ICY interfacing.

[Haamu]

Hello,

story (skip to question if needed):
I saw an interesting ad on cgtalk few days back, which had cleverly put phrases like "create new tools", "fast" and "codeless" together in same context...huh? I was stumped; You can't make tools without code! So after watching through all the videos I could find, completely mesmerized, I was already hooked and devicing how much time I could save with that thing.

I'm long time max user, haven't had any real interest in XSI before. I use massive amount of scripts when I work, most of which I've made myself. There's really no off the shelf solution for the countless specific things you might want to do, and I'm guessing that's exactly why ICE was created. It's really nice to see some real innovation still happening here, whereas Max development seems to be stalling quite badly.


Question:
I don't have time for extensive testing right now so it would be complete waste of 30day trial (why no free ple. like maya?), and I just wish to rid myself of these haunting questions.
As awesome as I think it is, ICE apparently does not directly create geometry, which is bit of a drawback for me.

Is it not possible to write custom scripted nodes or make ICE otherwise communicate with conventional scripts? E.g. Max's particle flow has script operator, which nicely works around all possible limitations in the system.

How about other way round, can scripts call ICE structures? Say, you have a global scope array of values, that you input in ICE and do some magic there. Then finally have ICE output a new array back into the script and continue with the usupported stuff.

Other than that I'm greatly impressed with the system. I hope it will be developed further.

[Zac-Donald]

I have no doubt that ICE will be able to create geometry in multiple ways in the next months using different work-a-rounds. You might want to wait and see what people do with it during the next month since it was just launched, before you start the trial.

[kim aldis]

ice currently can't add or create geometry and I wouldn't think there'd be anything really useful in the way of workarounds but I've no doubt at all that future releases will. Ice is simply too powerful for there not to be.

[thiago]

Other than that I'm greatly impressed with the system. I hope it will be developed further.

It is under constant development and I doubt they will ever stop to do develop ICE as a platform.
"create" geometry as a simple scatter, can be done as a particle emission that stick into the surface. Not sure if that helps you.

The way I see Softimage development now, is that in the future all the other parts of XSI will be open to this new platform. They are not gonna just add a "create object" node in there. For sure we gonna have a whole interface of object creation and more...
The way we group objects and all the management of procedurally throw objects in a scene...All the stuff with models and actions expecting certain objects... etc. All this is gonna be revisited eventually, just because of a procedural geometry creation interface.
If you see the particle interface, it's just crazy how much you can take from that and apply in deformation..
So I believe the "procedural modeling" interface, will also bring other things with it...

I see it happening like that because they don't need to think in a muscle system anymore, or a update in the shrink wrap algorithm, the famous push operator or invent a new kind of weightmap... etc.
Now it's time to think what they gonna open to the platform, and how the gonna open it.
But it always worth to wait. :)

Good luck learning XSI|ICE! It's pretty fun!