Zoltan Arvai, a Microsoft MVP in Client applications development, is a regular contributor to the article content on SilverlightShow. Having authored 7 articles on MEF, MVVM and WCF RIA Services, and with more articles coming along the way, Zoltan is one of the most appreciated and highly rated authors on SilverlightShow.
Meet Zoltan!
Q. Zoltan - please briefly introduce yourself to SilverlighShow community
A. Hi! I’m Zoltan Arvai from Hungary. I’ve been a .NET developer for 7 years.
A couple of years ago I came close with Microsoft Hungary, and started doing workshops and presentations for them. This is where my trainer carrier have started. One of the biggest Hungarian training center, called Szamalk Zrt., gave me a job offer. By that time, I was working with EvoPro Hungary Ltd so a full time job was just not possible for me and the whole training thing was pretty new. Couple of months later I fell in love with teaching but I wanted a freelancer status so I could have my own classes and own projects. So right now I’m a freelancer, developing Silverlight, WPF applications, consulting and teaching here in Hungary and sometimes, to my biggest surprise, abroad.
As opposed to millions of people I love working, developing on the bleeding edge. It is very important to me to be up to date with current technologies, even with betas, to have a very broad view on the developer platforms and possibilities. I’m always trying to do and learn something new, so Silverlight with its 8-9 months release cycle is pretty ideal for me. :)
My other pet is Team Foundation Server. It’s an exciting and a very advanced product. I’m addicted to methodologies. No matter what form of methodology we are talking about. Maybe it’s architecture or just project management, controlling the team work, I like to keep these things on the right track. I prefer clean solutions. My newest pet is the Windows Phone 7 platform. ”Great future, I see.” as Yoda would say. :)
Q. How did you decide to start out with Silverlight, and what helped you reach your currently level of fluency with this technology?
A. I’ve been working with Silverlight since the very beginning. Oh no, wait, even further back. As a matter of fact I started with Avalon. Avalon was the project name of WPF. I started experimenting with it very soon, maybe in the pre alpha times, or CTP I don’t really remember. By then it took me an hour to put a grey rectangle on the screen and write hello world with red on it. As you know WPF debuted as part of the .NET Framework and proved to be very effective when it came to UX design. Even Microsoft liked the concept so they created the little brother, WPF/E (WPF Everywhere, Project name of Silverlight). Well the 1.0 version was a little glitch in my Silverlight carreer, by that time it was nonsense to me.
Javascript is something, let’s say I don’t prefer. But I really liked the 1.1 extremely buggy alpha version, since it projected the future. As a WPF developer it wasn’t that hard to move to Silverlight. By that time I knew that this was going to be a successful story. Ever since I’ve been working with it, teaching it in classes, talking about it on conferences, writing articles and tutorials about it. Taking your skills beyond ”hello world” in Silverlight takes practice. But the best you can do is to surround yourself with people who are smarter than you, people from whom you can learn. There are a quite a few books, presentations, screen casts, blogs, and articles out there that can really help you. SilverlightShow has a huge amount of very valuable information.
Q. One of your recent articles covered Push Notifications in Windows Phone 7. What are your thoughts on Windows Phone 7? Have you already developed anything for this platform?
A. Well I think Windows Phone 7 is probably the best thing that Microsoft could do in the current situation. Future and current trends are very clear. Smartphones will definitely be the most important platform ever. You know smartphones are personal. My phone is with me almost all the time, it knows where I am, in some cases, what I do. It knows what I like what I prefer, it can hear me, see things, it knows what and who I’m interested in, what my friends are doing. I’m always connected to the internet, to my ”digital identity”. It can warn me on the way home not to forget to drop by in the closest supermarket since I’m out of fresh milk or something like that. :) It can also offer me related products and tell me about discounts, etc. It’s definitely going to change our lives.
Now Windows Phone is on the right track. Microsoft finally realized that consumer phones must be created and user experience is the differentiating factor. So they trashed Windows Mobile and now we’ll have great phones and, thanks to Microsoft, we can develop applications for WP7 in very familiar environments and platforms. Silverlight and XNA is just the perfect choice I believe. I also think that it is going to be very hard to cut a bigger slice from the smartphone market share. Google’s Android based phones and Apple’s IPhones are very popular choices and, let’s admit it, they are also great. The healthy competition will do some real good to this market. I believe firmly in this platform, so the answer is yes, I’ve written and still working on some projects, both XNA based games and Silverlight consumer applications. The development experience is pretty great. I could transform my ”Desktop / Web” developer skills extremely easy to be a WP7 developer.
