I want to develop a web Application to manage school administration. But I am not clear, which technology should I use to create Rich Internet Application. At present, develeopment company is going with ASP.net tools and say this is a web app.
However, there is page refresh on every click. Which I don't like. Please suggest, where to proceed.
Regards
Piyush
In a plain ol ASP.NET application, the screen refresh can be eliminated by using update panels in webforms, or by using more of the AJAX capabilities of the MVC Framework. (Wikiplex on Codeplex has a crazy simple example of this that I love in their example application).
Between Silverlight and Flash, however, Silverlight is designed to be a Web Application, or (at times) a desktop application in a Browser, while Flash is more of a multimedia framework.
Since it is school administration, which doesn't require random people to have Silverlight installed, Silverlight and JS/AJAX are both good options. Both are built into the ASP.NET set of libraries, and have good support in Visual Studio. Flash is a bit of an outsider in this regard.
So to sum up:
In general JavaScript/AJAX > Silverlight but either would be great.
If you are totally against any screen refresh, I'd lean more towards Silverlight.
If the developers are very familiar with Flash and communicating with the server with Flash, then I'd say it is also an option, but with no knowledge on all three and I'd not pick Flash first.
It's a little late for a response to the initiator. But I didn't agree with some of the answers I read here, and since google is popping this up in the search returns I felt compelled to align my answer on this board.
First off, a lot of developers create applications in ASP.NET and AJAX because they are comfortable with it not because the technology is better. For years, what it took one person to develop in the Adobe Flex framework, took many developers to create in AJAX. There is no more AJAX versus Flex, there's just Flex. Who wants to write a hundred lines of code in AJAX to do the same thing it takes 1 line of code in Flex. That's why the cost with an AJAX deployment was always more expensive.
On ASP.NET, you're right. Who wants to see the page refresh every time you perform an action? Not to mention, no browser follows the HTML implementations exactly, so how .NET app looks in 1 browser would look different in another browser, in addition to differences witnessed with the various browser versions.
Most ASP.NET, heck even Coldfusion Applications are clunky in HTML. So with that in mind all of my gui development is done in Adobe Flex. When a flash file is created, it compatible with 99% of the browsers (safari on mobile devices excluded), and all of the Operating Systems are supported. I'll tell you what, you can get silverlight to work in linux but have fun while it loads 50MB of MS software on your Linux OS. Not to mention Flash has a 99% market penetration. So it's already preloaded on machines and browsers.
Plus flash supports most picture formats and SVG natively. Did you know Silverlight became notorious because they don't support even a GIF/BMP formats? Forget SVG if you're using Silverlight because it doesn't support it. What's SVG? Look it up! It's only the most useful way to create any image you want right in a flex app. For instance maps are useful in SVG. AND its just a small file. Not a horrendous large file.
Language, Flex is just easier to program in and there's a lot of support for it. Think I'm kidding, try for yourself. You can download a Free version of flex for 30 days. Forgive my ignorance but I don't even know where to begin on loading Silverlight. I've got visual studio and all that but I gave up after trying. But that was 1 year after it's release so it may have changed.
And frankly, if you or a developer can't afford flex, DON'T BUY IT. Instead, download eclipse, which is free forever. Adobe Flex is only a special iteration customized by Adobe anyway. So you probably won't get charting and some other components in eclipse, but most of the GUI components are built right into Eclipse.
Programming ASP.NET is just old hat. It's clunky, difficult to work with and it's possible to corrupt the files. It's not possible at all to corrupt your project files in Flex/Eclipse because they are just TEXT files. How cool is that?
Don't get me wrong, I program in ASP.NET too. But I like it for Application server processes; you know like the middleman between database and client. SQL server feeds ASP, and ASP feeds my Flex GUI using SOAP. Big plus is even though I hate Visual studio crystal reports, it's still way better than nothing. Believe me you want to have some reporting possibilities like CF Report Builder or crystal reports. Some people just like a nicely formatted PDF and that's all there is to it.
I've released hundreds of applications using various technologies. Right now this is my favorite. But frankly, a good GUI can be done anywhere if you are a 'kung fu' master. I'm saying that although it's easier out of the gates to program in Flex compared to .NET/Silverlight, someone with some mad .NET skills can put a Flex app to shame. But as it stands right now, there is no argument that Flex is the stronger candidate with the most market penetration, flexibility, and ease of use.
Good luck all...
