Off-browser Silverlight is pretty cool. And one of the first questions that got raised is ‘where is it running’. Obviously, it’s not running within the browser. I mean, duh. But if not there, then where?
I got the first inkling of an answer listening to Mike Harsh’s presentation on off-browser Silverlight at MIX ‘09 (slides, demos and video can be found on his blog). Actually, it wasn’t his presentation, per se, but his answering of questions afterwards. I actually give Mike props for doing this presentation. Off-browser Silverlight is very easy to configure and use. It’s difficult to find 60 minutes worth of content when the user experience is ‘right-click to install…right-click again to uninstall’ and the developer experience is ‘check a box in the property sheet’. But from his easy handling of the questions afterwards, it’s pretty obvious that he knows the topic well.
One of the post presentation questions Mike was asked was related to the ‘how’ of off-browser Silverlight. Mike intimated that, basically, the Browser control was used. In other words, when installed on a client’s machine, launching the Silverlight application actually launches an application which hosts the Browser control. The Browser control then hosts the Silverlight application.
As it turns out, this is “basically” true. For the Windows platform, it’s precisely true. From Chris Hayak, the application which is launched is called SLLauncher.exe. This is also the application to which you will attach if you want to debug Silverlight apps in off-browser mode. For the Mac platform, the SLLauncher application runs webkit, which in turn hosts the Silverlight application.
The one downside of this approach is that, at least during the beta, there is no way to distinguish between multiple, simultaneous off-browser Silverlight apps. They all appear in the process list as ‘sllauncher.exe’. Hopefully this will change before SL3 goes gold.
One of the first questions that arose from the announcement of off-browser Silverlight was “What will happen to WPF?” The obvious source of this concern is that since Silverlight 3 can run either as part of a Web page or installed in an off-browser mode, why would there ever be a reason to write a WPF application? And since Silverlight seems to be the technology that has all of the new features, is there a possibility that WPF could languish as the ugly step child in the client application development world ?
First, let me assuage your concerns. WPF is not dead. This opinion is based on a couple of on-going development projects at Microsoft: Expression Blend/Design and Visual Studio. In the case of Expression Blend/Design, the entire application was written in WPF. For Visual Studio, the code editors are being re-written in WPF. If nothing else, the investment being made by Microsoft to these products in WPF should demonstrate its on-going commitment to the technology. And there is the on-going integration of WPF into the Live Messenger client to add more fuel to the argument.
Going forward, I see WPF and Silverlight moving ahead more or less in lock-step. Features in Silverlight that are successful will find their way into WPF (Visual State Manager, for example). Features in WPF that are useful will move into Silverlight (element binding, based-on styles). Since the products have different audiences, each technology will be driven forward with a different set of priorities. VB.NET and C# already have this So don’t give up on WPF because of all of the excitement from Silverlight. As a WPF developer, I found some of the excitement generated by Silverlight announcements a little odd (applause for element-to-element binding? WPF has had that for a while now). What playing with Silverlight 3 that I have done so far suggests that the disparity between the two features sets is going to be much less in the future.
That having been said, there is still the open question of when WPF should be used instead of Silverlight 3. I can see three main cases at a minimum, but they all revolve around the same restriction in Silverlight…the security sandbox.
Access to local system resources
Off-browser Silverlight runs in the same security context as on-browser Silverlight. So if your application needs access to the local file system or the communication ports on the client machine, then Silverlight is not going to work out. In fact, this limitation can be extended in include any feature that requires full-trust to operation. Silverlight doesn’t work in full-trust mode. WPF can.
Full 3-D graphics capabilities
With Silverlight 3, perspective 3-D is available. But that is not the same as the complete 3-D capabilities that WPF has. It’s nice, but if you’re looking for fully rotational 3-D images, then WPF is the choice.
I’m sure there are more differences. It’s not like WPF is not a compelling choice for a development platform. In fact, I approach the choice between the two the same way I would between WinForms and ASP.NET. Because ultimately, the decision between Silverlight and WPF will be based on the specific requirements of a project. If Silverlight (either on- or off-browser) is sufficient, then pick Silverlight. If not, then pick WPF. Regardless of your choice, you’re not in danger of dead-ending with the technology. Both areas will flourish and grow for many years to come.
