is where you'll be able to find the new content for me - http://blogs.msdn.com/agileer.

Hard-Coding Crazy?

Doug Seelinger — Mon, 15 Sep 2008 14:52:19 GMT

The current application that I'm working on is pretty configuration-heavy. There's even configuration to drive the configuration. There's configuration that drives the UI that drives the configuration.

Even so, when the product owner recently said, "I think we need to move some of this out of configuration and just hard-code it." Well... let's just say that I didn't agree. I do now.

I recently added a whole lot of flexibility by providing a "plug-in" configuration chooser enabled by... you guessed it - more configuration. We then deployed the application last week and my app worked nicely.

However, another application that consumes my code did not fair so well. It failed each time it was called because of the additional configuration. Yes, it was a "communication" gaff on my part, but it could have been avoided since we didn't really NEED extra configuration to achieve the plug-in capability. I could have just hard coded something to the affect of - in MODE1 do X, and in MODE2 do Y. That way we could have executed just fine in production without failure.

Ayende has similar feelings here.

Red, Green, Refactor (then Integrate and Commit)

Doug Seelinger — Tue, 09 Sep 2008 15:14:50 GMT

xDD (TDD, BDD, DDD, etc.) emphasizes the standard "Red, Green, Refactor" mantra in order to prevent the buildup of technical debt. However, without the addition of Integration testing, or user-acceptance testing, and then committing the code to source control, you're only partway there. I'm not saying that cycle should execute all five every time (RGRIC, RGRIC, RGRIC), but it should definitely happen when a feature/story has been completed at the very least (perhaps RGR1, RGR2, ...RGRn, IC). Don't forget that Integrate & Commit or you will end up with some very heavy technical debt rather quickly.

I'm writing this down in order to remind myself. I've also placed it in my desktop background via BgInfo.

A couple of Unit Testing Frameworks for .NET

Doug Seelinger — Tue, 12 Aug 2008 06:35:00 GMT

(archived from my old blog from my pre-MS days)

One of the great things about ALT.NET is the variety of choices it affords. However, this can also become an albatross around your neck when trying to choose a particular unit testing framework, mocking framework, IoC container, etc. I mean, if you haven't tried most if not all of the choices, then you don't know what you're missing. Of course you can read about the different options out there, but that's no substitute for trying them out for yourself. Having said that, there's still a lot of utility in reading about them to exploit them to your advantage. Thus, I recently moved out of my normal comfort-zone, (NUnit, Rhino.Mocks, Windsor Container) and tried some other technologies.

What follows is a short description of my impression of a couple of Unit Testing frameworks that I've played with in the past couple of weeks. Note that this is not intended to be an complete portrayal of each tool, but rather how I felt when attempting to use them.

NUnit

This is the old standby, but there's a 2.5 alpha that I hadn't tried. I liked a lot of the changes that have gone into it, but found that too many things that I depend on for integrating with NUnit don't work with 2.5 yet.

XUnit

I like XUnit. It's simple. ALT.NET takes a bad rap for being hard to adopt, so one thing I try to emphasize is simplicity. This is XUnit's strength. If I were starting someone off in Unit Testing, I'd tell them to look at XUnit. I have to admit that I do like the Constraint model in NUnit, however. It just reads more naturally. Also, pretty much everything works with NUnit, but not necessarily with XUnit (like TeamCity, currently). So eventually, I went back to NUnit 2.4.8. It worked fine with TestDriven.NET, however.

So the moral of the story is that software (including OSS) has momentum. Once you get used to a certain way of development, it feels painful to switch, even if the other applications have some nice new features.

Conspicuously missing here is MbUnit, which I've tried in the past but again went back to NUnit eventually. I'll have to try it out again soon.

Generic Type names

Doug Seelinger — Tue, 29 Jul 2008 17:28:00 GMT

The type name that you need to use in config files, such as for IoC containers, can be quite complex. For me, it's easier just to spike them than to try to remember the format gobbledy-gook.

public class GenericsTypeSpikes
{
    [Fact]
    public void The_name_of_a_complex_type_can_be_found_in_Type_AssemblyQualifiedName()
    {
        Type t = typeof(IList<ISource>);

        Assert.Equal("System.Collections.Generic.IList`1[[Blah.Core.ISource, Blah.Core, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null]], mscorlib, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089", t.AssemblyQualifiedName);
    }
}

NDepend

Doug Seelinger — Tue, 08 Jul 2008 00:14:25 GMT

The first time I heard of NDepend was on Hanselminutes (show here). It sounded intriguing, but at the same time the show went into a lot of detail and used a lot of terminology that I wasn't familiar with, even with a computer science degree. So it was with a little trepidation that I downloaded a copy and ran it against an application that I was working on at the time.

To be honest, a bit of information overload was involved. I could see that NDepend had gathered a lot of statistics regarding the code base, but I had no idea how to use it. I didn't look at NDepend again for a few months.

The next time I saw NDepend was when I was looking at CI-Factory last year. It was included as one of the tools in the continuous integration mix. Same experience - information overload.

Just last week or so I took one more look at NDepend to see if it could help out with the code base I had inherited at work. I had many places that needed refactoring, but it was hard to tell where to start. My initial (3-minutes) reaction was the same - too much information after the analysis was complete. I didn't know where to start. But this time I stuck it out for more than 5 minutes - and that was all it took for me to fall in love (this time) with NDepend. NDepend is largely based on "CQL" (Code Query Language), that sorta-kinda looks like SQL, but goes against code instead of a database (hmm... "LINQ to code" can't be very far off). NDepend has several canned CQL statements that it runs as part of the analysis to help identify coding faux-pas, and the first query, with the description of "Quick summary of methods to refactor" was a treasure-trove of bad juju. It identified methods that were too big, too complex, too... just about anything that can be quickly called out, and it picked the worst offenders to show.

NDepend showed me where the worst code was first. That's what I got for sticking it out this time. Now there are many, many more features that I have hardly scratched the surface of, and I hope to be able to cover those in later blog posts. For now, just try out http://www.ndepend.com and check out your own FBI-most-wanted list.

Is "if/else" better than "if/return"? If you want to easily refactor, then yes.

Doug Seelinger — Wed, 02 Jul 2008 02:10:01 GMT

Look at the following piece of code:

private void DbMove(CopyInfo info)
{
    string targetDb = GetTargetDB(info);
    if (targetDb.Equals(string.Empty))
        return;

PrepDbsForCopy(info.SourceDb, targetDb);

    GetAllData(info);
    if (IsNoSourceData())
        return;

int upperBound = GetUpperBound();
int rangeCount = GetRange();

if (CopyNotCurrentlyPossible(_isDisconnected, _isHomogenous))
return;

DoAsyncDbCopy(info, upperBound, rangeCount, _isDisconnected, _isHomogenous);
}

Now this is a standard piece of code with several "short-circuit" returns in the code. Short-circuiting is good, but usually only if it appears as the first instruction in the method, and only once. Otherwise, refactoring the code (as above) is a bear. Select a section of text to pull out (extract method), and if it contains a "return" or break, or anything like that, then you'll most likely get a complaint from your refactoring tool. ReSharper 4 complains "Extracted block has exits to different points Proceed with refactoring?(sic)". Besides the pain from lack of proper punctuation (just a typo, I'm sure), the pain of not being able to refactor when you want to is even worse.

However, you can refactor if you do some manual refactoring (and this is still easier than manually extracting the method). Instead of:

    string targetDb = GetTargetDB(info);
    if (targetDb.Equals(string.Empty))
        return;

...

Try the following:

    string targetDb = GetTargetDB(info);
    if (targetDb.Equals(string.Empty))
        return;
    else
    {
        ...
    }

Then you can start using your refactoring tools like "Invert If"

    string targetDb = GetTargetDB(info);
    if (!targetDb.Equals(string.Empty))
    {
        ...
    }
    else
        return;

And "voila," your else can be tossed as cruft. Do so with the other "if"s and yes, you have more indented blocks of code, but indented blocks of code can be easily refactored with "Extract Method."

