I promised some more comment on the conference on XNA game development last week. It was a fascinating event, not so much because of sparkling content (though it was good), but because of the underlying themes. This conference was for the academic community, both students and lecturers. The idea was to promote the adoption of XNA on computer science courses. Presenters included members of Microsoft’s XNA team, Nick Burton who is a graphics programmer from Rare, and Peter Molyneux from Lionhead.

XNA is Microsoft’s new framework for games development. It is a .NET wrapper for DirectX, and a replacement for the earlier Managed DirectX libraries. You can create games for both Windows and XBox 360. The initial target is student and hobbyist developers, and the development kit is free except that to deploy to an XBox 360 you have to join the XNA Creator’s Club, for a modest subscription fee. Even then, you cannot distribute your 360 game except as source code to other club members; you would need to find a publisher or get the game on XBox Live Arcade. A future XNA studio product is planned which will target professional developers.

There is an obvious benefit to Microsoft if it can establish XNA in universities. It gets students familiar with the Visual Studio/Windows platform and helps foster the next generation of Windows and XBox 360 game developers. But what’s in it for the universities?

Here the context is important. There is a crisis in computer science, with dramatic falls in the numbers of students applying for courses. I spoke to Rob Miles, lecturer in computer science at the University of Hull. “The number of people applying to courses across the country has dropped by something like 40% over the last few years,” he told me. Including games programming in the syllabus is a bit of sugar to attract new students. “It will help me motivate students to want to learn to program,” said Miles.

Microsoft says XNA will help to address another crisis, this time in the games industry itself. Developing a game has become extraordinarily expensive. The industry has moved towards the Hollywood model of big-budget titles that can make or break their publishers. Gamers on the other hand complain of too many sequels, lack of innovation, and poor gameplay. Microsoft likes to talk up XNA as the future YouTube of games, recovering the community and excitement that was around thirty years ago, in the days of the BBC micro and other home computers.

Here XNA is competing with Flash, which already has a considerable community of game developers. XNA is vastly more powerful than Flash, but for small, easily deployed games Flash has many advantages including cross-platform support and browser integration.

Is XNA just another hobbyist game programming tool, likes others including AMOS and STOS (for the Amiga and Atari ST), or more recently DarkBASIC? There is a range of opinions here. Peter Molyneux enthused to me about XNA, saying it is “a fast, efficient, better language” than earlier more compromised tools, and “built with ambition in mind,” so that XNA programmers can create the next Populous (the game which made Molyneux his fortune).

On the other hand, Molyneux undermined his evangelism by also stating that C++ remains the language of choice for professional development. If this is the case, then XNA will always be a hobbyist niche, which implies that serious students of game programming should not waste too much time on it.

Not everyone takes this line. The Rare team has spent time with XNA and appears genuinely impressed with its capabilities. Nick Burton showed us both the dramatic reduction in code and the impressive performance of XNA code. While Rare is not yet using XNA commercially, it does expect to do so in a few year’s time. This might be some sort of mixed code, using XNA for game logic and C++ for graphics heavy lifting, or might be pure XNA. A factor for Rare and any professional game studio is the extensive C++ code library which it has built up over the years, which makes any quick move to another platform out of the question.

My view is that the productivity win makes XNA or something like it inevitable in games development at every level. Timing is another matter, and C++ will be hard to shift.

Finally, a couple of hot snippets of game gossip. One is that Rare is working on a Live Arcade title. I got the impression that it is fairly well advanced, though nothing is yet formally announced.

The other is that Molyneux talked up innovation in Fable 2. He told me that it has an unique and surprising feature which he had to fight to include, and which will be unveiled in March at the Game Developers Conference in San Francisco.

Microsoft’s XNA site is here. Note that if you want to develop in XNA, you are well-advised to get hold of the wired XBox 360 controller. It is a decent controller and works nicely with XNA both on Windows and 360.

Yesterday I attended a conference on XNA game development. More on that soon; but first a quick note on what happened when, fired with enthusiasm for XNA, I installed the release build of XNA Game Studio Express on Vista.

At least I tried. You have to install Visual C# Express first (since for some reason XNA does not work with the grown-up Visual Studio 2005). That’s OK since it works side-by-side. Visual C# Express went on OK, but setup for XNA Game Studio Express failed towards the end of the install with an error about a file that could not be run. The install then rolls back completely leaving you with nothing.

I ran msiexec with logging to see more detail. Here it is:

Error 1721. There is a problem with this Windows Installer package. A program required for this install to complete could not be run. Contact your support personnel or package vendor. Action: RegisterWithCSExpress, location: E:\, command: “C:\Program Files\Microsoft Visual Studio 8\Common7\IDE\VCSExpress.exe” /installvstemplates

It occurred to me that even an install without the templates might be of some use, or perhaps I could manually install the templates. So I ran up Orca (MSI editor), opened the InstallExecuteSequence table, and deleted the RegisterWithCSExpress action (yes, I’ve wrestled with MSI before).