ASP.NET does provide some special sauce for AJAX, so it's entirely possible to eliminate the page load without leaving ASP.NET. If you leave ASP.NET, the options are endless: Java, Silverlight, JQuery, Flash, etc. ad infinitum.
Silverlight and Flash are great for richer user interfaces, however I dont think they should be used for the "whole" interface or application. Its best to narrow down these parts and decide how rich these need to be, then evaluate which technology to use. If you just want to prevent full round trips to the server on a page, AJAX is a good option. I would also recommend Silverlight over Flex, since you probably/already have the tools.
Related
I know this question was asked before, but almost 2 years have passed and the business requirements are a little different.
We are in the beginning of developing a mid-sized application and are divided which technology to use for the front end, WPF or ASP.Net MVC 3?
We are not an IT-company, but a business company with an IT-department who can outsource programming tasks, while the business core shall remain within the company. I did spend a lot time searching the internet for the answer, and I partially did succeed, but since the question is so important, I thought I ask here as well.
Of course, before someone can answer the question I need to specify the requirements and environment for the application at stake:
Infrastructure:
We have a pure Windows environment. Each user will have either Windows XP SP3 (currently) or a future version of Windows (if we skip from XP to Windows 8 remains to be seen, but let us assume that the user will use Windows 7 next) installed.
We are aiming for a service oriented architecture, meaning we only want to run/show on the client machine what is really needed. This is especially important since the databases are far away from the clients (USA/Europe). We plan on using WCF for cross machine communication between user system (brower or WPF), application and database server.
We expect the main user group to be around 30+, but since we are a growing company there should be no issue scaling up to 100 users. The users are spread over three main locations over the world, while we want the option to support smaller locations as well. All those locations are connected to the same intranet.
UI Experience
The new system is supposed to replace existing systems which are desktop applications (Winform). The number of screens are likely to be around 100+ with many labels, comboboxes, graphs. I like to call it an expert system b/c we expect the user to spend multiple hours a day with it, the user is expected to do interact fast with it (many clicks, multiple dialogs pop-up and close etc.) and the application will contain a lot of business logic (mostly mathematically).
Some limited interaction with Excel is required. At this stage only importing formatted data out of .xlsx file into the application in order to work with the data. This we expect to happen often.
Copy&Paste from Excel or other applications into our new application is a requirement (no pictures, just text).
We will use a vendor control library for a richer UI experience.
The users are used to desktop applications for their daily work (current systems/Excel etc.).
Tablet or smartphone support is not a requirement.
Deployment
If we were to use a WPF application we would likely either deploy it in CITRIX or use Click-Once.
Here are arguments of the two opposing factions:
Pro Web:
Deployment is much easier. All the requirements can be done in a web application directly, and if not we use ActiveX or make a separate desktop application for the missing parts. Also, the IT world is going to stop doing pure desktop applications and everything is moving to HTML 5 (Windows 8).
Pro WPF:
Web applications use many different technologies which makes it more difficult and costly to develop and maintain (HTML, ASP.NET, CSS, JavaScript, JQuery, AJAX). There are major deficiencies in a web-applications, mainly
Considering the various browsers and versions.
Screen resolutions
No hardware support for graphics (business graphs, point graphs with 200+ points)
Restricted access to local hardware (importing files, creating files, printing)
Keyboard shortcuts
Point #1 is also worrying since the browser is more out of control since other web-applications in the company (not expert systems) are used, and we fear conflicting interests with the new application (e.g., we must use a browser version where ALL applications run/render fine).
I know there is no black and white on this, but I would be interesting in the following:
Who was in a similar situation and how did they solve it?
(there is a nice article at http://karlshifflett.wordpress.com/2007/12/20/reasons-for-choosing-wpf-over-aspnet-for-very-large-project/, but the problem is that the article is 5 years old :(
How much more expensive is a web solution?
For development assume that the programmers are equally skilled in both (we can outsource this). For maintainance assume that we will internally support this where we have limited knowledge in ASP.NET and WPF. We know WinForm/WCF using C#. We would have to train/learn either technology.
How easy can a web application do Excel interaction, printing etc.?
I read a lot about "ActiveX hell" and I am wondering where we stand today?
Deployment
I have used Click-Once quite successfully in the past, although some team members mention that Click-once can be an issue. Any experiences?
Future?
The system is supposed to last 5+ years. We can not target HTML5 at this time (WinXP only up to IE 8). Where does Windows 8 stand on this?
Other thoughts?