Deleting Objects from Active Directory

Doug Seelinger — Thu, 19 Jun 2008 02:33:22 GMT

Active Directory is funny.

Or should I say "permissions in Active Directory are funny."

I find that whenever I'm "stuck" on a problem for a good chunk of the day (AD or not) , about 80% of the time it's a permissions issue. A couple of days ago I was stuck on an issue that I knew was a permissions issue, but I couldn't figure out why. I mean, here I was, supposedly something of an development expert in directory services, but I couldn't get my DirectoryServices code to delete a simple AD object. I could create objects, but I couldn't delete them.

Now, at this particular client they've tightened down the screws in their environment substantially, so simply increasing my permissions willy-nilly was out of the question. No domain admin privileges here. I'm not even sure any of the developers on this project even use Active Directory Users & Computers, they seem to only use home-grown LDAP tools for examining AD. They are able to delete items with one particular home-grown tool, but it was written in VB - in all it's COM, late-binding, who-knows-what's-really-lurking-under-this-IUnknown-interface glory.

Has anyone ever tried to peer into the ObjectSecurity class? It's not for the faint of heart. And it wasn't something that I was going to be able to decipher in an afternoon. So I downloaded the old Windows 2003 admin pack, installed it on my XP development box, and went into all the gory details that are AD ACLs.

The question was, why could the old VB code delete the object when my DirectoryServices code not do it? According to AD Users & Computers, the credentials I was using did not have "delete" permission on the object. I knew that. That's what "access denied" usually means. How could I explain this to the fellow who was successfully deleting with the VB code and expected me to do the same? DirectoryEntry.DeleteTree() simply would not work.

However, upon further examination, I could see that the OU that I was dealing with did allow my credentials to create and delete child objects. Hmm... One more quick test confirmed it. Although DirectoryEntry.DeleteTree() failed, with the OU I could do DirectoryEntry.Children.Remove(objectInQuestion). That's just crazy. Boggles the mind. But it works.

Of course, if the AD admins want to save another developer further grief (and there's no reason to believe they would - I know some sadistic infrastructure folks), then they'll just add the delete permission and avoid further confusion.

Starting out with Agile

Doug Seelinger — Thu, 12 Jun 2008 02:03:09 GMT

I heard recently on Hanselminutes (115, I think) something to the effect of "Alt.NET has a lot to contribute, but the barrier to entry is pretty high." I know that didn't necessarily mean "agile" in particular, but perhaps that was actually the intent.

Is it hard to "do agile"? I'm sure there would be plenty of voices on both sides, but I think there could be more effort put into making it easier or at least better explained. Sure all the "sexy" agile info is out there on the cutting edge, but there are plenty of folks who still need the basics.

My own knowledge of agile principles & practices came over a long period of time. I didn't have the advantage of having a coach. In fact, I remember reading something on MSDN around the 2000-2002 timeframe in which Anders Hejlsberg was writing about good coding practices, and in particular he mentioned how he used NUnit for his unit tests.

Unit tests? I thought that was like, run the application. If it doesn't blow up it was a successful unit-test; one scant step beyond compiling. So I didn't learn agile overnight. Even now I continually tweak my practices in an effort to learn the best way to develop the agile way. I'll probably continue to do that as long as I write code.

After thinking about my own experience, I believe that testing is probably the best way to introduce folks to agile concepts. I mean, if it worked for me without a coach (and for the longest time without blogs or many other agile resources), then perhaps testing is a great wedge for breaking folks thought processes into the agile arena.

So start testing.

While I Was Sleeping...

Doug Seelinger — Thu, 05 Jun 2008 03:42:55 GMT

It's incredible how fast things can change if you turn off your developer technology radar for a few months. I was pretty comfortable with my knowledge of what was out there around fall of '07. I'd explored pretty much all the different tools and technologies that had tickled my fancy and even though I wasn't an expert in all of them, I knew that I could pick any one of them up pretty quickly if need be. So down went my radar for awhile.

Holy cow.

Apparently not everyone has been as idle as I have been. Here is a list that I compiled in a single afternoon of some things (I'm sure there are plenty more) that I'd like to try out in the not-too-distant-future to see if they can assist me in my daily chores. What follows is not anything near a full review (basically I've pretty much only heard of them at this point), so my descriptions of what they are may be WAY off:

Gallio	Unit testing framework agnostic unit test runner
Git	Successor to Subversion???
Team City	Continuous Integration competitor to CruiseControl.Net from the makers of ReSharper. And, yes, it's free.
XUnit	The "opinionated" unit testing framework
Pex	Continuous testing
Spec#	C# with built-in specifications - gives the user of a framework a better idea about how a class/method will behave.
ASP.NET 3.5 Dynamic Data	Automatic CRUD that may actually work beyond simple scenarios

How about you? What kind of technologies have you excited right now?

Testing Changes to Untested Code

Doug Seelinger — Tue, 03 Jun 2008 21:54:00 GMT

Let's face it, not everyone has automated unit tests for their code. You may, however, have inherited some code and are responsible for making changes to that code or for creating additional functionality. How do you handle such a situation?

The best way is to find places in the code into which you can place a pry-bar (not a technical term) and wedge a couple of components apart. This is the best way, but it's also messy.

Pry-Bar Method

I believe this comes from Working Effectively with Legacy Code (Robert C. Martin Series), by Michael Feathers, though I haven't read the book personally (I need to get myself a copy). But I do understand the concept.

Let's say that you have a class Engine, which contains a class Carburetor. Let's also say that you have to modify some functionality of Engine. What you need to do is isolate Engine so that it can be tested all by itself. We don't care about the Carburetor class at this time - it's to be tested elsewhere (eventually). Here are the appropriate steps to "pry away" Engine's dependency on the Carburetor class:

If it doesn't already exist, create an interface for Carburetor's functionality, ICarburetor, such that ICarburetor contains at least all the methods/properties that are called by the Engine class. Refactoring tools like JetBrains ReSharper make this very easy.
Modify Engine's constructor so that it accepts an ICarburetor parameter and set a private reference to an ICarburetor object(there are variations on this theme, but this is my favorite).
In your test, create a mock object of the ICarburetor type, either "manually" or by using a mock framework like Rhino Mocks.
Set up any expectations of the mock object if required (that's a future post, or just Google Rhino Mocks).
Pass your ICarburetor mock object into your Engine constructor.
Do your test on Engine.
Confirm your expected state (and/or interactions with the mock object)

I hear what you're saying, "That's an awful lot of work for testing some code that's not been tested before!" Yep. However, if you ever intend to get the whole system under proper testing, then you've got to do it that way. What makes it more painful is when you don't have only one "dependency" (ICarburetor) but perhaps dozens (Carburetor, shocks, exhaust, fuel tank, oil, transmission, etc.). Whew!!!

Sometimes you may only be responsible for making a change to code ASAP and then go away. In this case, you're not likely to lose sleep over the fact that someone has written some untested code (gasp!). All you're responsible for is a small tweak. What do you do in this situation?

Lazy Method (perhaps too lazy)

For some reason, I find myself in this situation more often than not, perhaps since as consultants we're not often called in after the fact for long-term code maintenance. You certainly could use the Pry-Bar method above, but it may involve creating a lot of "interface noise" (at least the client sees it that way) when all the client wanted was the web-page widget to appear blue instead of red (ok, that's a little extreme, but you get my point). So what do we do? In this case, I'll sometimes resort to the "lazy" method. In this case, I really do mean lazy. In agile terminology, being lazy is not always a bad thing. However, in this case "lazy" may not be such a good thing, but hey - you do what you can to make people happy.

So here you simply create a class (these classes tend to "want to be" singletons, but don't have to be) that take data from the "untested" code and return data to the untested code (or don't in the case of "void"). You can test the "lazy class" all you want this way. Again, not the best way, but perhaps the quickest for short-fuse work.

