I'm guessing that it's quite common and pretty hard to get around having an very large stack of SOPS during a typical modeling session. I guess my question is how tall is common? Just curious as to when I should begin revising my methods.

That's a bit like asking “how long is a piece of string?”
With the “modelling in a window” approach that's been introduced recently, one of the by-products is massive networks. As someone who was using Houdini since day one, I tend to work in both the viewport and the network - but mostly the network. It forces you to conceptualize a little more before diving in, and tends to result in more efficient/powerful networks. However, there's nothing wrong with just diving into the viewport - it works very well that way - but you will get huge, dangling networks. It was really designed to attract users that are used to that non-procedural model approach - but with proceduralism. You can always collapse them as you go to make things neater. However, if this is starting to bother you, and you want the “next step”, try working more in the network. It may seem odd at first, but it's worth sticking with.

Oh - to answer your actual question , it's not unusal to have hundreds of SOPs…it really does depend on what you're doing.

Cheers,

JC

Hey thanks for the answer John, I'll certainly head towards the network side. Sorry about my vagueness :>

Personally, when the file takes about 200MB to open I usually start a new network, sometimes sooner. Coming from max I feel more comfortable modeling in the viewport, although I did give the network approach a go when it was mandatory in 4.0. That was painful let me tell you…

I think I should clarify what I said a bit.

“Pain” is relative. If you were trying to reproduce exactly what you would do in the viewport then I agree - *that's* painful. I wasn't suggesting to Scott that was the way to go and somehow you need to “feel the burn” to be a real Houdini modeler. There's a conceptual leap you need to get over so that you aren't *thinking* like modelling in viewports, in that stream-of-conciousness way(“hey - I think I'll move *that* point now!”). That's key. I suspect you found it difficult for that reason. However, having come from the network-only background, I can attest that there's more power to be had in designing networks rather than twiddling in viewport-only. I'm not intending to sound haughty about it - it's just a fact. Personally, I use both - the modeller has gotten really good with certain things like when you need to get stuck in and model/tweak curves, attributes, etc. - the building blocks as it were. That used to be awful. Once those are there, I'll use the network for building the backbone of the net - it's far more efficient. I'll flip back and forth constantly.

Not to belabour it (too late! ), but I didn't want to come off sounding like Real Men Use Networks …I would just recommend that learning network-only *thinking* would be a great next step for someone becoming concerned about the messiness of their networks.

Cheers,

J.C.

“Real Men Use Networks” I like that, should be the official Houdini slogan. I should clarify too. It was painful because 4.0 was not a good modeling environment (at least not for polys), that's all.