It's great to hear all this talk of rendering and FX, and scalability, two things that have seen paradigm shifts and multiple evolutions several times in the last few years, so that large firms can be catered to, when we already have our own solutions, or Houdini, we use for those that are very unlikely to be matched (because even if they are, we won't own the source).
In the meantime, animation, interaction, the general feeling of an app, ease of use, clarity, the whole animation field and everything related, languish and are stuck in the realm of models, tools, procedures and concepts that last saw a generational hop with SOFTIMAGE|3D 3.0, and haven't seen one since. This leaves the large shops with a feature set we were already forced to develop quite a few years ago, and that needs to be close to the rendering engine (which is NOT going to be MRay, the one currently bundled with all products), and is absolutely useless unless you have that (rendertime injection and all), and doesn't even nudge all the static things that we actually DO use Soft or Maya for, rigging and animation. At the same time, the small boutique will still not be able to do the major scale stuff, because even with all the tools, frameworks and scalability in the world they won't have the infrastructure and deadlines to crack LA in two and sink it, and their bread&butter tools languish, and their staff still places keyframes on the same six bloody function curves they did fiften years ago, on animation rigs that feature, as their most complex technology provided by the app, IK that is the same that was made available to the first Jurassic Park. Of course I'm biased, I've always worked on a broad spectrum of things, but always around characters/creatures, so this might be a pointless and blindsided rant. But more likely, in the endless chase for bigger numbers and better headlines in reviews, the major software providers forget that as a large studio we can get ourselves a better fluid simulator for a movie in less than a year with three people than maya was able to provide in seven with a team of six. What IS phenomenally expensive to create (the user facing and enormously complex scene graph and data management side and the UI and GUI around it) is what both us and the smaller shops REALLY use day to day. And when the small shops need a multibillion particle sim they'll have three days to do, and will therefore pick up the first 400$ plugin that does it for them over an API built over a framework that scales like a Caterpillar mining truck, but can't be used without three engineers full time on it. As for ICE being the now and this massively scalable thing being the next, two years on Walking with Dinosaurs with a shitload of deformation and pipeline work relying on ICE beg to differ that large datasets are the next, and would like to raise their hand and affirm that some of us don't blow shit up all day for a living, and kinda relish a solid data model and a good UI :p