Whichever option you choose, it's better than not testing your code at all, which is probably why they called you in for help in the first place.

Selling Agile (testing)

Doug Seelinger — Thu, 08 May 2008 06:47:00 GMT

(archived from my old blog from my pre-MS days)

It's amazing. The battle for agility is far from won. Folks are still not writing unit tests. Some folks don't even know what unit tests are, or have a badly mangled definition of them. I'm not only finding this as a consultant at client sites, but among some of my own colleagues as well. A developer asked me last week, "What is this agile thing?" I seriously thought that by now people would have at least a basic understanding of what it is. But alas, it not currently so. This isn't to castigate anyone. I'm more than happy to help folks out with "this agile thing".

So what is Agile? I point you towards the Agile Software Development Manifesto, and some further explanation, and the "original wiki" houses a great deal of info on agile. There is much info from people more eloquent than I. Those who don't know what "agile" is usually know what "Google" is. Use your search engine of choice and explore. But back to testing...

Test your doggone code, people!

Why? Because I like to sleep at night. That means that not only do I have to test my own code, but if I depend on your code, you better be testing yours. What? I don't take your word that you've done an excellent job as a developer? Frankly, no. I don't trust me, and IMHO I consider myself somewhat more than competent as a developer. Therefore, if I don't trust myself, why in the world should I trust you? I read somewhere (I think it was in Ron Jeffries' book, Extreme Programming Adventures in C#) that without testing, the average developer could only write about five lines of code before introducing a bug. My own folly has born this out in the past.

So if you don't test your code and I'm left to work with it, then I have to find the seams within your code and crack it open to wedge in support for testing, as found in Working Effectively with Legacy Code, by Michael Feathers.

It's all about confidence

Before I embraced the agile lifestyle, I never really knew if my code was doing what I thought it was supposed to do. I hoped and prayed that it would. And really, at that point, prayer was my only option, because I didn't know whether the testers (when we had them) were doing an adequate job. What did we do when we didn't have testers? Well, testing turned out to be the weeks and months after going live on a production environment - not the best time to do your initial testing. Demos and integration testing were high stress times for me. I didn't even want to be around when they happened. I don't enjoy getting yelled at (or looked down upon as incompetent) any more than the next person.

So I took the time to learn about unit testing. NUnit was the only thing available to me as a .NET developer at the time and it seemed to work well. Over the years I started with writing unit test - bad unit tests to be sure, but tests nonetheless. I recently had someone tell me that the code that they had written was too difficult to test and thus it was better to not have any tests at all. I fail to see the logic in that (besides the fact that there are really very few situations that cannot be adequately unit tested if designed correctly in the first place). Bad unit tests are SOOOOO.... much better than no unit tests. Even if your unit tests are brittle and return false-positive results, at the very least you are thinking further about how your code will be utilized. It could be argued that you can have a false sense of security with badly written tests, and while that my be true, what, if not a false sense of security, is someone operating under when they have NO tests?

Now, with proper testing in place, I have no problem ripping out huge sections of code and replacing them when requirements change. Why? With the press of a key I can run all my tests and confirm that yes, all is well. That for which I am responsible... just works.

In my organization we're going to start trying to bring all the developers up a notch or two. If you write tests then that's the beginning - the very beginning. There is so much more to understand about improving the quality of code. But testing is the very first thing. If you're not testing your own code - just stop. Turn around, and take some responsibility for quality without cowering behind some large object when your code goes live to the public.

Coding Blind

Doug Seelinger — Fri, 25 Apr 2008 17:42:44 GMT

I remember a cartoon I saw in the early 90's regarding the way that software projects were sometimes run. The punch line was, "You guys start coding, I'll go find out what they want."

The industry's answer to that was the onset of "Big Requirements Up Front" - (BRUF). Waterfall. RUP. UML full-trip code generation/regeneration - and like most things, it was something of a pendulum swing - reacting to the cowboy coders with a methodology that prevented writing any code that was not useful. Get all those requirements up front - they'll never change. Implied in the "They'll never change" was that any analyst worth their salt would make sure that you're asking the right questions, then requirements never change. Requirements only change when you've gotten the wrong requirements.

But requirements do change, even if you've done a good job up front. No one has all the info. Decision makers don't always have the right info. Things change. And in RUP - if you were doing it "right" you had oodles of documents to modify if requirements ever "changed." Not to mention the fact that you're most likely going to hit your client up for more money - they always love to hear that. Change requests anyone?

In reaction to that, agile & XP arose. Power to the Programmers! But there are those who have taken that pendulum swing back too far as well. We don't need no stinking Architects! The problem with this approach is that you ignore resources that are already available here, sometimes called the "Not Invented Here" syndrome. I confess that I've been guilty of this. I love to code. I'd rather be coding than analyzing any day. It's not fair to the customer, however, to ignore the groundwork that other's have done in order to get my coding fix. A few months ago a client had a project where they wanted to provision all accounts from a single request "front end." I automatically thought "new web site," and for a long time ignored the fact that they already had a MOSS site available internally. These account requests fit fine within the paradigm of a SharePoint list. I almost went with a custom web site in spite of that, but in the end, I could not justify the fact that I'd spend a lot more on this custom web site than I would with MOSS. In the end, I did get to do a good bit of coding, since we had to implement a custom MIIS extensible management agent to get into and out of MOSS. I did end up getting my coding fix.

So the moral of the story is - don't assume that you're going to do it better than anyone else who's already done the same. If you have the opportunity to exploit the collateral of others, do so where it makes sense.

The Wonder of GO

Doug Seelinger — Fri, 28 Mar 2008 20:20:09 GMT

Note: There is tech content in here, I promise!

I really suck at math. Some may think that's odd for a "computer guy," but it's true. But perhaps it's because of this suckiness that keeps me coming back to it - the challenge of it. In college, I didn't have a problem with computer science classes (well, except for numerical analysis - math) since I tend to have a pretty linear/logical thought process (or so I wish to believe, perhaps). I knew in 6th grade (around 1977) what I wanted to do - computer programming. The fact that I've stuck with it for so long is unusual, given the number of friends that changed their majors, multiple times, during just the four years of college. Probably only 25% of those people are doing anything even remotely close to what their degree was in. Again, I'm the unusual one. So I've been the "logical" one, the steady one, for quite awhile.

Surprise, surprise - I was in the middle school and high-school chess clubs (in addition to football, basketball, track & field, so don't start stereotyping me, thank you very much). My dad, a West Virginia coal miner at the time, taught me how to play chess when I was in elementary school. He says he doesn't remember, but it's true. My best friend in middle school & high-school, and he was my best man as well, was also on those chess teams - another "logical guy," whom we thought of as our own Bobby Fischer. I went to a several tournaments, but in high-school I caved under pressure, and quit competing in tournaments. I don't remember my USCF rating, that was almost 30 years ago. I won some and I lost some. I've continued to play throughout the years, and have taught my kids how to play (no Bobby Fischers here, so far). Bottom line - chess is fun and challenging. So chess is about the only traditional board game I ever "got into."

I have another confession: I like to watch anime with my kids (heck, I like to watch it alone, too. There, I said it). I tend towards the Shonen variety, even though I have mostly daughters - Sailor Moon gives me the heebee-geebees. So about a month ago I started watching one called "Hikaru No Go" (Hikaru's Go - see Wikipedia). The plot was pretty clear after a couple of episodes. The story centers around a kid (Hikaru) and his entrance into the world of "GO", an ancient Asian board game.

Now, I may have heard of Go before, but if so, I don't remember. I get the same reaction from all my American family & friends - "What's Go?" In China, Korea & Japan Go is "the bomb." They follow it the way Russians follow the chess world, which is WAY more than Americans do (see Searching for Bobby Fischer: The Father of a Prodigy Observes the World of Chess).

