Channel 9 Live at MIX10: Dean Hachamovitch on IE9

What about WebGL? Is IE9 or future browers from Microsoft ever going to support it? If not (for obvious reasons ), are you guys planning for something for 3D on IE which doesn't require a plugin?

Good to see Giorgio Sardo doing cool things! Cheers mate!

I asked Dean this question at PDC 2009 and the answer was something like getting 2D first was important. I can certainly understand that. A full SVG implementation from one version of IE to the next doesn't look like a small task. Still, 3D in the browser will be very interesting.

Now as for background compilation: how does the CLR handle this? Isn't the CLR also doing background compilation? There is now background GC, so background compilation would be a nice complement. In particular since it always seemed to me like the CLR limited the kinds of optimizations it applied because it couldn't do them in time - it's RT/JIT anyway - but with a combination of JIT and background compilation there's less of an excuse.

Oops, when I asked about 3D on IE, I completely forgot about XAML I think it would be great if someone wrote a library to convert WebGL to XAML. No?

Charles, there's nothing wrong with D3D (in fact pretty much every 3D application that I run on my own machine is more performant in DirectX mode than OpenGL - possibly chalk that up to the driver support for my older GPU or the coding chops|unequal-experience-with-both-platforms of the authors or the apps), but that's not the point.

The point is that we're talking about a web browser... for loading the worldwide web. It doesn't matter if we could get faster 3D apps (or bigger 3D apps running at the same speed) if IE9 had built in support for D3D rather than built in support for OpenGL (or its little brother WebGL). The point is that Microsoft will never program and support DirectX and Direct3D for MacOSX or Linux or Android or anything else. Neither will you open source it so that the people responsible for these other platforms can make it work for themselves (as I assume that OpenGL does). Microsoft will never make IE9 for anything besides Windows Vista and Windows 7. As much of a market share as Windows has, that excludes a huge number of people as far as proposing to web developers that they use this (theoretical) built-in IE9 D3D support to build 3D web apps.

Apps written in WebGL's subset of OpenGL are going to run in Safari, Opera, Firefox, and Chrome on Windows, on MacOSX, on Linux distros, and on a heap of mobile devices. Apps written in D3D are not. Again, it's not a question of which technology is better, but what is going to actually be usable by people on the platform of their choice.

(Optional rant, available at the bottom of this entry, to be inserted here if so desired.)

Now if all that you meant is that the IE team would be calling D3D behind the scenes in order to run WebGL web apps in the same way that Silverlight for Windows calls DirectX for hardware acceleration for Perspective 3D and HD Video playback and Silverlight for Intel Macs calls OpenGL for hardware acceleration of those same things to provide reliably equal experiences, then that's fine. I just meant to point out that all four of the other major browser vendors are onboard for WebGL and are going to make it a reality, whether IE stays relevant or not. It's going to be a lot less work for IE9 to support WebGL than to have all four other browser vendors have some way to interpret D3D web apps into WebGL apps or whatever would need to happen. The point is that these other browsers care about working on every platform and care about these major features working across platforms whereas Silverlight calls itself cross platform but as of yet the only real platform other than Windows that runs the feature-complete, up-to-date version of Silverlight is Intel Macs, so that even if Silverlight 5 has a major breakthrough in 3D graphics, WebGL will have a much farther reach and deliver much more of this "code once, run everywhere" dream. Moonlight is sadly always going to be a year behind Silverlight because Microsoft understandably does not publish the Silverlight source code to Novell (nor do you take the approach of simultaneously providing an official Microsoft coded Silverlight runtime for Linux as well as allowing Novell to continue to reverse engineer Silverlight with Moonlight into an open source implementation with their one year lag). This means that if authors of Silverlight apps are keeping their apps up to date with the latest version of Silverlight, then Silverlight apps will never work on Linux. That is incredibly frustrating for people who want Silverlight (and consequently Moonlight) to succeed.

While I'm on the Silverlight rant, can I just ask why there isn't any Codeplex project for porting Moonlight to PowerPC MacOSX? I would be stoked to see that being encouraged here on Channel9 - relying on the developer community here to extend Silverlight support past where Microsoft is willing to go. (Honestly, Silverlight has been ported to PPC for the XBOX 360 and has been ported to OSX for Intel processors - why not just combine those two pieces of work and finish the job?) The Moonlight team is already stretched too thin trying to keep up with Silverlight's current iterations to bother about also porting themselves to another platform, but someone should definitely do it. The Moonlight source code is just laying there, ready to be worked with. If I were a coder, that's exactly where I'd be spending my time.

Moving back to the issue of what will be the web standard for 3D, WebGL does seem to be the clear winner, but I don't know that Google has given up on its own O3D either.

Am I wrong? Please discuss.

Warning: the following is a tangent and a bit of a rant. Skip it for logical flow.