What important things am I missing?
Thank you!
I know this entry is long and not an easy question. So I think you for reading and thank you even more for constructive feedback. Thank you!!!!
I would not build a business application in WPF, especially if your goal is to have it last for 5+ years. Silverlight is in sunset phase now - Win8 apps are betting on JavaScript and HTML5 now, though you are correctly noting that HTML5 support is not universal across all browsers and platforms (see http://caniuse.com/)
Let me try and address some of your concerns above to hopefully persuade you to build a web app:
Considering the various browsers and versions. Yes, you would have to do that. However, for enterprise applications most of the time you can find an acceptable solution if you use industry-standard web technologies and don't use esoteric HTML5 stuff that is not universally supported. It won't be a slam dunk, but it is very much doable.
Screen resolutions. You can address this by utilizing what is commonly known as Responsive Web Design. Once again, there is broad community and industry support for CSS frameworks that allow you to achieve responsiveness. YUI and Bootstrap come to mind as two examples.
No hardware support for graphics (business graphs, point graphs with 200+ points). Well, here is where HTML5 Hardware Acceleration may help you, but I'd say that libraries like HighCharts can easily handle 200+ points graphs - see this example
Restricted access to local hardware (importing files, creating files, printing). Fair point. I'd argue that working with files is MUCH easier with things like Socket.io, Filepicker.io and Zip.js, but "enterprise" requirements may get in the way. And as for printing, you can create "printable" version of your pages or generate PDFs and Excel Exports on the server side. Not ideal, but very much workable.
Keyboard shortcuts. Have you used Gmail app? It is full of shortcuts and keyboard-based interactions. This applies to any application - if you need keyboard interactions, you'd have to build them regardless of your choice of WPF or Web
I have read a range of articles on advantages/disadvantages of Microsoft Silverlight framework in comparisson to Flash.
Fact that there were two version of Silverlight in the past 18 months worries me, as well as the fact that over 97% of web browsers already have Flash pre-installed.
I'm a .NET developer and I'm very happy with what I can do in the framework. At the moment I feel like I have to put my own preference to a side and decide whether I need to integrage flash with .NET instead of using Silverlight with WPF.
Did anybody try integrating Flash with .NET? What challanges did you come across? How easy was it in comparisson to working in Silverlight?
I have also read about recent talks between Microsoft and Adobe. Whatever way I go It feels very unstable. Can HTML 5 really compete with what's offered by Flash and Silverlight?
Thank you
I have integrated both Flash and Silverlight into my ASP.Net applications. I have to say that working with Silverlight was by far the easier way to go. Communication with the server is easier, initial setup was easier (along with a ASP.Net application or a simple HTML page). Integration into the ASP.Net application was the easiest part (as I'm sure you imagined it would be). If you're a .Net developer, then this is the way to go.
Also, I wouldn't worry about either of these technologies going away anytime soon. HTML5 is not going to take over. It'll take some of the market share, not all.
EDIT
Here are some links to other SO posts on the subject:
Which is the future of web development: HTML5 or Silverlight(or other RIA framework)?
Should I Abandon Adobe Flash for HTML5 and ?
Can HTML5 do most of what Flash does today?
This is basically the way I pick web technologies:
do you need to support every available platform made now, in the past and in the future? HTML4 (yes, 4!)
Do you have to make stuff that is not possible with HTML4, and are ready to sacrifice some user base? HTML5
Can you limit yourself to desktop users (no mobile devices) and need a very interactive application (more so than a "website")? Flash or Silverlight
Do you need to integrate with .NET? Silverlight
Are you familiar with .NET (more so than Flash)? Silverlight
Do you want to reach the absolute maximum number of users? Flash
Do you know Flash/ActionScript (more so than .NET)? Flash
Do you need even more features than what Silverlight provides, and can you limit your deployment to intranets? WPF/XBAP
The whole debate around Silverlight being dead is completely flawed in my opinion: those that thought Silverlight would be completely multiplatform really were not connected with the reality: just by looking at Flash it was clear from the beginning that iOS & co. would never support Silverlight.
HTML5 will probably be the real "universal multiplatform" environment (what is HTML4 now), but with all the nightmares we all know of html. If you are targeting desktops and need more interactivity, better tooling, unit testing & co. then Flash and Silverlight will still be the first choice.
Silverlight's user base is quickly coming close to Flash, so between those two it really comes down to the one you know better and the ease of integration with an existing backend.