I was unprepared for how interesting Hikaru No Go would be. Perhaps it was my ties to traditional board games, but I was really sucked into the story line. It was also refreshing that it's neither dark nor was their a true "bad guy" in the traditional sense. It still has the same basic motivation that many Shonen storylines have, which is "try your best and get better and better at what you do."

About halfway through the series, I actually started to get interested in the game itself. I thought, "Hey, if it's that popular in east Asia, maybe there's something to it. I saw a few decent introductions online, but I really wanted a book. I found the ONLY book on Go available in the town of Winchester, VA - Teach Yourself Go. Now that I've bought that copy, there are no more books on Go available in Winchester - I've looked. The book is very dense. You CAN learn Go from it, but it's like learning math from the textbook alone - heaven help you if you miss a concept.

One thing that this book attempts to explain is why there aren't very many computerized Go "engines," and why the existing ones aren't very strong. The author believes that it's much more difficult to build a Go engine than, say, a chess engine, which now can pretty much beat any world champion. There aren't any Go programs that can even beat strong club-level players, much less pro Go players (and yes, there are professional Go players).

Seriously? No one can write a strong Go engine? I thought that was hard to believe - until I started to play go. I'm the one reading the book. I've taught my children just the rules - and they beat the living daylights out of me. Not just once in awhile, but about 80% of the time! There is no randomness involved, no dice or anything. I thought that I just didn't know enough (and I don't) and that a little more studying would help. So I went to this section that tries to explain how to capture stones. Again, this book is a dense, textbook style rather than conversational (which I prefer any day of the week), and I played through the examples, and see this concept of a "net" on the edge of the board. I see it WORKING, but I can't say WHY it works. It's almost anti-logical! I was up at midnight last night raving like a lunatic about how I couldn't see WHY this "net" thing would work (my wife and oldest daughter can confirm that I was on the edge of lunacy).

So I'm only about 1/8 of the way through Teach Yourself Go and already I see a concept that I have NO IDEA how you would code, since why it works is beyond me. My wife and oldest daughter tell me that I don't need to know why it works, only that it does. That is NOT good enough for me. It seems like there should be an easy explanation, but I don't see it. Not that it couldn't be done EVER, but I have a feeling that you'd have to be a 9-dan Go pro AND a sharp software developer in order to write a strong Go engine. I have a feeling that I know how many of those are around - none.

That doesn't mean that I'm not going to try, of course. It looks like the perfect opportunity to get into Silverlight or WPF.

It's this question of "Why does this work!?!?!?" that is both driving me crazy and fueled my desire to learn this game, as well as some of the "science" involved here. I'm looking forward to an off-site long-term project with a Chinese colleague who has promised (whether he knows it or not) to become my own personal Go tutor.

Come to think of it, the other colleague of this three man team is Russian. Hmm... maybe I'll bring my chess set, along with the Go stones.

"Heroes Happen Here" Notes part 2

Doug Seelinger — Thu, 13 Mar 2008 18:07:42 GMT

Yes, I know. I have a weird definition of "tomorrow." Here are my notes from last Thursday's HHH event in Harrisburg, PA:

Session 2: Office Development

Zhiming Xue http://blogs.msdn.com/zxue

Talked a bit bout how VSTO isn't an extra in VS 2008, but that it's built in.

One thing I noticed about each of the presenters - they need to discover the Consolas font! Scott Guthrie uses it, should be good enough for these guys.

He had an interesting use of template expansion. Before the demo he created some snippets and bound them to keywords inVS (like you do with "ctor" to expand a constructor). So he'd type something like snip1, then "tab" and get his snippet inserted. I don't think I've seen that one before, but maybe I'm behind on the power curve.

I noticed that he spent a good while demoing the "auto include imports/using" feature, which has been in VS for three years now.

Showed using WPF inside an office Panel in word, which was cool. It took me awhile to realize that it WAS WPF - just goes to show what we "expect" from software nowadays.

Session 3: WPF & C++ (huh?)

This session surprised me. I really liked the first half-hour. Not the most authoritative speaker, I could almost feel the BS oozing. Yes, WPF can do some 3D things (on Vista), but that's not its strength. WPF isn't challenging any 3D game engines out there, yet. But the demo was really nice. I almost had an epiphany... almost. But I'm still not sure where I'd use WPF in a LoB app. I had a similar experience but GOT the epiphany back at Tech Ed 97, where I finally understood how the Web could be used as an application platform (there really weren't too many of them around back in early 97).

The last half-hour was a total waste. Completely unconnected to WPF, the speaker began on new features of C++ in VS 2008. At least I think they were new features. I haven't used C++ in about 10 years or so I can't really say. Another famous quote from 97 Tech Ed (from Don Box), "I used to code in C++, but then I had a family and had to go home at night."

So that's about it. There were supposed to be four sessions but one was either combined with the others or just left out (a little confusing). It was supposed to be about SOwhAt (my term for SOA), but I guess it just didn't happen. I don't have anything against web services, but I just don't see the big advantage of WCF and w-death-*.

"Heroes Happen Here" Notes

Doug Seelinger — Sat, 08 Mar 2008 04:16:36 GMT

I attended the "Heroes Happen Here" event in Harrisburg, PA yesterday. Not a bad trip overall - the best takeaway was a free copy of Vista Ultimate and Windows Server 2008 (installing now for the first time), though this one is just a one year evaluation license (with a license key). They only included Visual Studio 2008 Standard, which is pretty useless to me since we already have Team Dev, with the promise of Team Suite in the near future. I also got a copy of SQL Server 2008, which I might install some day, but honestly, I don't know what big advantage that it holds right now, especially since they didn't even mention SQL 2008 in the Dev Track for the launch event for... what was it? Oh yeah, Visual Studio 2008 (released 4 months ago), Windows Server 2008 and SQL Server 2008!

Stuff I Learned:

PA Interstates around Harrisburg are dangerous.
Registration was one hour later than I thought. I thought my track started at noon, but that was just registration. Turned out to be a good thing since I was lost on the PA Interstate system.
The name tag they give you to stick on your shirt was just large enough to cover the blue printer-ink stain I got on my shirt 5 minutes before I left home for the event.
Microsoft RoundTable was very cool (in the Partner Expo area). It shows a panoramic view of an entire room for video conferencing - set it in the middle of a conference table and it will show both a panoramic view of your room in Live Meeting, but also track the current person speaking (whoever's loudest, basically). Runs about $3K, but much cheaper than others in its space (according to the blue-shirt I was speaking with).
Overall, the Partner Expo was very small, with about 1/2 of the exhibits staffed by blue-shirts (Round Table, Windows Mobile, etc.)

Dev Track Keynote:

5 minute video - complete waste of time. 1.5 hours may be too long for a keynote, but 5 minutes is too short. I was not inspired, which is the whole point of a keynote, right?

Dev Track Session 1 - David Solomon davidsol@microsoft.com:

This was my favorite session. Good info, good speaker.
Lots of emphasis on RIA (aka Silverlight) and collaboration. This year "collaboration" means devs collaborating with Expressions designers. No one mentioned the fact that Expression lacks collaboration features itself - like any kind of communication with Team Foundation Server.
Mort wants more wizards, apparently.
VS has better CSS & Javascript capabilities now, though I think FireBug has a better debugger, still.
Still pushing SOA - "SOA is important" without any reason why in the talk.
My random thought while watching a demo - why doesn't the HTML editor automatically insert quotes in your attributes after you type - attrName= ?
I didn't realize the Comment/Uncomment toolbar button works in the HTML and XML editors!
Lots of interesting things regarding the AJAX extender controls and validators, though I'm still not convinced ASP.NET AJAX is the right way to go.
Integrated Code Metrics in VS 2008 looks cool - Cyclometric Complexity, Maintainability, Coupling, etc.

Notes on Session 2 & 3 tomorrow.

