The Media Center Sandbox

Hi

I am new to Media Center development, so bear with me.

My intention was to investigate developing a media center addin built using XBAP or Silverlight platform.

I do not see much info on doing this most XBAP related material is timestamped in 07 or 06. In the 5.3 media center sdk, it states

The following features have been deprecated:

• Hosting for Microsoft .NET Framework 3.0 Extensible Application Markup Language (XAML) browser applications (XBAPs).

It looks to me that in order to produce a seamless addin for media center I would need to adopt the MCML route.

Is there any route that would allow me to produce XBAP or Silverlight bases addins.

Any comments or advise would be very welcome

Regards

Vinay

Hi Vinay,

MCML is definitely the best route - you can use XBAP at present, but there may not be support for it in future releases of Media Center.

Cheers,
Andrew

Hi

Thanks for your reply.

If I was restricted to having to display WPF or Silverlight based material in media center, what are my options? There is an existing code base that the client is looking to reuse as much of rather than adding another new (MCML) UI component.
I see host based addins being mentioned, but I get the feeling that this is regarded as legacy, would using Silverlight inside  a html page  be possible or is that just hacking.

Regards

Vinay

XBAP has been deprecated, so while you can use it now, there's a good chance that a future release will result in your app being broken.

HTML is still considered current - Media Center acts as a host for MSHTML so you can theoretically use Silverlight in an HTML based add-in.   Best thing is to try and see what happens.

Cheers,
Andrew

Hi Andrew,

Coming to this, I wanted to avoid having to spend time investigating avenues that others know to be unworkable. I also wanted to have my understanding of the current state of MC development confirmed. So thanks for taking the time to answer my questions Chris.

I am surprised to see XBAP being removed since, from a wider perspective, it would appear that WPF/Silverlight as presentation technology stack would be a natural companion to the presentation needs of a highly visual product such as MC. It also suprises me that there seems to have been so little outcry from developers that this has taken place. But those are discussions for another thread.


Regards

Vinay.

The thing to remember is that Media Center is an application first, and a platform second - this means that primarily the workings of Media Center are there to support Media Center, and extensibility for media providers to augment the solution with their own "pages".   MCML was born out of the internal markup system that Media Center has used since MCE2004, so pre-dates WPF and Silverlight by a long margin.  Also, if you have a look at MCML and how it integrates into Media Center, in my opinion at least it is much better suited to the job.  I look at what is involved to make my add-ins look great on Media Center, and using WPF I just can't get the same effects, or can't get them without a lot of effort, where in MCML they are effortless.

Cheers,
Andrew

Hey Andy.

I also asked me what is the current state of XBap. On a Conference I heard from a developer that Xbap will be absolutely thrown out of Media Center?! I also read this at certain Websites.

Today I see a plugin called Yavido which is developed in Xbap.
It's cool. A lot of animations.
Is it actual that Xbap will be unsupported in the future?


Currently I am developing an mcml Addin, but I doubt a little to start with it.

regards, tronic

tronic,

From earlier in this thread: "XBAP has been deprecated, so while you can use it now, there's a good chance that a future release will result in your app being broken."

So, yes you can use XBAP today and produce Media Center add-ins with it.  However, in a future version of Media Center, Microsoft will remove XBAP support and at that point your add-in willno longer work.

Cheers,
Andrew

Hi Andy,

I made a Windows Media Center Application for Vista using Xbap . But i read from some internet articles that Xbap support in MCE has been deprecated. that means my application will not support for newer version of MCE? What will be solution ,if microsoft deprecated Xbap support?

If microsoft want to deprecated Xbap , why they introduce Xbap ?

       

I don't have any inside knowledge on this, but this is my opinion.  

When the eHome team developed MCE2004/5 they decided to implement an extensibility platform.  They chose HTML as it had a large developer base and third party content companies could easily develop stuff for Media Center without any specialist skills.

With MCE2005 they had a lot of feedback from the developer community that we wanted to be able to create applications that had the same visual effects that Media Center had.  Media Center did this itself using the markup language developed by the eHome team that we now know as MCML (actually it was something like it, but not exactly it).   The team re-worked the language to make it usable by third parties, and also started re-working Media Center to use the new language internally (all of the horizontal scrolling lists inside Media Center are the MCML we use to develop, for example).

In the same way that HTML was used to grab a wide developer base, I believe that the eHome team implemented support for XBAPs as they anticipated it too would have a wide developer base (although at the time of implementation it wasn't even released).  I imagine that also it was reasonably easy to implement in the same fashion as HTML (via embedded MSHTML or something like that).  I also think that there was probably a push from some launch partners to allow them to use something that they were developing skills for elsewhere in their organisations, rather than having to learn a new skills set.

Once Vista Media Center had been in use for a while, the eHome team will have had to make some choices on development environments for the platform - should they continue to support three platforms and the cost entailed with that, or reduce the number to allow resources to be redirected elsewhere?   In terms of broadly supported technologies, HTML would definitely be there, but XBAPs aren't as prevalent as they were expected to be, and most serious Media Center developers are using MCML rather than XBAP because of the better integration.  

So my guess it was better to reduce the confusion in the range of competing environments available, direct more resources to the MCML development environment, and deprecate the XBAP environment in favour of something that worked better with Media Center.