The big announcement at yesterday’s keynote was the availability of an off-line and off-browser mode for Silverlight 3.0. Just to clarify the terms, off-line mode indicates that Silverlight will not require a live connection to the Internet in order to continue functioning. Practically, this means that a Silverlight application will able to be installed on the client’s computer. Off-browser means that Silverlight apps will no longer appear to need to be running within any browser.
As a point of clarification, this last feature doesn’t mean that Silverlight will have free reign over your computer. Off-browser Silverlight apps will still run in the same security sandbox that they have in the browser. It’s just that you won’t need to explicitly launch the browser to start the application. In fact, the installation process for the Silverlight application includes the ability to place links on the desktop on the system tray
So what are the ramifications of these changes?
First, it opens up a world for Silverlight applications. While I’m a huge fan of WPF, there are a number of roadblocks to wide acceptance. It needs to have the .NET Framework installed (3.5 or 3.0). Which, in itself, implies a Windows platform (according to the Mono project, they are not planning to support WPF). Full WPF applications need to be installed. Yes, you can use XBAP to deploy the application, but that model has implications on security (it runs in the browser-driven security sandbox).
Silverlight, on the other hand, runs on both Windows and Mac. And while it requires an installation step, to call it an install is really overkill. It’s basically an XCOPY from one place to another on the client’s system. The conclusion, though, is that WPF and off-browser Silverlight do hit different target markets.
The second question is how does it change your Silverlight development process.
The goal of off-browser Silverlight was to have the same application run in both modes. And it certainly appears that is the case. However, being able to run the same application off-line and on-line does have an impact on your approach to development. For example, UI pieces might be different. The ability to retrieve data for a combo box, for example, might depend on being able to call out to a WCF service. That isn’t happening in off-line mode (although it would if you were just off-browser). So to create a solid off-line scenario, you may need to provide additional functionality or enable/disable functionality that is currently in place. This is not an insurmountable problem, but it does require some care and design effort.
All in all, off-browser/off-line Silverlight is an interesting idea and I’m looking forward to not only working on apps in this space, but also to see how others take advantage of it. Exciting times indeed.
Have you ever been dissatisfied with the code that is automatically generated by Visual Studio in response to various commands that you perform. For example, in VS2005, I got used to using the prop snippet to create properties. This snippet created a property declaration complete with a private backing variable. However, in VS2008, the prop snippet was modified to create a property declaration in the form of the automatic property declaration. That, to me, was annoying.
(Note to Microsoft: Please consider the existing snippets to be part of what you consider when looking at backwards compatibility. Create more snippets, sure. But please don’t change what the current snippets do…especially when the ‘better’ snippets are not actually better)
Or, consider the fact that, as more and more work is done in WPF, the need to ensure that business classes implement INotifyPropertyChanged grows. Adding or modifying a snippet to include the code to call PropertyChanged could be quite useful. It certainly has been for me.
Little known by most developers is the location for the snippets that Visual Studio provides. Or, what might be more interesting to intrepid developers, the snippets used by Visual Studio refactoring. All of these are found in %ProgramFiles%\Microsoft Visual Studio 9.0\VC#\Snippets\1033. The keyword snippets can be found in the Visual C# subdirectory, while the refactoring snippets can be found in the Refactoring subdirectory. Take a look at what’s there and you may be surprised how you can increase your productivity for mundane (and frequent) tasks.
So to set up this problem, an application that I'm currently working on needs to process data that is stored in an Excel spreadsheet. The creation of the spreadsheet is actually performed by a scientific instrument, so my ability to control the format of the output is limited. The instrument samples liquids and determines a quantity. That quantity is placed into the spreadsheet. If the quantity in the liquid is not successfully read, then the literal "Too low" is placed into the sheet. The application opens up the spreadsheet and loads up a DataSet using the OleDb classes (OleDbConnection and OleDbDataAdapter).
This seemed like a fine setup. At least, until some of the numeric values in the spreadsheet were not being read. Or, more accurately, the values in the spreadsheet were not making it into the DataSet. Head-scratching, to say the least.
After some examination, the problem became apparent. When the values were being successfully read, the column in the DataSet had a data type of Double. When the values were not being successfully read, the column in the DataSet had a data type of String. Now the difference between success and no-success was nothing more than the contents of the spreadsheet.