ScottGu and his faithful Blog

Doug Seelinger — Fri, 29 Feb 2008 16:56:01 GMT

If you want to learn a new Microsoft developer tool or technology, what's the first thing you turn to? MSDN? Books? Google? My blog (well, of course you do, but besides that)?

Me? I'm starting to learn to really appreciate the value of what gets posted on ScottGu's blog (Corporate VP of Microsoft's Developer Division). ASP.NET MVC Framework? I learned about it there. jQuery Intellisense and Ajax? I learned about it there. Silverlight? Well, if I wanted to know more about it (and I will eventually), I can learn about it from there, too.

I've subscribed to his blog for a good while, but it's almost better to look at it not as a subscription, but as a big reference library of cool, Microsoft developer... stuff. Subscribing and seeing the posts come out one at a time is almost anticlimactic, somehow.

Oh, and if you don't find what you need from Scott, you can always subscribe to my blog too: Feed your aggregator (RSS 2.0)

ReSharper 4.0 & LINQ

Doug Seelinger — Thu, 28 Feb 2008 22:34:25 GMT

Unfortunately, as of today (2/28/08), the ReSharper 4.0 EAP still can't handle LINQ (or partial methods, apparently). However, I just can't give up ReSharper's refactoring capabilities and other goodness, so here's how you can disable the code analysis and IntelliSense in the ReSharper -> Tools menu dialog:

If you don't use LINQ (or partial methods), just ignore this advice - use ReSharper to the hilt!

If you're not yet using ReSharper for VS 2008, I'd still advise that you download the latest nightly build and give it a try. You'll not want to go back.

ASP.NET MVC Framework, Ajax, & jQuery

Doug Seelinger — Wed, 27 Feb 2008 19:17:29 GMT

I've seen a few examples trying to show how to combine the MVC Framework and Ajax. There are even a couple that show you how to use jQuery with MVC. However, they sometimes leave out a couple of details - like properly passing parameters to the Ajax controller and returning information from the controller - both are trivial if you've done it before, but if not... well, an example helps.

Let's have a view/page that insults you - that's always fun. As a bonus, you're insulted via Ajax using jQuery to call an MVC Framework controller. Let's start with the Controller, as all good MVC practitioners should:

namespace JqueryTest1.Controllers
{
    using System.Web.Mvc;

    public class HomeController : Controller
    {
        [ControllerAction]
        public void Index() { RenderView("InsultMe"); }

        [ControllerAction]
        public void InsultMe() { RenderView("InsultMe"); }

        [ControllerAction]
        public
void InsultViaAjax(string who, string what)  { // normally we'd return something
more sophisticated than plain text, // but, hey! It's just an example. Response.Write("Your
" + who + " wears " + what + "!" );  } } }

It's pretty simple - a couple of controller actions to handle the default view and the "Insult Me" view, which are the same thing in this instance, and an "InsultViaAjax" action that takes its parameters from the Request["xxx"] collections - very nice. Then my simple Ajax controller puts together an original insult from the parameters and writes it back out to the response stream.

Next we have the Site.Master page, which we've just modified to contain the references to the jQuery scripts. In order to have some useful intellisense we add a second reference to a "dummy" jQuery script, which contains only the information required for jQuery intellisense, basically empty shells with XML documentation comments. You can get the jQuery intellisense script from here.

<%@ Master Language="C#" AutoEventWireup="true" CodeBehind="Site.master.cs" 
    Inherits="JqueryTest1.Views.Layouts.Site" %>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" 
    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head runat="server">
    <title>My Sample MVC Application</title>
    <link href="../../Content/Site.css" rel="stylesheet" type="text/css"  />


                    <script src="../../Content/javascript/jquery-1.2.3.js"
type="text/javascript" /> <!-- this is a little trickery to get intellisense
to work well for jQuery --> <asp:Placeholder runat="server" visible="false">
<script src="../../Content/javascript/jquery-1.2.3-intellisense.js" type="text/javascript"
/> </asp:Placeholder>

</head>
<body>
    <div id="inner">
    
        <div id="header">
            <h1>An Insulting Sample MVC Application</h1>
        </div>
        
        <div id="maincontent">
            <asp:ContentPlaceHolder ID="MainContentPlaceHolder" runat="server">
            </asp:ContentPlaceHolder>
        </div>
   
   </div>
</body>
</html>

And finally the view itself, InsultMe.aspx.

<%@ Page Language="C#" MasterPageFile="~/Views/Shared/Site.Master" 
    AutoEventWireup="true" CodeBehind="InsultMe.aspx.cs" 
    Inherits="JqueryTest1.Views.Home.InsultMe" Title="Insulting Example" %>
<asp:Content ID="Content1" ContentPlaceHolderID="MainContentPlaceHolder" 
    runat="server">
<form id="insult-form" action="NA">
    <fieldset>
        <legend>Insult Me</legend>
        <label for="who">Your relation:</label>
        <input type="text" id="who" name="who" value="mamma" />
        <br />
        <label for="what">Footwear:</label>
        <input type="text" id="what" name="what" value="combat boots" />
        <br />
        <input type="submit" id="insult-me" name="insult-me" 
            value="Insult me now" />
    </fieldset>
</form>

<script type="text/javascript">
 $(document).ready(function
() { // when the document is all ready, $("#insult-form").submit(function()
{ // hijack the form's submission,  $.post( // and replace it with
an Ajax call '/home/InsultViaAjax', $("#insult-form").serialize(),
function(returnVal) { alert('Here it comes: ' + returnVal); }); return false;
// stop the actual html form submission }); // a little eye-candy to hide the form
until the ajax call is complete jQuery().ajaxStart(function() { $("#insult-form").fadeOut("slow");
}); jQuery().ajaxStop(function() { $("#insult-form").fadeIn("fast");
});  }); </script> </asp:Content>

Do you want to watch it all happening? Love those morbid details? Then check out FireBug, an add-on to FireFox that makes it easy to set breakpoints and step through the JavaScript code.

Technorati Tags: jQuery,Ajax,ASP.NET

Choosing a JavaScript Library

Doug Seelinger — Tue, 26 Feb 2008 15:14:23 GMT

Up until recently, I had pretty much ignored JavaScript. The last time I remember using it heavily was back in the old classic ASP days. But when ASP.NET WebForms came out, it handled all the nasty client-side stuff for you, so I was free to "think elsewhere". However, with the advent of Ruby on Rails, Monorail, and the ASP.NET MVC Framework, people are wondering if the WebForms paradigm (attempting to design a web-app the same as a stateful thick-client) was such a good way to go after all.

One side-effect of ditching WebForms is that you lose all those lovely WebControls - DatePickers, DataGrids, Wizards - even TextBoxes and Buttons. What did we use before TextBoxes and Buttons? Remember <input />? Yeah, it really has been awhile. At first blush, it would seem like switching from WebForms is a non-starter - that it would just be too complicated to do things "the old-fashioned way." And if that were the case, it would be a correct assumption.

However, while we were all in our own WebForms-induced sleep, the rest of the world went on. They did not have WebControls, and they didn't like working "the old-fashioned way" any more than we did. So what did they do? Up until about 2005, I'm not really sure. But starting around 2005, some general purpose JavaScript libraries began to emerge that were truly helpful. Prototype, jQuery, YUI, and Dojo in particular are browser-only (no server requirements), open-source libraries. I need to decide which one that I'm going to commit to, and I've found an excellent set of online slides describing the differences between the available libraries out there.

What about ASP.NET Ajax (Atlas)? Well, one problem is that it's tied to both the client and server for full functionality. It's also built mainly for WebControls, which aren't available in MVC context (no viewstate or postbacks). I understand that there is a JavaScript library supporting the client-side, but I haven't yet found anyone who's using it exclusive of server-side support, and certainly none that would defend it against something like Prototype or jQuery.

So I think I'll go with jQuery for now. It seems to have the most concise "syntax", has intellisense and ajax support (of course) in VisualStudio 2008, Granted, the slides I mentioned previously were put together by one of the authors of jQuery, but it still looks impressive... or maybe I just like jQuery's logo...
it looks like Devo.