The fact that Silverlight got a new major version every 9 months until now is just a sign of how much they are pushing it. Now that it's reached a mature stage you can expect to see larger intervals.
Frankly, AS3 is not hard to learn if you already know OOP. It will take a week or so. If you don't like all the frames stuff in Flash, you can create a single frame app and then manage everything from your custom AS classes.
I'm also a .NET developer, and I had no trouble learning AS3.0. Of course, one week is not enough to become an expert (it takes years to become an expert in any field). But if you simply need to create video or mp3 players, create drag and drop basic games/apps to add to an ASP.NET page, it's worth spending 20 or 30 hours on AS3. There are great video trainings out there . Seven or 8 hours training should take the 20 to 30 hours I mentioned. I went for AS3.0 a few years ago, rather than SL, simply because everybody has Flash plugin installed.
AS3.0 is typed (simple types like Number, String etc), but at least it's typed. There are plenty of functions, classes and methods allowing to implement hit tests, drag/drop, event listening (mouse events, keyboard events etc). Really cool and fun language.
Take care.
HTML5 has SVG and Canvas and video.
It's perfectly possible that at one point someone creative is going to create a very nice animation package that generates SVG files. There already are SVG generators there of course, but obviously they're not good enough because SVG and Flash is, so far, never mentioned in the same sentence.
But is ought to be possible. SVG does structured vector graphics, embedded scripting. The things you see done in Flash has to all be possible in SVG. Flash also has this awful notion of frames, which was a major design flaw from day one.
They should have just let you determine that you want to move an object from point A to B along a path determined by a line or curve or freehand path, and that the times of A and B can be anything and not just confined to a particular frame.
Then at playback, the faster your computer is, the better the frame rate ends up being. Slower computer, slower frame rate. As long as the object moves from A to B.
Then there are the bugs and just overall clumsy handling of Flash.
Flash can be done so much better.
So, I think that someone will at one point soon make an amazing SVG animation package that will just crush Flash.
I'm against Silverlight because it's Microsoft. Microsoft means proprietary. They do whatever they want to do. You've already mentioned different version numbers. This means your customers have to have the right version downloaded. You can count on your customers having to download major upgrades, and before you know it their entire .net install needs updating, before you can show your animations and applications.
Silverlight also doesn't work on Linux. It's supposed to, through Novell's efforts with Mono, etc, but in practice, in the field, it just does not work where you need it to.
I don't know if and how well, in practice, Silverlight works on the Mac, but I don't trust it.
Eventually, I think, that future HTML5/SVG (Canvas maybe?) is the way to go. It'll even do 3D using OpenGL accelerated graphics... (but I don't know if that's portable enough).
In the meantime, Flash is your safe bet, and it's almost guaranteed to run anywhere.
I wonder if anyone has created an animation package that outputs to swf files that's better than the Flash IDE. Shouldn't be too hard, given Flash IDE's clumsiness.
Microsoft recently announced a "change of direction" with Silverlight with more emphasis on mobile rather than desktop.
Flex/Flash and SilverL. now are very similar... using webservice for the clientt/server comunication you can work well with both.
Sure, for .NET developer with Visual Studio to use SilverL. is much more FAST and you need C# only.
But, Flash is more available on PCs, also for mobile devices... you could think for Flash if you need portability.
My concern is that novice users will turn their back to a website which asks them to install Silverlight.
One of the reasons I think they might be scared of installing Silverlight is because they are not aware of what Silverlight actually is.
What's your take on this?
You can see the website riastats.com for information on install base, the main thing is if Silverlight gives an advantage such as the NBC Olympics coverage in United States uses Silverlight because of the Smooth Streaming technology plus the ablity to have a Live DVR where live streams can be "rewound" - all these features are probably better supported by Silverlight has the Client and Server ends can be well integrated (with Microsoft solutions).
I may be a little biased as I am a Silverlight developer, but with a less than 5MB download for runtime, that helps sell it more - especially as it may be a better alternative than the Flex download for Flash if this is required to create a solution that would have worked in Silverlight as it is a larger download.
Check out the silverlight.net showcase you'll see who and where it is being used - you can then make your decision based on this - rather than a few opinions (including mine!).
Well Silverlight is installed on about 50% of pc's now. So it depends on your market. Normal web users (ie. non-developers) tend to install anything you prompt them to if you can clearly and succinctly give them a compelling reason to run your application.