Now the obvious path to follow is how the data type of the column is determine. Some research brought me to what I believe is the correct answer. The Excel driver looks at the contents of the first 8 cells. If the majority is numeric, then the column's data type is set to Double/Integer. If the majority is alphabetic, then the column becomes a string.
Of course, this knowledge didn't really help me. As I said at the outset, my ability to control the format of the spreadsheet was limited. So I needed to be able to read the numeric data even if the column was a string. And, at present, the cells containing numbers in a column marked as a string were returned as String.Empty.
The ultimate solution is to add an IMEX=1 attribute to the connection string. This attribute causes all of the data to be treated as a string, avoiding all of the cell scanning process. And, for reasons which I'm still not certain of, it also allowed the numeric data to be read in and processed. A long and tortuous route yes, but the problem was eventually solved.
I just finished teaching a Silverlight/WPF course last week. As is true in almost every course I teach, there were questions from people who are trying to use the technology in the real world. It's what I love about teaching...that I always learn something about how people use the tools they have been given.
In this particular case, the problem was relating to walking the control hierarchy in a user control. The user control contained a DataGrid and the goal was to get from a particular cell within the DataGrid back to the user control itself. The code that was written to do so looks approximately like the following.
public UserControl GetContainingUserControl(FrameworkElement element)
if (element is UserControl)
The problem with the code is that, in the case specified, the Parent becomes null before the UserControl is found. For most developers, this is a strange and unexpected turn of affairs. For most hierarchies, it seems completely reasonable to presume that every element in a hierarchy will have a non-null parent except the root of the tree. But if you read over the description of the Parent property closely, you will see that it could be null. From MSDN,
Parent may be a null reference (Nothing in Visual Basic) in cases where an object was instantiated, but is not attached to an object that eventually connects to the Silverlight RootVisual, or the application object.
Ouch. Sure it's documented, but still. Ouch.
Fortunately, there is a solution, found in an incredibly useful class named VisualTreeHelper. This class exposes a number of methods that can be used to navigate up and down the visual hierarchy in a WPF interface component. There is, for example, a GetChild method that retrieves a particular child element. As well, of particular interest for this example, there is a GetParent method that retrieves the parent of a given element. This method will not return a null parent until the top of the visual hierarchy is reached.
In other words,VisualTreeHelper.GetParent(element) can be used in place of element.Parent in the above method with the result being more in line with expectations.
I’m sure we’re going to look back at 2009 and say “it was the best of times, it was the worst of times’ and it will no doubt be interesting. Here’s my predictions….
Social Networking Everywhere
Although online social networking companies are already struggling with diminished valuations, in 2009 we’ll see social networks break out of their silos and become essential platform elements that see their way into other online applications such as travel, e-commerce, job-posting boards, online dating services, CRM services, web based email systems, etc. Blogging is also changing, slowing down in fact. Micro-blogging with status update-esque features in FaceBook, Windows Live, and of course the explosion of Twitter will take on even larger roles. It’s as true today as it was back in 1964 when fellow Canadian Marshall McLuhan wrote “The Medium Is The Message”.
The Death of Optical Media
Okay, so you’ll still be able to walk into a video store to rent a DVD or buy a spindle of blanks at your grocery store but make no mistake about it – the death march is on, and that includes you too Blu-Ray. Blu-Ray will never see the adoption curve that DVD’s had. They thought they won when HD-DVD died, but if winning means dying last, then sure, you won. We’ll increasingly be renting our movies on-demand through our cable boxes, on our converged PC’s and XBOX 360’s via services like Netflix. Along with this, the rest of us will start to realize we don’t really need to own our libraries of movies. With IPod penetration as high as it is, it may take longer to realize we don’t need to own our music either – frankly we don’t own it anyway even though the pricing models try to convince us we do. I won’t go out and predict the death of DRM, frankly, I think 2009 maybe the year where DRM starts to get more tolerable once we are clearly renting our music and movies. The Zune Pass is making some inroads here but until Apple starts offering a similar subscription pricing, this may take a bit longer.
The Mac Air may have been a bit ahead of the curve with dropping the optical drive, but get used to it. Expect more vendors to do the same as they reduce size or cram in additional batteries or hard drives.