STSDEV & Visual Studio 2008

Doug Seelinger — Fri, 22 Feb 2008 03:10:11 GMT

STSDEV is a code generation tool on CodePlex for SharePoint, kind of like my PoShGen, except much more polished. It generates a starting project for several different kinds of scenarios - SharePoint Solutions, Features, & Web Parts. I tried to bring it up on a box with Visual Studio 2008, however, and the app crashed. It was trying to use the Microsoft.Build.BuildEngine.Engine object and fails. I noticed that MsBuild.exe seems to work fine, so I posted a patch on CodePlex that replaces the routine that attempts to build the project with code that shells out to console (it's already a console app, so it uses the same window). With that it works fine in 2008 and should work fine for VS2005, too.

You can get the patch for STSDEV here.

SharePoint Blog Tag Clouds

Doug Seelinger — Wed, 20 Feb 2008 23:27:28 GMT

As I've said before, my organization does an awful lot of SharePoint work. Since button-pushing really isn't my thing, I've got only a few options when it comes to coding most SharePoint projects:

I've started to duplicate this blog on our internal MOSS server, and there's something that I've missed about dasBlog that's not available in WSS or MOSS blogs - Tag Clouds. That sounds like a good place to get my feet wet in Web Part development. I know that there are already several web parts available that provide tag clouds for MOSS, but of course, they're "not invented here". No, seriously, this is purely for my own educational purposes, so even if I do provide a solution to the general public, there's certainly no assumption that it would be any better than what's already available. This project will let me explore several interesting technologies:

Web Part development (of course)
Code Access Security (CAS) in SharePoint
STSDEV
Microsoft Unity (though I'd rather play with StructureMap <sigh/>.)
VSTS 2008 TFS Team Build
any other random bits of new tech that catch my eye!

Technorati Tags: SharePoint,Web Parts,Unity

Microsoft "Unity"... Isn't

Doug Seelinger — Wed, 20 Feb 2008 19:41:04 GMT

I'm not dissing MS Unity, Microsoft's new IoC Container - I'm really going to give it a chance. Rather, I'm talking about the reaction to it on the Yahoo! [altdotnet] group.

I have just spent most of the morning following a thread on the [altdotnet] mailing list titled "Unity" Ioc from Microsoft has been released in CTP... And it has. But in retrospect, based on the thread, "Unity" was a very ironic name choice for the tool. This is a very long thread that is involving dozens of folks. It started off by comparing the functionality in Unity to functionality in other IoC containers, mostly Castle Windsor & StructureMap. The Spring.NET crowd has been strangely silent.

Later in the thread, however, the "altdotnet"ness of the list has begun to show its colors. Passions have arisen and tempered. Staunch defenders of each "side" (and there are more than two sides to this polygon) have held their ground while alternately being gracious one minute and on-the-attack the next. It's the ultimate Reality TV... except that it's not TV, of course.

I've wasted too much time trying to follow this thread and I declare that I am DONE with it. Not that the task wasn't worth it. I've learned a lot about IoC that I hadn't know before. I'm familiar with Windsor, and I've downloaded Unity and played around with it as well. The really great gem that I have just discovered is StructureMap. I just love the "fluent interface" it has - much like is available in Rhino.Mocks. Fun, geeky stuff.

I thought it would be a useful exercise to compare "Hello, World!" scenarios between the big-four (and, yes, I'm including Unity as part of that):

*Function*	*Castle Windsor*	*StructureMap*	*Unity*	*Spring.NET*
Putting an interface and concrete type into the IoC container	WindsorContainer. AddComponent()	StructureMapConfiguration. BuildInstancesOf(). TheDefaultIsConcreteType()	UnityContainer. Register()	N/A - Spring doesn't support XMLless config (that I could find). Those Java guys just have a love affair with XML, I guess
Getting a concrete type for a given interface	Resolve()	ObjectFactory. GetInstance()	UnityContainer. Get()
Default items are	Singleton	separate instance	separate instances

The fluent-interface of StructureMap is so nice. Yes, it's longer to type, but hey, that's what IntelliSense is fore, and that will make it easier for new developers to discover.

Incidentally, Unity looks an awful lot like Windsor in its syntax, and it was mentioned in the aforementioned thread that the PM on the Unity project is a big Windsor fan.

Technorati Tags: IoC,Microsoft,Open Source,Unity,Windsor,StructureMap

Apology: I just couldn't resist the lolcats thing, it was just too apropos. This is my one (1), lifetime, self-imposed limit.

Team Foundation Server 2008 and Continuous Integration - Screencast

Doug Seelinger — Thu, 14 Feb 2008 01:52:47 GMT

Wow, that was a lot of work. I always hate when a screencast of demo starts with this big huge installation of a bunch of different technologies, but doesn't tell you how to get the same environment, so I started from scratch and show you how to go from a generic W2K3 base image to a fully functional TFS 2008 box and THEN show the Continuous Integration sample. Due to the fact that we're going soup to nuts it's a pretty big screencast for me 34+ minutes - a new record for me. I only wish complete installations like this REALLY only took 30-40 minutes.

Screencast: Continuous Integration and TFS 2008

Technorati Tags: Team Foundation Server,Continuous Integration

How To Delete A Rogue Build Agent from VSTS 2008 TFS Team Build

Doug Seelinger — Sat, 09 Feb 2008 15:49:51 GMT

I'm pretty new to Team Foundation Server, so after my first (successful) install of it I tried to do a simple Build Definition (just an automated build configuration), but there are several moving parts involved, and somehow my VS Team Foundation Build service was set up to run as Local Service. I found that out when my simple build def refused to build (Local Service wasn't a member of the requisite security group). So I modified the service to run under the TFS service account and suddenly my "Build Agent" was "unreachable".

So since my build agent was then unreachable, I then tried to create a new build definition with a new build agent. The problem was that my original build agent was still holding onto port 9191, my default build agent port. When I tried to change it to port 9192, it complained, saying that I'd have to modify the port elsewhere so that the various moving parts could communicate. That was more than I wanted to mess with in my first foray into Team Build, so I tried to figure out a way to delete the old build agent. It took quite awhile for me to find, but I finally realized that TFS adds new menu items all over the place, not just in the Team menu. I finally found "Manage Build Agents..." under the "Build" menu. However, it only appears AFTER you click on Team Explorer. Once Team Explorer has been selected, it adds the appropriate menu items:

Build Menu before selecting Team Explorer

Build Menu after selecting Team Explorer

Now, even if you go back to Solution Explorer, the additional "Build" items will still be available. Every system has its quirks.

Team Foundation Server 2008 and Continuous Integration

Doug Seelinger — Fri, 08 Feb 2008 14:01:29 GMT

Why Should You Care? Continuous Integration saves you much heartache when the time comes to integrate your code with someone else's, whether that someone else is a colleague or an external system - like SQL Server, for example. Visual Studio Team System 2008 Team Foundation Server (man, that's a long name - I think it's URL is even shorter) has CI baked in, something that TFS 2005 lacked (and was basically burned by in the OSS community).

Why do I care? The decision has been made here at my organization to standardize on Team Foundation Server. Personally, I'm a CruiseControl.NET fan, but in all honesty, perhaps it's just that I'm comfortable with it. I know CC.NET has a pretty hefty learning-curve that goes along with it, but I've taken that curve and I'm comfortable with CC.NET, NAnt, NCover, etc.

So let's put together a TFS build-server and test it out. First, here are some links/videos regarding TFS 2008 & CI:

And that's about it. I smell a screen-cast coming.

Caveat 1: I still wonder what will happen when we engage with clients (we're the Microsoft consulting arm of EMC) who do NOT have licenses for all of the technologies that we're using. In my experience, we don't hang around extensively after we've built the software and maintain the application, though it has happened before (has anyone ever been a "security blanket" before?). But by and large, after we've delivered a project we go away. Inevitably, the client asks, "How do I maintain this thing?"