It doesn't tend to matter outside of the work place environments if you use silverlight or flash. What's more important is that your application will solve a problem the user has, and you can communicate that well.
They do not know what is is, marketshare is around 40%. I'd say it's not ready if you want to reach everyone.
However if Flash or javascript is not an option, give a firsttime visitor window informing your clients about silverlight.
I think you're going to have a lot of users on the public internet that don't have it installed. If you're really going for a mass-market site, and you aren't doing a crazy amount of media-rich type content, I would go with something like JQuery to make your Javascript easy and cross-platform.
I've been developing with XHTML, CSS and Javascript for about 4 years now.
I love it a lot and hate it a little. I've looked into Flash and Silverlight a bit, but as a developer, I'm not too keen on them.
One reason is that they lock you into a vendor and generally, into using that vendor's tools. E.g. Adobe Flash or Microsoft Visual Studio, etc.
Also, Silverlight seems to mix content, layout/styling and behavior and into a single markup language, whereas I like the XHTML way of separating them out in code, but bringing them together in the user's web browser.
I also applaud the usability of the web, e.g. back button, hyperlinks, etc. which are set-in-stone standards that people are used to dealing with.
However, I'm seeing a lot of industry support for Silverlight and Flash. As far as .NET Developer jobs, I'm seeing less jobs for front-end/.NET developers and more jobs for Silverlight/.NET developers.
Will HTML developers still be employable in the future, or should I consider moving to a proprietary platform such as Silverlight?
While Flash/Silverlight skills may be worth developing, I think you will find that general web development skills will still be required for some years to come. Mobile apps in particular seem to place more emphasis on good, basic web design without dependence on plugins and or client-side code. Eventually, I would expect web standards to evolve to subsume the best (or at least most used) features of proprietary plugins. The web, at least, seems to be a place where people tend to favor solutions that maintain independence over lock-in to specific vendor technologies.
No, I think that idea will never fully catch on. The problem is really about the platform being developed on.
Look at how accessible the web is. Almost any machine can get on the web. My phone, my iPod, my laptop, my 11 year old PII machine, my gaming tower, all can access the same web.
The devices I have are not the limit to what can reach the web either. I think just about every gaming platform and cell phone can get on the web, as well as thin terminals running any OS imaginable. I'm sure there are others also.
The big thing looks like it's going to be the mobile market in the next few years. Some mobile devices can run flash, but it isn't used much because of the poor support & performance. The only way that the mobile web can work is by using pure standards based solutions, because that's really the only baseline that can be trusted to exist.
No matter what proprietary technologies come out, I can always rely on the fact that my XHTML pages will still render successfully on whatever device decides to access it. The same can't be said for flash or silverlight.
At the same time, I can also guarantee you that there will be a bigger market for flash and silverlight because the web is becoming more "media rich" in some niche markets (YouTube, Adobe Air, Hulu, Google Gears, etc. to name a few examples). There will absolutely be a market for it, but I wouldn't say it will defeat XHTML and web standards because the web is constantly being redefined.
No matter how much Flash or Silverlight try to take on, the technology will move so fast that the only baseline that I think will remain will be standards like XHTML and CSS.
Flash has been around for years and still hasn't taken over. I think that is one good example of how hard it is to replace XHTML.
Go for server-side development of any kind, but I wouldn't become a Silverlight or Flash specialist.
<CrystalBallMode>
To be honest I can't see it happening. Other than the reasons mentioned by tvanfosson and DanHerbert, the XHTML + CSS + JS stack just grew mature enough so that things like AJAX and jQuery make pretty much all the lightweight client side stuff easy with these tools (as opposed to things like streaming video, heavy computations or sockets etc.)
Common technological inertia will just guarantee that the existing things will stay around. People are much more likely to use something that has been around for a while and has been extended to meet the latest requirements than to use something totally new. Of course there are great paradigm shifts every now and then like the native to managed code transition but I don't see that happening with Flash or Silverlight.
</CrystalBallMode>
My hope is that what comes out of all of this is a new standardized web platform truly suited to building the web applications that people want to see with tools that developers really want to use. I see all of the effort going to trying to shoehorn these legacy web technologies into the "Web 2.0" model and I just wish that this effort could go towards making a truly revolutionary "Web v.Next".
Don't get me wrong, I really like what jQuery is doing to make Javascript client code easier, but it's still Javascript and my personal preference is to work with strongly typed languages with productive development tools.
