Hey guys,

Just a quick post to wish you all a happy new year. May your resolutions be fulfilled! :-p I hope I’ll achieve mine.

Sorry for the long delay for the previous article. Correlation is a very tricky technology. It’s still very vague and such. I hope I explained it clearly. Other hickups for the delay were, well, the holidays and some strange wordpress problems… which still need to be fixed, actually. :p

Anyway, Happy 2008!

Correlation is one of the key concepts in Workflow Foundation to understand when developping workflows. Most of the time people are confused about what correlation exactly is. They can’t really describe it, they can’t really explain it. Unfortunately, when you do not fully grasp this concept, it can lead to some major confusion when developping advanced scenarios and custom activities.

So what is correlation? Correlation is the machanism created to map an inbound message for a specific instance (sent from an external source) to the specific handler within the activity. There are two levels here:

• specific instance
• specific handler within the activity

Each level has his own differentiater. The specific instance (known by the external souce in the outside world) is defined by the CorrelationParameter and the specific handler within the activity (note that a workflow is an activity too) by the CorrelationToken.

But how is this implemented? Well, let’s first recap the basic objective: we are trying to establish external communication. The Workflow Foundation achieves this by using interfaces as a “channel”. The implementation of such a “channel” interface is a service which is added to the workflow runtime. This is actually the heart of correlation. As there are many instances (in the outside world) the service can work on, Workflow Foundation needs to know with which instance it must work on. This is why we need to define the CorrelationParameter in our “channel” interface. This is how an interface for external communication can look like.

[ExternalDataExchange]
[CorrelationParameter("taskId")]
public interface ITaskService
{
  [CorrelationInitializer]
  void CreateTask(string taskId, string assignee, string text);
}


In this example we define a CorrelationParameter “taskId” that will help us differentiate tasks - the instances - in the outside world. This name will stand for the correlation paramater throughout the whole interface. So when you use this name in a member (eg. CreateTask) as a parameter, Workflow Foundation considers the parameter as a correlation parameter for that member. Note that a correlation parameter can be a guid, int or whatever.
You also see a CorrelationInitializer attribute in there, which decorates the CreateTask member. When calling a member with this attribute, Workflow Foundation will know that the correlation parameter is initialized with this member. It will then reset internal pointers.

Now that the channel is defined, you can use it within your activity. You can use the channel directly by querying the runtime for the correct service. Workflow Foundation however gives you another possibility to ease working with different instances and handlers for those instances WITHIN your activity. This is where Correlation Tokens come into play. A correlation token encapsulates the correlation parameter and a value, together as a correlation property. For our task example, we could create a correlation token named “taskToken” which encapsulates the correlation property of property parameter “taskId” with the value 25. In other words this correlation token will hold the reference to the task 25.
Workflow Foundation also provides two activities that use this correlation token: CallExternalMethodActivity and HandleExternalEventActivity. A common way to use these activities consists of overriding their methods when developping custom activities (most sharepoint activities are implemented this way). Another method is to use them directly by setting the correct service, member name and parameters. As you probably already figured out, the first activity is able to call external methods. The HandleExternalEventActivity, on the other hand, will wait and listen for a specified event (after which it can trigger methods). Both of those activities use the correlation token to determine which instance they have to work with. It, however, makes most sense with the HandleExternalEventActivity, as an event cannot handle parameters, while the method called from a CallExternalMethodActivity specifies which instance to work on by one of the parameters.
To resume, here is a little schema of how to look at correlation tokens:

I hope this article helped you guys understand correlation a bit more. Untill next time..

This week I am starting a new series of articles to give you guys more insight in the Windows Workflow Foundation framework. The Windows Workflow Foundation (WF) is a framework which was released last year as part of the .NET Framework 3.0 alongside the Windows Presentation Foundation (WPF) and the Windows Communication Foundation (WCF). It allows you to model workflows: flows of logical (and easy-to-understand) work units. Here’s a basic example.