The Rise of the NetBook
If 2009 is the year of doing more with less, then this will surely be the NetBook’s year. Mainstream hardware manufacturers hate these and their small profit margins, but Acer and Intel will be raking it in building market share if not large bottom lines. Who knows, MS may learn to love the NetBook if they can get Acer to start shipping Windows 7 on them this year as well. Be prepared to see these everywhere in 2009, but don’t expect to see Apple make one (ever).
The big story at the end of 2008 has been the global suicide of the original Zune 30s. I predict that tomorrow they’ll be they shall rise from the dead but it might take until the 2nd for everybody to figure out that they need to entirely drain the battery. The big news is that there won’t be a Zune phone with the MS brand name on it, but the Zune UI will come to Windows Mobile (6.5?) turning legions of touch based smart phones into music players almost as good as an IPhone. The bad news is that without an App Store to vet software quality, crapware will continue to be the source of reliability issues for the Windows Mobile platform. The good news is that without an App Store, Windows Mobile users will have lots of choice in the software for their devices, not to mention lots of choice in devices, carriers and plans. The battle between Good and Evil may morph into the battle between Reliability and Choice.
Get your head out of the gutter, that’s not what I meant. What I did mean is that 12-24 months from now, it will be difficult to purchase a digital frame, LCD monitor or phone without an onscreen touch capability. Windows 7 will light these devices up and we’ll start to not think about the differences between Tablet PC’s and Notebooks as they just converge into a single device. With the advent of Silverlight, WPF and Surface computing, MS has been banging the “user experience” drum for a while now but when touch starts to be the expectation and not the exception, we’ll have to re-engineer our applications to optimize for the touch experience. This may turn out to be bigger than the mouse or even a windowed operation system.
Flush with Flash
In 2008 we’ve been teased with sold state hard drives but with less than stellar performance at outrageous prices, they’ve been on the fringe. In 2009 prices and read/write times will both come down in solid state drives, but with the increased capacity of USB memory sticks 32gb, 64gb +, we likely won’t see SSD drives hitting mainstream this year. Instead I think we’ll see an increase in the behavior of people keeping their entire lives on USB flash memory sticks. Hopefully we’ll see sync & backup software such as Windows Live Sync, Active Sync, Windows Home Server, etc. become more aware of these portable memory devices that may get synced from any device in your mesh.
Camera flash will have to have a new format as SDHC currently is maxed at 32gb. With the increase in demand for HD video recording on still and video cameras, we’ll need a new format. As such we’re seeing rock bottom prices on 2gb chips now. Maybe somebody will come out with a SD Raid device that lets us plug in a bank of 2GB SD Cards.
Growing up in the Cloud
Cloud computing is going to be a very long term trend. I think we’ll only see baby steps in 2009 towards this goal. In the consumer space we’ll see more storage of digital media in the cloud, online backup services and the move of many applications to the cloud. Perfect for your Touch Zune Phone and Touch NetBook without an optical drive eh? IT shops will take a bit longer to embrace the cloud. Although many IT Data centers are largely virtualized already, applications are not all that virtual today and that doesn’t seem to be changing soon as developers have not whole-heartedly adopted SOA practices, addressed scalability and session management issues nor adopted concepts such as multi-tenancy. As we do more with less in 2009, we won’t see that changing much as a lot of software out there will be in “maintenance mode” during the recession.
Maybe, Just Maybe, this is the year of the Conveniently Connected Smart Client
Adobe Air & Silverlight are mainstreaming web deployed and updated rich client desktop apps. It’s hard to take advantage of touch interfaces and massive portable flash storage within a browser. All of these other trends can influence Smart Client applications, potentially to a tipping point. We’ll hopefully see out of browser, cross-platform Silverlight applications in 2009 to make this an easy reality on the MS Stack.