In the meantime, I think tools like Silverlight and Flash have a lot to offer and help you do things more easily in some cases than in other web technologies, and there are some things you simply can't do any other way. But I don't think Silverlight or Flash or any other technology is the end game, just a step in the right direction.
Consider for a moment that you can manipulate a web page using Javascript, (X)HTML, and CSS with a great deal of overlap in functionality and yet ALL three technologies remain in prominent use today. The reason for this is because all three languages are different tools meant to solve different problems and no one of them can serve as an adequate replacement for the other.
Its the same thing with Flash / Silverlight vs these existing web technologies. In fact, I work in a dev shop that builds Flash based e-learning. One of our current products was originally built to use a purely Flash-based solution for navigation, etc. However, as the product has continued to evolve we have actually moved a lot of the functionality from the Flash-based e-learning module and into regular html pages.
In other words, I don't think that we'll be abandoning the current tools that web developers use any time soon. For the most part I see Flash / Silverlight as additional tools that will solve particular problems better than we were able to solve them previously.
Neither one is going to win out anytime soon. I expect which one is used will depend entirely on the purpose for many years to come.
The reason you're seeing so many job offerings for Silverlight of late is because it's a relatively new technology and just recently gained some momentum.
Though, I do expect Silverlight to make quick work of Flash.
I sure hope so. And yes, I think they will. There will be some development on legacy (XHTML/CSS/JS) apps for re-tuning purposes, but I think there will come a day when new apps are simply not created on those platforms.
Mobile phones are the issue right now. Flash isn't available on many of the major phone models. And their browsers are all over the map. Luckily there's Webkit (iPhone and G1).
If Silverlight makes it to a web platform then it will be a nice viable alternative to the hodgepodge of technologies that are currently in use. FYI, Microsfoft says Silverlight on Android is very possible. On the iPhone, hard to say, Apple is weird about such things.
AOL recently created a RIA version of it's email client in Silverlight. Looks nice and there's no Javascript errors to worry about. From a developer standpoint, that's huge.
Since the launch of Silverlight 2 I was expecting a lot of full blown Silverlight applications popping up but still there seem to be little evidence of this. Does anybody know of such applications out there in the wild. And also what would be the obvious applications you would develop in Silverlight. I would say mail clients are bad examples as they just as well could be written as a web/ajax app. As Silverlight is far more powerful than web+ajax possible candidates should be impossible/akward implementing as a web/ajax app.
The ones that comes to my mind is
Photo and imaging editing apps
Reporting applications
Office applications, Word/Excel...
Edit:
Added from posts
Games
The point isn't that the app need to fill the whole screen just that it isn't just a small part of a webpage, or you could call it a full blown application running inside the webbrowser, only using the webbrowser as a host.
I think the Medical app that Microsoft itself developed shows pretty well what could be achieved with silverlight http://www.mscui.net/PatientJourneyDemonstrator/
As for image editing then as I understand its a bit difficult as Silverlight lacks a Bitmap API to be able to do per pixel image editing...
Edit:
I noticed you added Word/Excel to your question and there comes the problem that Silverlight doesn't have a rich text editor built in and there hasn't been real good examples of custom implementations. There is one http://www.codeplex.com/richtextedit but I haven't seen any applications that actually use it.
I'm working on one in the medical domain.
This started as an update of a Mac classic application but due to the amount of work involved, broadened to considering other toolkits. I convinced them to go for an initial WPF desktop port to be followed by a Silverlight version.
I don't know one so far, but I could imagine that it could be used in a kind like the fullscreen video playback on youtube.
How many fullscreen desktop apps are there? Most application don't need the entire screen. If you don't want to be distracted by menus and taskbars and so you go fullscreen. Another type of applications that can use fullscreen are games.
You are limited in fullscreen to certain key presses such as arrow keys, tab, enter, and space so this rules out some of those types of apps. They have done this for security reasons so an app can't hijack the screen and record the keypresses, but I wish they could come up with a scheme to sufficiently warn the user then allow it if they consent.
An application Microsoft seem to like to show case is the AOL mail client written entirely in silverlight.
Personally I follow the rule is if you would not write it in flash you would not write it in silverlight preferring AJAX in most cases. In the past most large flash application have failed such as the flash word processor (cant remember the name) while AJAX enabled applications such as google documents have taken off.
Finally I believe until moonlight (linux and mac support) has been released and more general users have silverlight downloaded developers will be reluctant to use it widely even for smaller apps and gadgets.