People already have this (infuriating) perception that Microsoft = suck (e.g. Windows Live Messenger for Mac). Consequently, having something awesome that only runs on Windows does not generally get the reaction of, "I have got to get a Windows 7 PC!", but rather, "Great, Microsoft broke this innovation so that it only works on Windows!". Examples include: 

1. Photosynth. Even though the Photosynth viewer switched development from D3D to Silverlight last year so that Macs can view synths as well, the synther which you actually make synths with remains Windows only. Considering that it is a C++ application, I really don't think that it's any technical issue that's stopping the team from porting the synther to OSX (regardless of the fact that the synther uses Windows Imaging Component and other things that are not on OSX and would have to be wrapped into the download or... less attractively for purposes of updating the synther... ported to use something completely different on other platforms). Is Microsoft allowed to have a killer app that only runs on Windows? Sure, but that just makes it irrelevant to people who aren't going to buy a copy of Windows for Parallels or Boot Camp. Essentially it's a opportunity lost to make Mac users think well of Microsoft and what is more an opportunity lost to use Mac and Linux users to strengthen Bing. To me, it reeks of insecurity too... amounting to, "If we make our most popular software for other platforms with full feature parity of their Windows counterparts, people might not buy Windows anymore!". Your approach should be more along the lines of making sure that people have a great time every time that they open a Microsoft program... a better time than when they used any other software that day.
   
   Consider this too: Photosynth is a part of Bing Maps' division. Bing Maps' competitor to beat is Google Maps. Google Maps, although it uses a proprietary plugin to run the Google Earth plugin, has done the work to ensure that that proprietary 3D plugin works on Windows, MacOSX, Linux distros galore, and coded custom apps for multiple mobile device platforms. Google doesn't care about anything except that people use Google and will stop at nothing to make sure that people are able to do so. Likewise, they are aggressively synthing photos from all across the web against their Street View panoramas. Their street level panorama coverage absolutely dwarfs Bing Maps'. Because of this, they can offer Flickr, Panoramio, Picasa Web, and general users of the internet worldwide the ability to see their geotagged photos on Google Maps today with no effort on the photo uploader's part. With Photosynth, you absolutely must have a Windows PC. You can't contribute from your Phone. You can't contribute if you prefer using a Mac or Linux. This is really just shooting yourself in the foot. Google will take photos from everyone and make itself stronger. Bing, while interested in this, doesn't have the panoramic coverage to make it happen. In many parts of the world, Google will beat Bing to market with this synthing of web photos onto the street level panos by a time difference of two years. Making sure that only Windows users can create Photosynths doesn't make Photosynth or Bing stronger. It only ensures that for many people, the first time that they see photos registered in geo-3D will be courtesy of Google. It's crippling Microsoft's ability to compete with Google when you only allow Windows users to make Bing Maps stronger via Photosynth while Google ensures that Sketchup and server-side synthing works for everyone. 
   
   
2. The same case could be made for the Zune software. That is one beautiful piece of code, but of course it doesn't run on a Mac. Public's reaction: Zune sucks because it doesn't come where I live [i.e. on a Mac; on Linux]. In fact, not only does the official media manager not support anything but Windows, but there's no other way to simply copy your files onto Zune like any other USB storage device (what does this mean for Windows Phone 7?). You have to go find some sort of hack to just put your media files onto your media player if you don't run Windows; how insane is that? Look at Zune's biggest competitor... oh wait, what's this? Their media manager supports their arch rival's OS? 

We can talk all day about how Apple has no choice but to make iTunes for PC because of Windows' awesome market share but at the end of the day Apple is making Microsoft users happy with Apple. Down the line that can easily translate into users leaving Windows for Mac. For contrast, most Microsoft products for platforms other than Windows are... not so hot, compared to their Windows counterparts. Office and Messenger for Mac, I'm looking at you. (Silverlight is the exception to the awkwardness, although the Silverlight for Mac installer is crazily complex compared to the Windows installer and even other Mac installers.) The bottom line is that those sorts of efforts don't generally please people so much that they are head over heels loving Microsoft and drop their Mac to run out and buy a PC. Microsoft's most influential evangelists to end users are their products for Mac (or Linux... if there are any) - Office, Messenger, Silverlight, and perhaps most of all in more recent times, Bing (Bing actually happens to get everything about right, I'm happy to say). They alone are what determine incentive for people to buy Windows 7. Happy Windows 7 commercials will be ignored if the last Microsoft program that people used on their current platform was inferior in any way. (After all, if "Microsoft" can't build a decent instant messenger (Mac users' impression) then why would I want an entire operating system or anything else from "Microsoft".) You want each Mac program Microsoft puts out to be making users think, "It would be so incredible if these guys made an entire operating system. I'd go buy it in a heartbeat.". You want people to not want to close your program because it's so great, not some sort of 'poison the well' strategy where you make sure that the Mac version of Messenger is so terrible compared to the Windows version that people finally cave in and get a PC. I mean, what kind of strategy is that? How do you hope to compete with other instant messengers on the Mac like Skype with a strategy like that? Make sure that your management for Mac development understands this concept.

Good rant Nate! I love your effort at elaboration!

Yes, a lot of cool technology has come out of Microsoft, except I don't know why many of them die a sudden death Probably because of the fact that Microsoft is trying to restrict their usage to the Windows platform. But with the web and web apps coming of age, I doubt that it is a wise decision - the OS might not remain that important and essential anymore. Time will of course tell.

While it is asking a lot of Microsoft to themselves port their applications to other platforms, it does make sense for them to play along with others when it comes to the web.