The units are represented by blocks, so-called “activities”, you can use by dragging them on your workflow designer. And right there, you get the main advantage of workflows. This methodology of creating a process by assembling logic building blocks together is pretty powerful. Especially when you envision a developer sitting down next to a business guy and interactively building the workflow together with him in just a couple of hours. And this without loosing track of the objective as they are working with small logical blocks.
For example, you can use a Send Email activity without having to know how it is actually sent or which methods you have to use. You just fill in a couple of properties and off you go. The big hick-up, unfortunately, is that there aren’t many blocks to choose from. So, we have to create them ourselves. And this is where the frustration begins. Most of the time the business has these complex rules set up and wants them translated to workflow technology. Creating all of these abstractions of the inner workings to the actual logic they’re producing, can be quite hard and sometimes just plain impossible. Why? Well, we have this framework we have to work around, which has very specific rules and complex, new logic to get code to run. And I’m not yet talking about parallel execution of our blocks.

Another great advantage is its pluggable system of services. When we want to run a workflow, the Workflow Foundation can add services to extend its possibilities. One of the most interesting services already developed, is persistency. It gives us the possibility to have the workflow, its whole instance, being interrupted and saved on a data storage. This way, lengthy processes that are waiting on some user interaction can be cleared from main memory and brought back to execution only when needed!

All in all, Windows Workflow Foundation is quite a nice technology, which can be very powerful. Unfortunately, development is proofing to be quite challenging. This is why I decided to give this series a try to get things cleared up and help you guys with some hints and tips. Hope you’ll like the series. Stay tuned!

Back again for a SharePoint branding issue, and this time we’ll tackle the calender. More specifically its calender views: the day and the week view. Let me first show you, how the week view should look like:

As you can see, an appointment can cover multiple hours. This way you have a graphical overview of how much time appointments will take. This, however, is how it looks like on the other MOSS master pages:

This is not what we expected. Appointments now only cover a small fragment of the time to take, while the rest stays blank. It is very confusing for the user. And some appointments seem to pointlessly be in different columns. The day view experiences the same kind of problem. What is this weird voodoo? We did not change anything to the calender, now did we?
Well, no, we didn’t. But we did change the looks of it by changing how the looks are parsed. As I already mentioned in another article a couple of weeks ago, the other MOSS master pages use a (correct) doctype declaration, while the default master page does not even contain one:

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd“>

Make no mistake, this is a declaration HTML pages should contain. Otherwise, you have no clue how browsers will render the page. Internet Explorer falls back to its “quirk mode”, where everything is “kinda” allowed. Other browsers will just try to render the page. And this is what we experience in the calender views. The inner box of the appointment has a CSS property: height equals 100% to make the appointment cover the whole time slot space reserved in the table of the view. In the doctype declared version, this is only supported if all parent elements also have a CSS height equal to 100%. Of course, this is not the case for all its parents. That’s why the appointments get shortened.

Is there a solution? Well, no solutions, only workarounds, I’m afraid.

• Remove the doctype declaration from the master page. I don’t really advice it, but it can do the trick. Make sure other browsers still display the master page correctly. Future browsers can break the layout however. (this is what SharePoint did for the default.master)
• Rewrite the views. You could make the views compatible with current standards, but it’ll cost quite some time to develop.
• Have SharePoint show the default.master for the calender views. It’s an ugly workaround, as you’ll have different layouts in your site, but it works.

Once more, a very unfortunate branding issue. Aspecially, because it could have been prevented by generating correct HTML. Hopefully this kind of bugs will be resolved in future SharePoint versions. Till next time.

The ASP.NET page life cycle, control or web part life cycle for that matter, is an important concept to know and to understand in ASP.NET web development. ASP.NET, the successor of ASP, introduced the concept of object oriented development. The complete page and its controls are now objects which behave in a child-parent relationship and go through a complete life cycle of their state. It’s pretty crucial to know what stages they go through and which events are called as the web is a stateless medium and we are trying to make a statefull environment out of it.

The key concept is the view state. This is an object, stored within the page, which enholds the states of all current objects. It is loaded, processed and saved during the life cycle. More specifically, it will load all data into the controls right before entering the load event. From this event on, you’ll have controls which are loaded with all previous data. You can manipulate those just until the PreRenderComplete event. Once this event has passed, the viewstate will once again be saved into a hidden control on the page. So things you change during the render event will not be taken into account.

Here is a graphical overview:

Couple of remarks with this diagram:

• The dark gray events are only available for the page object.
• The green area indicates events where the viewstate is loaded and avilable in the controls. Once changed the controls will be saved back to the viewstate.

As I said, the viewstate is a pretty important concept to grasp. So, please take time to fully understand it. It’s crucial to know when and where you can interact with it or you’ll be in a whole world of pain when creating interactive web parts or pages. Untill next time, take care!