The convenient (and cop out) way to do things is to hand them a VM and say, "Have a nice day." But that's not right from a software licensing aspect. Perhaps we simply leave them the code and say, "Have a nice day," but that sounds pretty rude as well. That's why it's nice to have a very basic development environment using a lot of open source software, no licensing problems. Here's an example.

Caveat 2: Another down-side of going the TFS route for Continuous Integration is the fact that you have to have Visual Studio 2008 (Team System for Testers, Developers, or Suite) on the build machine in order to run tests, thus the build-machine is not as "pure" as it could potentially be (plus that's another license). However, most build-servers I've worked with are not as pure as the driven snow, either (though much less costly).

Caveat 3: TFS has a HUGE footprint. SQL Server (and attendant services) + WSS 3.0 + TFS Core + Team Build + VS 2008 Team Dev would not install on a VM with a 32 GB drive. Well, that's a bit of a fib. When I tried to install Team Build, it gave me a warning that performance wouldn't be as up to snuff since I didn't have enough disk. The footprint for a standard CC.NET install is miniscule in comparison. Thanks, MS, for kitchen-sinking me yet again.

I was even bitten by this one TWICE. I re-installed from my standard base VM image, added a second drive on which to install VS 2008, and changed the order of install - installing VS 2008 last. When it came time to install VS 2008, it wouldn't let me install it onto my second 32GB drive, since Team Build (or Team Explorer, not sure which) had already installed several VS2008 components on the C drive. However, VS 2008 didn't complain about lack of space as did Team Build. After finishing the VS 2008 install, I still had 21GB left on my C drive.

Team Build must expect to use A LOT of disk if it complains about disk in that environment.

Screen-cast To Come: I'll see if I can get a screen-cast out soon to show off the CI capabilities of TFS - I'm still waiting for VS 2008 to finish its install... once again.

Scrum and XP from the Trenches

Doug Seelinger — Wed, 06 Feb 2008 22:10:02 GMT

A colleague, Daren May, has turned me onto a great Scrum resource, Scrum and XP from the Trenches, which is the best online Scrum resource that I've seen yet. It's a free PDF book about one ScrumMaster's experience with Scrum. Most of what I've read about Scrum so far has either been heavy on theory or very "Pie in the Sky"ish. I was so impressed by it that I've emailed everyone I know who's interested in Scrum.

The author, Henrik Kniberg, has a blog that's just dynamite as well. Lots of good content on agile and Scrum in there.

As I said, I was very impressed, so much so that I decided to pay for a hard-bound copy, which is pretty out of character for me if it's available in a free form.

Secured MOSS/WSS Blogs and RSS Readers

Doug Seelinger — Wed, 06 Feb 2008 21:31:33 GMT

It doesn't look like Google Reader or FeedDemon are going to work out for subscribing to any of my secured corporate blogs (which currently only consist of my own blog). Let's give the Windows Common Feed List a try. Nope. IE doesn't support "feeds with passwords" as it puts it. Let's try Outlook. It looks like Outlook is the only reader (that I know of so far) that handles this kind of feed. Wow. I didn't know Outlook had it in it. Am I ready to leave Google Reader yet? Mmm... not really, but I'll just have to have separate areas for public and private feeds from now on, however.

Windows Live Writer

Doug Seelinger — Wed, 06 Feb 2008 19:10:21 GMT

Windows Live Writer is definitely the way to go for writing your blog. I've tried both BlogJet and WLW, and besides the fact that WLW is FREE!, I didn't really see any advantage to BlogJet. Plus WLW works seamlessly with my MOSS blog even though it's HTTPS, etc.

Why not MS Word?

Can't stand how it pops up 3 different windows in order to just do a blog entry.
Doesn't integrate with IE to blog with
Won't post to dasBlog (my public blog engine)

The SharePoint Driven Blog

Doug Seelinger — Fri, 01 Feb 2008 02:03:39 GMT

If you've followed my blog for very long, you've seen that it's mostly about agile software development. I took a short "sabbatical" over the holidays and started playing around with SharePoint (both WSS 3.0 and MOSS), and I posted maybe 5-10 blog entries on these subjects. Since then, I've seen a dramatic increase in my blog traffic - particularly from search-engine traffic. It's now 9-1 SharePoint referrals. It's not like there's not a lot of places out there to get SharePoint info. I know that I haven't had a hard time finding SharePoint info in the Googlesphere:

but there you have it. People are so starved for SharePoint info they're coming for my meager offerings (I must be 18,100,001).

It's all about marketshare. I remember seeing something similar in 1994-95, when I was mostly a Borland C++ guy. I loved Borland, but Microsoft Visual C++ seemed to reach a tipping point in marketshare around then and I decided to hop into a different boat - turned out to be a really good move. Whatever happened to Borland anyway? Funny, I don't even feel the need to look them up now, it seems so irrelevant. But I did, and it looks like the used-car dealearship of the internet:

But unlike Borland C++ vs Visual C++, I don't feel quite the same tension. I don't need to abandon "agile" in order to do SharePoint. At least I hope not. Agile hasn't made nearly the same in-roads in the MOSS world as it has in other areas of software development, so I kind of consider it a challenge to "agilify" (I just made the word up, I believe) the SharePoint world. There's only one question left...

Should it be MosScrum or ScruMoss? For some reason, I'm leaning toward's ScruMoss.

Technorati Tags: agile,moss

Only Mock Thyself

Doug Seelinger — Wed, 30 Jan 2008 17:19:15 GMT

I knew that conventional wisdom states that you should only mock your own classes, but I was tempted by all the beautiful interfaces within Microsoft.Office.Interop.Outlook. I'm starting to pay the piper, however. Rhino is slowing down a good bit because it has to emit so much "Reflection"ed stuff each time I CreateMock<AppointmentItem>(), etc.

Oh well, live and learn... or re-learn.

The Dearth of Outlook Add-In Info

Doug Seelinger — Fri, 25 Jan 2008 18:53:55 GMT

My current project has me working on an Outlook 2003 Add-In (funny how clients aren't always up with the latest-and-greatest, huh?). With it, I should be able to do just about anything in Outlook... if I only knew how. The problem is that this project has a really short duration (2 wks.) and I've got to figure out how to do some things that I've never done before. Usually I can find out all I need to know after a few sessions with Google, but info on Outlook Add-Ins (beyond the simple "Hello, World" style, which I can already do - there's a wizard, for heaven's sake) is pretty rare. Most searches bring back references to Sue Mosher's Outlook books, but after looking at the contents, there's nothing in there about Add-Ins so far as I can tell. The one book specifically about VSTO Add-Ins doesn't have any Outlook examples that I could find. Sheesh. I'm still installing VS 2008 on a box that has Office 2003 installed, which was another pain. I tossed my copy of Office 2003 awhile back and had to go the the company software download area to get a copy. So once 2008 is installed, with both support for Outlook 2003 & 2007, I'm hoping that they have some stellar documentation.

Anyone have any great tips for Outlook Add-In development?

Custom MOSS Workflow with Visual Studio 2005, with Guest Guru Arya Parsee

Doug Seelinger — Thu, 24 Jan 2008 23:28:32 GMT

Last but not least is workflow in MOSS is created using Visual Studio 2005. The same dude that did the SharePoint Designer screencast, Ted Pattison (http://tedpattison.net/), has another screencast posted on Channel 9 that shows the process of creating a custom workflow in Visual Studio for MOSS (or WSS 3.0 to be precise). This is exactly what I needed to finish off my crash course in MOSS workflow! Thanks, Ted.

Note that the version that Ted uses of the wizard is slightly different than what most of us would have installed today. He was using a beta 1 installation of the VS tools for MOSS workflow, but I didn't have any problem following along. He takes a couple of minutes to work through a problem that was in the beta, but appears to work fine in the released version without any workaround required.

