RE: Using middleware (RE: [GD-General] Prices of well-known 3D engines)

[Paul R. <pa...@so...>]

> Me too, and the main reason we had to ditch it completely was because we
> couldn't use just the functionality we wanted from it and do the other
bits
> ourselves. With the middleware we are starting to see now, this seems to
be
> less of a problem.

Which swings back to the original topic (sort of). The architecture of the
engine itself is a huge factor to consider for price. Our middleware is
incredibly monolithic with an "all or nothing" approach. It has super high
level abstraction to support several different platforms with super tight
coupling to it's subsystems' interfaces. Along with very little flexibility
for customization. For a PS2 only game that will never ship on another
platform, this is not good. Especially when the interfaces are most
obviously catering to a "von Neumann" machine, it's really bad news. (Don't
ask me, I wasn't working here when they bought it.)

We have a lot of tech on top of the middleware. But it's mostly hacks to get
it to do what we want it to do. This is the last project we'll use it on and
a lot of our technology will be rendered useless. Aside from my tinkering,
we don't have an in house effort to build an engine from scratch (that
government idea sounds interesting...). So now we're likely going to dive
into a new middleware solution. But this time we're battle scarred, seasoned
shoppers. So they'll have to show they're fancy demos to some other doe-eyed
n00bs. :op

Just be aware that if the engine supports multiple platforms, there's a high
probability you'll want to get to the guts and optimize for your specific
game (that's right) on each specific platform. So easy paths to
customization is a factor as well. Like Neil mentioned, any middleware
that's been around the block and used by more than a handful of teams, will
likely have these features by now.

// Paul Reynolds
// Lead Programmer

// ICQ:   102704602
// AIM:   pbreynolds
// Yahoo: reynoldspaul