Today I’ll be reviewing the SharePoint SDK’s. I’m not saying they are bad, but I’ve never found any good overview about ‘em. And I guess such an overview can be usefull, especially for people starting with SharePoint.

Before we begin, let me explain what a SDK is. A SDK is a Software Development Kit. It’s a set of resources designed to help understanding a specific technology. Sometimes a SDK is necessary to allow development with such a technology. (I’m thinking about the Java SDK) So, what are those resources? Well, it depends. They can include a lot of things. But most of the time, you’ll find documentation (CHM’s and the likes), samples, tools, templates, … you name it. Really usefull stuff.

SharePoint SDK’s come in different flavours. There is the online MSDN section with all the documentation, together with links to seperate downloadable resources. Or you can directly download the SDK. This download will include the documentation, samples, etc, but is suited for offline use, as an internet connection is not required once installed.

So what are the SharePoint SDK’s? First off, there is the Windows SharePoint Services 3.0 SDK. This SDK will only cover WSS related topics. And although it’s quite usefull, the offline installer includes only a couple of files, more specifically: the actual WSS 3.0 reference and a compilation of technical articles. Both can also be found on the online version.
The Microsoft Office SharePoint Server 2007 SDK is much more interesting. It includes both WSS files, and adds much more reference on the MOSS-only technologies. Also, it features some nice samples with source code. This gives us the opportunity to catch a glimpse at how the guys at Microsoft are using SharePoint, which is quite nice if you’re starting off with SharePoint and have no idea how to begin development. The MOSS 2007 SDK also includes the Enterprise Content Management Start Kit (ECM Starter Kit). This starter kit gives some in-depth information on the new ECM features and platform in Office SharePoint Server 2007. Moreover the starter kit is the collection of some white papers and code samples.

Next is a handy overview for the SDK’s:

In short, the MOSS SDK completely covers the WSS SDK. And even if you don’t have MOSS, the MOSS SDK is still very interesting and worth the download, as the samples give a great in-depth look at coding for SharePoint. ’till next time!

This final blog post (for now) in the SharePoint Branding Issues series is one regarding web parts and their branding issues. I’m assuming most of you know what web parts are. They are like mini applications within a web page. As they are so small, you’ll often add several of them together on the same web page. In other technologies, such applications are called widgets or gadgets or whatever. But basicly, their ideology is the same. A nice example is featured on live.com.

To get back to .NET, web parts are a new technology within ASP.NET (and unified in version 2). In previous versions of SharePoint, they used some custom made framework. And now in WSS 3.0, they are slowly migrating to the new ASP.NET 2 standard, which allows far more easy web part development. Finally to be complete, a few more statements regarding web parts. Web parts are placed within horizontal or vertical web parts zones. These zones can define the chrome style used for the web parts within. The web part chrome, for those who do not know, is a fancy word for its border decoration. Furthermore, the web parts need one single web part manager, per page, for.. well.. managing purposes. Finally, all of these concepts have a corresponding class within the System.Web.UI.WebControls.WebParts namespace. And basicly this is, in a nutshell, what web part infrastructure is all about.

Now, the thing with SharePoint. They decided to overwrite all of these classes (just as we would for a custom branding job in ASP.NET). But somewhy they made the decision to make the new classes all sealed. So basicly you can not change anything, except, of course, if you rewrite every single class… I know… And as one of my hungarian colleagues would say in such a situation: nice!

The chrome in SharePoint consists of 5 fixed chrome types (Default, None, Title and Border, Title Only and Border Only). These are implemented in an internal SPChromeHelper. What this class does, is adding the following structure as chrome:

<table>
    <tr>
        <td>
            <table>
        </td>
    </tr>
    <tr>
        <td>
            <div>
        </td>
    </tr>
</table>

Depending of the style the first row, which is the header, will be generated or not. Accordingly CSS classes are attached to the rows, the inner div and the inner table. As you can imagine, this kind of structure can be really restrictive. It is very difficult to add gimmicks, like rounded corners, to such a composition.

