Game engines consist of multiple components that combine to form a playable game; such as graphics, physics, sound, lighting etc. Instead of starting from scratch for every game, the other option is to buy and customise ‘middleware’ therefore saving a lot of time. Producing sequels is a clear way of saving time and money by re-using the same engine and just updating and improving from the previous game.
The Unreal engine is popular for next-gen development and has been used successfully to produce memorable games such as Gears of War, Bioshock and Mass Effect. Companies don’t always want to use another engine so if they choose to create their own, it may allow greater flexibility but is obviously more time consuming.
On the other end of the scale are more simple and accessible ‘point and click’ engines. Game Maker is an example that I know of. They are currently using this on the Dmu Games Programming course to teach basic coding, but engines like these are very limiting as to what can be achieved. You can, however, create games quickly and as we all know, sometimes the most simple of games can be the most fun and addicitve.
When it comes to expectations for next-gen engines, there will always have to be compromise to what can be achieved because, as said, players will never accept slower framrate over better quality.
Looking at a feature on a next-gen engine, it seems that the improvements are being implemented through shaders, speculars, normal mapping, motion blur and (body)physics. These improvements are even better complimented with the advancement of technology and the shiny prettyness of HDTV, hoorah.