Incremental, Value-Based and Agile Software Development
Many of my customers began large-scale re-writes of their key software assets in 2008, many of them against my recommendations. For most of my key customers in 2008 and into 2009 I’m an advocate of providing incremental value in short iterative releases, not major re-writes that take 6+ months to develop. Even if your application is written in PowerBuilder 6 or Classic ASP, avoid the temptation to rewrite any code that won’t see production for 4 months or longer. We can work towards componentized software by refactoring legacy assets and providing key integration points so that we can release updated modules towards gradual migration. It is difficult for software teams in this economy to produce big-bang, “boil the ocean”, build cathedral type projects. We simply can’t predict what our project’s funding will be in 4 months from now, or if we’ll be owned by another company, scaled down, out sourced or just plain laid off. That is of course unless you work for the government. Government spending will continue if not increase in 2009, but still, try to spend our taxpayer money wisely by delivering short incremental software releases. It allows you to build trust with your customers, mark a line in the sand and move onward and upward, and let’s you move quickly in times of fluid business requirements and funding issues.
Incremental, Value-Based software development isn’t easy. It takes lots of work, creative thinking, and much interop and integration work than one would prefer. It might easily seem like an approach that costs more in the long term, and in some cases you could be right. But if a company has to throw out work in progress after 6-8 months or never sees the value of it because of other changing business conditions, then what have you saved? Probably not your job anyway.
In the excitement of PDC, it slipped my mind to let everyone know that the book on which I was a co-author was actually shipped at the beginning of October. The title is the terse, yet incredibly descriptive MCTS Self-Paced Training Kit (Exam 70-503): Microsoft® .NET Framework 3.5 Windows® Communication Foundation (PRO-Certification). There is a bidding war for the movie rights and I'm hoping that George Clooney plays me in the adaptation. :)
For those of you wondering how the actual release might have slipped my mind, the reason is that I'm not involved in the steps that takes place at the end of the publishing process. Most of the book was written in the first half of the year. Since July, I have been reviewing chapters and responding to editor notes. But since the middle of August my tasks have been done. And, I'm afraid, when it comes to book writing, once I'm done, I mentally move on to the next task. So I wasn't even sure when the publication date was. But it was released and, based on the numbers that I've seen so far, it seems to be doing quite well. If any of you have the chance to read it, I'd be thrilled to hear any feedback (both good and bad).
DevTeach Montreal is less than a month away but it’s not too late to register. This is a great conference with sessions covering .NET FX, Future, SQL Server, VSTS/Team System, Silverlight, Agile Development, Software Architecture, ASP.NET.
Aside from the great content, build up your professional network by rubbing shoulders with the speakers in an intimate conference. The list of speakers is particularly impressive this year. From MS you’ll get to see Elisa Flasko and Carl Perry from the Data Programmability Team, Beth Massi and Yair Alan Griver. Of course .NET Rockers Carl Franklin & Richard Campbell will also be there with fellow MS Regional Directors Tim Huckaby, Joel Semeniuk, Stephen Forte, Jim Duffy, Guy Barrette and yours truly.
But wait, there’s more. Each attendee will get Visual Studio 2008 Pro and Expression Web 2.0 full copies along with the entire DVD set covering all sessions from TechEd 2008 Developers conference from Orlando this year. It’s Shamwow!
If you attended TechDays, your included coupon is now worth $350 off the price of DevTeach. If you didn’t, you can use this code:TO000OBJSHARP good for 50$ off. Sign up here.
Earlier today, a colleague had an issue with the publishing of unit test results into a TFS instance. The publication process, which is typically done manually at the click of a button, was no longer available. Specifically, the Publish button was actually disabled. There was no obvious error message indicating what, if anything, was wrong. This lack of information made identifying the problem a challenge, to put it mildly.
The solution, at least to identifying what the problem was, is to use the command line version of MSTest. If you execute the command MSTest /? in a command window, you will see that there are a number of options which can be used to execute a set of unit tests and publish them to a TFS server. For example, the following command will execute the unit tests in the TestLibrary.dll assembly and publish the results to the TFS server located at http://TfsServer:8080
MSTest /nologo /testcontainer:"TestLibrary.dll" /runconfig:"localtestrun.testrunconfig"
/resultsfile:"TestLibraryResults.trx" /test:TestLibrary /publish:http://TfsServer:8080
/publishbuild:"DemoTestBuild_20081103.1" /teamproject:"DemoProject" /platform:"Any CPU" /flavor:Debug
In this particular situation, running MSTest generate an error that indicated that the drop location for the build could not be created. An error that was, thankfully, quite easy to correct. But difficult to identify without using the command line tool.