What can we do then? Well, basicly, applying tricks and hacking our way in. To achieve this, once again, our friend is CSS. By adding some classes to the right place and by using some specific CSS selectors, one can manage to achieve some degree of customization. A possible solution could be to add a specific CSS tag name to the declaration of the zone tag in your page. This way you can reach its inner web parts through the use of CSS selectors. In most cases you can define your own zones, so this can potentially be a solution for the problem. Another, more obvious solution, consists of simply overwriting the fixed chrome type’s CSS styles. The issue here is that every web part has to be set with the correct fixed chrome type to display your customizations. Of course, you still have the same limitations as the previous solution, as the structure of the chrome is still the same. More over, you usually cannot reach all generated HTML tags using only the fixed CSS class names.

While writing this article, I found this post written by Daniel Terborn. Daniel tackled the same problem and supplied another possible solution. He decided to create a base web part which implements all of the layout. Finally, with some CSS hooks, he manages to get really nice web parts. The problem with this solution is that it only applies to custom written web parts. In some situations this can be acceptable (by only using custom web parts), but most of the time you’ll want to use the out-of-the-box web parts too, for obvious reasons.

Conclusion? There is no clean and neat solution for web part branding within the current iteration of SharePoint products. We can only hope the sealed classes will be opened up in future versions. Until then, happy SharePointing.

Application pages are those special pages within SharePoint that share functionality accross sites. Some examples include the site settings, the list settings, each setting page within those settings, the All Site Content pages, the personal settings pages, etc. A way to directly recognize one of these pages is to look at the URL of the accessed page. If the url contains this segment “/_layouts/”, you are dealing with an application page. Why? Well, the /_layouts/ directory is a virtual directory created by IIS and is available to every created SharePoint site. You can access every page within the layouts directory by simply adding the path to your site url. For example: http://www.host.com/subsite/_layouts/settings.aspx. But where does this path redirect? The _layouts virtual directory is mapped to a fixed directory within the SharePoint hive, more specifically: C:\Program Files\Common Files\Microsoft shared\web server extensions\12\TEMPLATE\LAYOUTS\. So, every page within this directory is accessible through the special \_layouts\ directory.

Another way to recognize these pages is that, if you use a master page other than the default one, the application pages still seem to use the default master page. I write “seem”, because although the appearance seem to be the same, it is a different master page. This master page is also located within the layouts directory and is called application.master.

If you want a complete make-over of your SharePoint site, chances are you’ll want to change the appearance of those pages. One way to do this, is to edit the application.master master page. This can be a valid solution if your whole farm uses the same master page, as all application pages of the sites on your server use this system master page. So, if you change the master page, the changes will be visible on all instances. Most of the time, however, you have different sites on your server with different branding needs. So this approach is usually out of the question.

My solution consists of altering the application pages themselves. Although this is not a recommended practice, as Microsoft is not supporting this (Service packs could overwrite the changes), sometimes it’s the only option left. Let me show you what to add:

<script runat="server">
protected override void OnPreInit(EventArgs e)
{
base.OnPreInit(e);
MasterPageFile = SPContext.Current.Web.CustomMasterUrl;
}
</script>

Basicly we’re assigning, at run time, a new master page to the application page. Application pages can use the SPContext object to find out the context in which they are called. From this object we extract the current site (a SPWeb object) and assign the CustomMasterUrl property as the application page’s master page. This will do the trick just fine.

An important note to this solution, regards the master pages themselves. Master pages for application pages need to support two extra content place holders, PlaceHolderPageDescriptionRowAttr and PlaceHolderPageDescriptionRowAttr2. As these place holders don’t seem to do much, you can just add those two tags within a hidden ASP.NET panel to your master page.

<asp:ContentPlaceHolder id="PlaceHolderPageDescriptionRowAttr2" runat="server" />
<asp:ContentPlaceHolder id="PlaceHolderPageDescriptionRowAttr" runat="server" />

If you are using this technique, be sure to add these tags to the default.master master page in the SharePoint hive and this for compatibility reasons with the other sites. Of course, do not forget to add them to your custom master page too, as this is usually the reason, in the first place, to apply this solution.

A nice extension would consist of making the Application Pages Master Page of a site selectable through the UI (in MOSS) for selecting a master page. You could then use the SPWeb property bag to save the master page url.

Untill next time, happy sharepointing..

For those of you tracking my blog. You’ll probably have notice the small lack of updates. Well, I can firmly put the blame on this orange thingie. The Orange Box of Half-Life 2 got released last week (retail, that is; the steam edition is about 2 weeks old now). I was looking forward to the next installment of Dr. Freeman. And I must say this second episode is very much worth the wait. It’s a very exciting single-player adventure, and you really sense Valve has taken full advantage of all possible gameplay elements and graphic novelties within its Source engine. Also, the new “achievements” mode within the game give episode two more replay value, which is nice to have. And though episode one was quite short, this new episode ensures a much longer game time.

