Here's a talk by deepmind about this AI https://www.youtube.com/watch?...

A signed integer overflow for timing - scary...

There is a couple of open source projects doing this already: http://csnative.codeplex.com/ https://github.com/xen2/SharpL...

One of its unattractive features to programmers who are probably more familiar with c#/java is its verbosity. Having to declare ever function twice in an interface and implementation section gets tedious and seems unnecessary given Delphi has provided RTTI for years. At least with c++ you could in theory code everything in a header file - although I've heard that is sure fire way to make a summoning grid and summon an elder one.

Still miss delphi sets, they were so much more powerful than c# flag attributed enums.

It could be the mortality rate for Ebola has been overestimated in any case, mild cases may not be recorded at all - there is something of a stigma with Ebola. The most recent case with the Nigerian diplomat who skipped quarantine and holed up in a hotel and then recovered after a few days indicates there are some who get mild symptoms (and are still infectious however - see Port Harcourt outbreak).

This article http://cosmiclog.msnbc.msn.com/_news/2012/03/23/10831821-mysterious-cloud-spotted-on-mars says that it has 'since' been found in images going back to the 12th.

The problem with all of these cross platform frameworks, they work great if your writing a game or a very simple app that consists of a single screen. However as soon as you go multiple screen you find that the UI metaphors each OS has fight each other, namely Android use of a hardware back button and it use of a hidden menu, these just don't gel with iOS way of doing things. You end up having to write two apps within one app anyway, and in javascript...

Anti-cheat efforts mirror DRM efforts in that in an open system that can run arbitrary code you cannot make it work. Its pointless. Theres only two ways you will get a reasonable working anti-cheat system and that is to raise the bar of access to run arbitrary code, i.e. a closed system such as a console. OnLive's video delivery method may offer some hope with regards reducing cheating on the PC, but you still allow some pixel scanning cheats, but such cheats are very limited in what they can achieve. I hope that OnLive realise that one of the major advantages that they have over traditional disc/download games is that they will be relatively cheat free, and they segregate their multiplayer servers from non OnLive players.

Although it'll be nice to have a proper centralised player matching service on the pc, much like live on the 360. The biggest concern for me is the mandatory 360 controller compatiblity to get games for windows certification. For RTS games this is going to result in horid UI's which have to work with both a mouse and a controller. Also as the games will most likely be cross platform with the 360, the controller will take precedance in the UI design. So games for windows will most likely mean that all the pc gets in future is 360 ports that require a controller plugged into the pc to play decently.

Ultimately while this seems like its a bold new push for windows games, in reality I can see this reduce the distinctiveness of PC games against 360 games (the control method) and hence push more people onto the cheaper wholely owned microsoft platform ... the 360.