The hacked install ran OK, and to my surprise the templates were present. This might be because of the earlier failed install, or because I also ran:

VCSExpress.exe /installvstemplates

from the command line. This command, which tells VC# Express to refresh its template cache, is the one that fails. On my system it causes VC# Express to crash, which is why it trips up the XNA install, but it appears to install the templates before it crashes so it is kinda OK.

My official advice therefore is not to try and install XNA on Vista, but the above may help if you are determined.

Microsoft knows about problems with Visual Studio on Vista; we are promised a “Vista Support Update” next year. In the meantime it does mostly work if you run it with elevated permissions.

I noticed that the Microsoft folk were running XNA on Vista yesterday.

Microsoft has posted a CTP (Community Tech Preview) of Windows Presentation Foundation Everywhere, a cross-platform browser-hosted runtime for XAML, its XML GUI language. There’s a download for Mac as well as Windows, but sadly one of the most intriguing aspects of WPF/E is not yet included. You can code the CTP with JavaScript, but the promised cross-platform .NET runtime is not in this release. It is still planned though:

After the December 2006 CTP, we will also enable a managed code programming model using a subset of full CLR that will enhance the programmability side of the browsers to enable more performant and more scalable Web applications.

The quote is from the WPF/E Architecture Overview, which is a great place to start if you want to know what’s in WPF/E.

See also this interview with Forest Key from back in March, if you have not read it already.

I’ve downloaded but not tried the samples yet.

Postscript: Mike Harsh has some more info and a sample WPF/E video on his blog. The humour in the video is somehow typically Microsoft…

At the official Vista launch yesterday (UK version) Microsoft’s UK Managing Director Gordon Frazer called the launch of Vista, Office 2007 and Exchange 2007 Microsoft’s “biggest launch to business ever,” following up with further extravagances such as “a new era in business computing.”

Clearly these launches are exceedingly important to Microsoft, but I doubt they will prove the most significant in its history. Maybe that honour should go to Windows 95, which saw off the threat from OS/2, or maybe Excel 5.0 and Word 6.0 in 1994, which as I recall was the end of serious competition in spreadsheets and word processors until the Sun-sponsored Open Office in 2002. Or maybe the arrival of Microsoft .NET in late 2001, which has proved a remarkably successful answer to Java for server-side computing.

Success for Vista will be more about maintenance than breaking new ground. Success will be persuading businesses to upgrade from XP, or dissuading those with Windows fatigue from switching to the Mac. Office is the same. I mostly like the bold new user interface in Office 2007, though I’m suspicious of Microsoft’s motives, but where can you go when you already dominate the market? Staying still will be a big achievement.

That said, I am impressed with what Microsoft is doing with SharePoint, its portal technology. At the launch we were shown how this can aggregate diverse sources of information. I consider business mashups behind the firewall to be a big growth area, and SharePoint is well-placed to benefit.

What about Vista? It’s a decent product, but I’m anticipating much anguish in the first months after its launch. Two reasons: drivers, and UAC. Many drivers for Vista are not yet done, many will never be done. Users will try to upgrade and find stuff does not work, or even worse their systems will not be stable. They will blame Microsoft, and Vista will have to live down a bad reputation. Then after a year or so the drivers will be there, the OS will have had a few fixes, and the world will realise that it is actually pretty good.

The second reason is UAC (User Account Control), the new security feature which means users run with reduced privileges most of the time. UAC is a genuine step forward in security, but breaks many applications. Some will turn it off and lose the security benefit, others will suffer the compatibility issues. In a year or two maybe software vendors will have fixed their applications to play nicely with UAC.

I’ve been using mostly Vista for several weeks. It is more enjoyable to use than XP, but there are still annoyances which leave you wondering what its creators were thinking. For example, Vista Media Center is excellent – though I had to switch off Aero to stop it flashing – but why does the music library apparently forget its index from time to time, so all your albums disappear for a while?

Other things are just not done yet. I plugged in a SmartPhone and couldn’t figure out why Vista could only see it as a storage device. Answer: Windows Mobile Device Center is still in beta, and has to be downloaded separately.  

Little things perhaps, and your annoyances will be different from mine, but they spoil the overall effect. And if I were running a business network, I would leave it at least six months before rolling it out.

Overall I still think .NET Framework 3.0 is more significant than Vista or Office 2007, though it was hardly mentioned yesterday. It includes, after all, the two remaining “pillars of Longhorn”: Windows Communication Foundation and Windows Presentation Foundation. It is these that may support Microsoft’s platform through to the next generation of applications, even though they were invisible at yesterday’s event.