This is not the only game in the Orange Box. No, there are two other great games in the box. First game is the VERY long awaited Team Fortress 2, so Fortress players… rejoice.. And there is Portal, the game based on Narbacular Drop. It was already refreshing to play such a new style in its early days. But with the Source engine upgrade, it really got a boost.

If your an avid FPS player, go get this box. It really has a great value for money. And as I just beat episode two yesterday, I’m already looking forward to episode 3!

Navigation can break or make your site, and more specifically, your site’s usability. One of the tools to achieve a good navigation experience is the breadcrumb. In opposite to a menu structure a breadcrumb is quite space saving. And this is, according to MSDN, because it displays a hierarchical path of hyperlinked page names that provides an escape up the hierarchy of pages from the current location.

Basicily, it shows the current node and all its parent nodes. For the sake of completeness, it’s also called an eyebrow. And I’m not making this stuff up.. :p All in all, I’m pretty sure you guys all know the beast. I, for one, find it pretty usefull… as long as it keeps consistency along the way.

Back to SharePoint now. A breadcrumb within ASP.NET is called a SiteMapPath. This is a control which uses an XML sitemap of your site structure to keep track of your every move. It then shows your progress in a nice little breadcrumb. To access the sitemap, we need to use a .NET object called SiteMapProvider. As this is an object in code, we can change the way such a SiteMapProvider works and feed him some custom links programmically in stead of using a XML scheme. This gives a whole lot of flexibility. And because of SharePoint’s inherent dynamic nature, this is why almost all SiteMapProviders within SharePoint are implemented this way.

WSS natively offers a couple of SiteMapProviders (SPNavigationProvider, SPSiteMapProvider, SPContentMapProvider and SPXmlContentMapProvider), each with its specific use within your site. MOSS adds about 10 more. And it’s in MOSS, the troubles begin. Say, you design your site and you have to choose the breadcrumb’s location. Done? Ok.. which SiteMapProvider will you use? The default.masterpage has two (event up to three) breadcrumbs integrated in its design, all with different SiteMapProviders. (I wonder why this design decision was made) The two providers used are:

• SPContentMapProvider: the default WSS SiteMapProvider which shows nicely all different items and lists
• CurrentNavSiteMapProviderNoEncode: a provider which parses pages, not as items in a Pages list, but as sections within the site, which is pretty clean.

My requirement was: I want both advantages: items and lists nicely shown and pages cleanly parsed. Unfortunately, the WSS SiteMapProvider does not know pages and the MOSS SiteMapProvider does not parse lists and items. So, I decided to develop my own SiteMapProvider.

A SiteMapProvider consists of two things. A class containing the logic and an entry in the web.config enabling the SiteMapProvider’s use in a masterpage. Let’s first start with the logic.
As this is a MOSS problem only, I created a class inheriting from a MOSS SiteMapProvider: CombinedNavSiteMapProvider. This one will show nice complete breadcrumbs, except of course for lists and items. The SiteMapPath control uses one particular property from the SiteMapProvider: CurrentNode. So this is the one to override. This is the code I came up with:

public override SiteMapNode CurrentNode
{
  get
  {
    SiteMapNode node = base.CurrentNode;
    SPListItem listItem = SPContext.Current.ListItem;
    string url = HttpContext.Current.Request.Url.AbsolutePath;
    bool isItemSelf = listItem != null && url.EndsWith(listItem.Url);
    if (!isItemSelf)
      node = SiteMap.Providers["SPContentMapProvider"].CurrentNode;
    return node;
  }
}

Two things to notice. I check if the url is a page by comparing the requested url with the current item’s url. If they both match, it’s a Page alright. Second thing, to parse with the default WSS provider, one can find the correct provider in the static property Providers from the type SiteMap.
The last piece of the puzzel is the entry in the web.config. This is pretty straight forward. Go to configuration.system.web.siteMap.providers and duplicate the entry of the provider you inherited from. Change the assembly details and you have yourself a working SiteMapProvider.

All in all, this was quite a challenging change to find a solution for and to implement it. But once up, it works like a charm and seems so easy. See you next time!