One thing that I did notice, however, is that there appear to be a lot fewer "activities" available in the designer than he shows in the beta 1 - maybe a bunch got dropped after that beta. Another problem is that the deployment seems to be completely different in the beta he's using and what's in RTM.

I never did get it deployed correctly, Arya Parsee even got together with me in Office Live Meeting and we still couldn't get the thing deployed.

Arya says that he thinks I messed up the MOSS installation somehow, but we may never know, since I'm now billable on a new project that is going to be taking all my time over the next couple weeks or so.

"Where everybody knows your name..."

Doug Seelinger — Mon, 21 Jan 2008 19:03:59 GMT

Stop me if you've heard this one... Windows Server 2003 walks into a bar and...

Cute.

Creating a Workflow with SharePoint Designer 2007

Doug Seelinger — Thu, 17 Jan 2008 17:00:16 GMT

The next step for me in understanding workflows in MOSS is SharePoint Designer. There is a problem with SharePoint Designer workflows, however. They can only be attached to a single Content Type or Document Library. It's not like the Three-State workflow, that is available everywhere and can be attached to multiple things - SharePoint Designer workflows are all one-off workflows.

Note: I suspect that you can tweak the xxx.xoml.wfconfig.xml config file to associate this workflow with another list or library as well, but I can't be sure without trying and, frankly I just don't want to mess with it.

Given the above caveat, SharePoint Designer 2007 workflows are pretty powerful for next to zero coding (but where's the fun in that, I say!). If, and only if, you have a workflow that is manually started, you can have a custom page pop up at initiation that can be used to gather additional information about the workflow specifically.

But why would you have separate information just for the workflow? Why not have it embedded into the list or library content type? Good question, and there's a good answer. Document libraries and lists can have multiple workflows associated with them, so you wouldn't want to carry around the extra data associated just with the workflow that may or may not be used for a particular item.

For example, consider a document library that centered around a project type document. There could be many workflows associated with the project - like initiation, funding, execution, post-mortem. Some may or may not happen, as in the case where a project was not funded, you'd skip the execution step.

I did find at least a couple of interesting posts on Designer-specific workflows, I'm sure there are many more.

http://blah.winsmarts.com/2006-8-SharePoint_2007__Authoring_custom_workflows_using_SharePoint_Designer.aspx

And here's a good screen-cast: http://channel9.msdn.com/ShowPost.aspx?PostID=197661, that's one I don't have to do.

WSS 3.0 Built-In Workflow

Doug Seelinger — Wed, 16 Jan 2008 00:01:26 GMT

Here's the promised screen-cast. It was a bigger deal than I thought it would be and I had a couple of false starts before finally getting something that worked - even then one of the workflow emails did not get sent for some reason or other. It's also about the longest screencast I've done to date at over 24 minutes. Covered in the screencast:

Creating a domain user enabled for POP3 (to allow them to send/receive email)
Creating an Issue Tracking List in WSS 3.0
Enabling workflow for the Issue List
Setting up a WSS 3.0 Server's email capability
Creating an List Item and starting the workflow
Monitoring workflow
Tracing an items progress through workflow.
Completing workflow items.

Enjoy the screencast!

MOSS Workflow Notes

Doug Seelinger — Tue, 15 Jan 2008 04:42:22 GMT

Workflow (in MOSS): The automated movement of documents or items through a specific sequence of actions or tasks.

WSS 3.0 has one default type of workflow available - the Three-state workflow. It's main purpose is to supply the Issue Tracking list template with its workflow, but supposedly any list that has a choice-type column with three or more choices (??? why >3 when it's a three-state workflow ???) can use the workflow. Though this workflow is here by default to service the Issue Tracking list template, a new list from this template does not have the workflow associated with it by default.

MOSS's built-in workflow types are:

Approval - can have one or more approvers
Collect Feedback - completed when others have sent feedback
Collect Signatures - gathers signatures (somehow), can only be started in something like Word.
Disposition Approval - participants can decide whether to retain or delete expired documents

If you want something beyond the simple workflow that's built-in, you have two other choices, as I mentioned in a previous post:

SharePoint Designer - ONLY if you're dealing with a specific list or library - can't be copied or moved to a different list or site!
Visual Studio Extensions for WWF, created by a professional developer. Hey, that's me! Ah, job security.

Adding workflow - In order of increasing specificity, here are the things in SharePoint to which you can add a workflow:

Site Content Type - any list (or document library) in a site that uses your content type will have the workflow available for it. This is the most general way to specify workflow, though it does have to be attached to a particular content type.
List or Library - any item in your list (could be multiple content types) will have the workflow available for it.
List Content Type - only items in a specific list that are of a specific content type will have workflow available for it.

WSS/MOSS workflow is also tightly coupled with a specific Task List (you choose which one) besides the list that is "generating" the workflow. Tasks (BigBossMan must approve/reject document X) are stored in this list.

Workflow History is stored in yet another list in MOSS/WSS.

Hey, I should now have enough info in order to create a new Issue Tracking list, enable workflow and to test it out. Screen-cast tomorrow! Should be very exciting.

Stuff I Learned about MOSS Content Types Today

Doug Seelinger — Tue, 15 Jan 2008 00:32:22 GMT

More Workflow today. In learning about workflow I had to learn about Content Types. I did find a good page on describing what Content Types are: Content Types. Yes, it's the MS Office Online web site. It had very good documentation. Here's what I learned:

Content Types can be defined for:

Documents
List Items
Folders (haven't figured out how to do this one, however)

A Content Type can specify custom solutions of features that are associated with it (not sure about this one, either)
Content Types have the concept of Inheritance (they call it a hierarchy, however).

The difference between a Siteand a Site Collection? A Site Collection is a Site that has Sub-Sites... I think, Microsoft's definition was kind of fuzzy.

One difference between MOSS and WSS? WSS has only one Feature: Team Collaboration Lists.

Three Ways With Workflow:

Doug Seelinger — Wed, 09 Jan 2008 17:27:21 GMT

The big world of actual "work" is rearing its ugly head at me and I need to put at least SOME focus on whatever is likely to be my next project. It's a safe bet at the EMC Microsoft Practice that that project will involve MOSS 2007. I'm pretty interested in the ILM "2" beta, which is heavily involved with MOSS, but chances are that it won't be in the EMC pipeline for awhile. I'm at least mildly interested in Windows Workflow Foundation, so that's probably the best place to put my focus on for now.

From my somewhat limited knowledge in MOSS - I'm not completely clueless, I just don't spend 40-hrs a week in MOSS unlike some of my colleagues - I see that there are basically three ways to "work" with workflow in MOSS:

Built-in Workflow to the MOSS
Workflow built with SharePoint Designer
Custom Workflow built in Visual Studio (or Notepad, if you really wanted to). My guide is Essential Windows Workflow Foundation.

For some reason I find it a big pain in the patootie to install MOSS. Part of the reason is that there seems to be a big long checklist of "stuff" (maybe it's just in my mind) that you have to go through in order to get the development environment right. So for probably the next few days I'll be putting together my own MOSS install. Pray for me.

Fortunately, while trying to figure out how to get POP3 to work with my demo VM, I came across what appears to be a pretty good description of how to create a MOSS demo setup.

Ajax & the Browser "Back" Button

Doug Seelinger — Wed, 09 Jan 2008 15:52:19 GMT

I'm watching another ASP.NET 3.5 Extensions video while waiting for information in order to begin work on a new proposal. This video shows you how to allow the user to go "Back" in an AJAX web-app even though the page doesn't really postback during AJAX updates. Normally, a regular ASP.NET app fires postbacks when clicking on buttons & what-not, but put those buttons into an UpdatePanel and you don't get a full postback of the page, only a partial update, which does not affect the browser history (and thus you can't go back). I still have nightmares about the back-button in web-apps, so if it goes away, I don't care too much, but at least it's another tool in my box.