Cheers,
Andrew

Also, with the caveat of having even less of an insight into the thinking of the Media Center team than Andy probably does...

I think it comes down to the question of remoting the experience (think extenders, a key component of Microsoft's strategy). The MCML renderer appears to have been designed from the ground up with remoting in mind, even to such an extent that it actually seems far more capable than what we currently see today. It's fundamentals are relatively simple, however it can be used to implement complex behaviors in unexpected ways while still maintaining a very small footprint for limited capability devices.

XBAP/WPF is more comprehensive, and certainly appears to offer more features (although many of the 'features' are actually just syntactic candy around largely similar base of primitives), however this comes at a cost on the client with regard to the fundamental platform capabilities and their cost of implementation.

At the end of the day, implementing a full fidelity MCML renderer on a device appears comparatively trivial verses implementing XBAP/WPF/XAML/Silverlight, and this is a very worthy goal (at least intellectually...).

Whether such capability is needed for what has essentially seemed unable to evolve beyond such a limited and closed ecosystem is actually warranted in hindsight is hard to say. I'm sure that if things had evolved along the lines the eHome team seemed to have anticipated/desired it would make perfect sense.

I would argue that, regardless of how it has been received to date, it is also still the only way forward that makes sense. If you want a renderer in every TV/STB/Tablet/etc. the best way to make it happen is to make it as cheap as possible to implement while still offering a compelling user experience.

James:

At the end of the day, implementing a full fidelity MCML renderer on a device appears comparatively trivial

I don't think it's that simple - just look at some of the more "popular" v2 Extenders available today, the animations aren't exactly smooth, and in some cases simply refuse to display at all. The Xbox360 has it right in terms of full fidelity, as does some other extenders but the cheaper models do not.

Speaking about extenders though, I'm still waiting for a little Windows client that'll act as one. Even if it required a fully fledged Home Premium, Ultimate or Enterprise Vista installation it would be gold worth for those wanting to setup a single VMC box acting as a "server" to share the tuners around the house.

emigrating:

Speaking about extenders though, I'm still waiting for a little Windows client that'll act as one. Even if it required a fully fledged Home Premium, Ultimate or Enterprise Vista installation it would be gold worth for those wanting to setup a single VMC box acting as a "server" to share the tuners around the house.

Oh emigrating, you do crack me up! :)

emigrating:

I don't think it's that simple - just look at some of the more "popular" v2 Extenders available today, the animations aren't exactly smooth, and in some cases simply refuse to display at all.

Heh, I'd love to look at some of the more popular V2 extenders... finding one in Australia is a bit challenging though, and I'm certainly not prepared to pay for one. As far as the animations on them go, well, from what I've heard some of these V2 extenders are kind of only halfway there. Instead of compositing the whole UI, they render the MCML to a buffer and then just use it as a video overlay. This means they don't qualify for the 'IntensiveRendering' capability bit and so a number of animations simply aren't triggered, and the ones that are are rendered at an unsatisfactory rate. Makes you wonder why they mandated V2 extenders... V1s could get the same sort of results using the pseudo-RDP they used.

I can't pretend to understand what's going on with all of this - when I really spent some time researching it in 2005 it looked like one of the new SoCs (from Cirrus Logic from memory) could do everything that was needed, allowing you to build a full-fidelity extender the size of a packet of cigarettes with a wall-wart power supply. It should have been pretty easy too - they could be a WinCE target and had a DX7 rendering pipeline.

At the end of the day, I think it's an issue with the 3rd party hardware manufacturers. They feel more comfortable shipping big boxes of air, thereby compromising the feature set due to the cost of shipping, manufacturing, and inventory, than they do shipping something small and sensible. Apparently they believe it gives consumers a better perception of 'value'...

If Microsoft Hardware shipped an extender I think you'd see something more in line with this. The fact that they don't is kind of funny (in a depressing way). My guess is that they don't see the market (and Microsoft, being the gloriously siloed business it is, they're not forced to), and so don't produce it. Sure, producing such a device might very well make the market, however that's not the way they appear to operate.

The Vista remote going MIA seems to suggest something's not quite right in the whole ecosystem...

emigrating:

Speaking about extenders though, I'm still waiting for a little Windows client that'll act as one. Even if it required a fully fledged Home Premium, Ultimate or Enterprise Vista installation it would be gold worth for those wanting to setup a single VMC box acting as a "server" to share the tuners around the house.

Ah, the legendary SoftSled, where art thou? There was compelling evidence that it already existed back in 2005, and nothing I've seen since then has lead me to believe that's not still the case. Unfortunately, I don't think it fits the 'Vision'. Microsoft's apparent goal is to get extender functionality into TVs/DVRs/STBs/etc, and they seem terrified that SoftSled will erode their ability to do so.

James:

Ah, the legendary SoftSled, where art thou? There was compelling evidence that it already existed back in 2005, and nothing I've seen since then has lead me to believe that's not still the case. Unfortunately, I don't think it fits the 'Vision'. Microsoft's apparent goal is to get extender functionality into TVs/DVRs/STBs/etc, and they seem terrified that SoftSled will erode their ability to do so.

The original extenders for MCE2005 were codenamed "Bobsled" - I think that all SoftSled was, was an in-software version of the extender for testing the Media Center code, rather than a product-in-planning, as the hardware wasn't available to test with.

Cheers,
Andrew