Q. Is it a problem that you don't have the actual device currently? Is the emulator enough for developing and testing applications?
A. Well, not having the device itself is definitely a problem. Some of my WP7 Silverlight based applications were easily developed on the emulator without a problem. Nevertheless I will definitely want to see them on an actual device before I publish any of them. Some of my ideas were immediately cancelled because of the heavy dependency of sensors. Let’s face it, simulation of the sensors, however it might be possible, is absolutely not satisfying or usable at all. As for game development I actually designed the game with the fact in mind that I don’t have the actual device right now. Performance is also a key point here. How good is the job the emulator does, when simulating the performance constraints? So we definitely need developer devices to make sure we are releasing quality products.
Q. You wrote an article about MEF. While that's a pretty powerful framework, developers are still pondering over why they should use it, how does it differ from PRISM, how do they use it side by side, etc. What would you suggest?
A. That’s not a surprise. The way I saw it in the beginnings, even Microsoft had problems positioning MEF right. MEF is a pretty interesting technology. It was designed to support extensibility and plugin-like scenarios but the technology was picked up in a lot of different area, like when some people use MEF to loosely couple the View and the ViewModel. It became very popular in a very short time. I think today we overuse it a little bit.
First of all I have to pin down that PRISM is not like MEF. PRISM is a very good library to support MVVM, communication between modules, with event aggregators, region management. MEF does not support any of that. PRISM is using Unity as an IoC framework, this is the part where MEF can come into play. Both can be used to build modular applications, PRISM suites this scenario the best. MEF, on the other hand, can also be used for later extensibility, in plugin-like scenarios. MEF suites this scenario the best. But if you want the same toolbox like the one that Prism offers you’ll have to add extra libraries to do that. They definitely have some overlapping.
As far as I know in Prism 4.0 you’ll be able to use MEF for modularity while still leaving Unity as an option. This story has another aspects as well. MEF is part of the .NET framework 4.0 and also Silverlight 4.0 while PRISM is not. There is probably no right or wrong here. Use the one you like and suits your scenario best. Here is my checklist. I want extensibility or modularity and I want MVVM and I also prefer the MVVM light toolkit, also I want to keep the xap’s size low. I go with MEF. I’m creating a composite, modular application, where I’m not planning to support plugins and PRISM’s MVVM support is good enough for me. I go with PRISM. If for some reason you have to use it side by side then I suggest to draw a line between modularity and extensibility. Use Prism for the previous one and MEF for the later one. But as I said before, in Prism 4.0 MEF will be used for modularity and a couple of other issues might be solved as well, so I’m really looking forward to the final version of 4.0.
Q. We have recently released a fully Silverlight-based version of www.completit.com - the corporate portal of the company behind SilverlightShow. However, porting an entire website to Silverlight is a step not many companies are yet willing to make. Why do you think? What else does Silverlight 4 miss to become fully trusted for LOB applications?
A. Really nice, good job guys! Well, one of the first issue is the penetration. At this time Silverlight 2.0 + 3.0 + 4.0 penetration is about 60%. It’s significant but it’s not enough. Most of the companies don’t want to risk that a potential customer might turn back when his browser displays ”Please install Silverlight”. Also Silverlight is not available everywhere. You can’t display it on Linux (Moonlight doesn’t really count, however it’s a superb project), you can’t display it on smartphones, at least not yet. Not all of the browsers are supported. At some companies the policy does not even allow the install of the Silverlight plugin. I know these are not huge slices of the market but the little numbers add up in the end.
Of course there is still the issue of ”why change something that works well”? Not to mention, that in many cases, business decision makers have no idea what Silverlight is. Plus there has always been a little resistance against Microsoft technologies in developer circles. However that really seems to fade away now. I honestly believe that Silverlight 4 has everything to be a trusted LOB application. The weakness is still the penetration, the availability of the plugin and a somewhat wider acknowledgement. I think trust will come, it’s just a matter of time. the platform is ready.
Couple of month ago I made a presentation about the future of the web, and about Silverlight. In the end I told the audience that I do not know what the Web 3.0 will be, but I do hope 4.0 will be a beautiful world without browsers. What I meant is, that in my opinion connected Out of Browser applications, and applications on smartdevices will take over the reign of websites. I really do hope so. This is where Silverlight can have a true Bing Bang.
Zoltan - thanks for the opportunity to introduce you to SilverlightShow community. We hope you will keep contributing to creating top-quality article content on SilverlightShow!