> On 15 May 2018, at 06:16, Matt Lind <[email protected]> wrote: > On this I believe you are way too close to Softimage because it is not > trivial either to follow a complex scene, or a character? not saying it is > not easier (it is) but it is not trivial either. > > I disagree. A graph can be traversed and relevant nodes displayed in a view. > Softimage, Maya, Max, are all node graphs under the hood, but the data views > such as the schematic merely display subsets of the nodes which contain > certain characteristics (parent/child relationship). Houdini has parent/child > relationships too, but there isn't a convenient place where they are > displayed in isolation of other properties of the scene. Tools could be > written to traverse and display only the nodes which define a parent/child > relationship. This is, in my opinion, a low hanging fruit that could be > addressed. > Tags and network isolation is a fantastic idea… still I am convinced this will be a manual task, the traversing part is very tricky as you are in a procedural world with nodes in various context interacting.
> As for networks and subnetworks. Great, you have a system. Most people do > not, or if they do, it will not be the same system as yours. THAT is the > point. > > Same as with Passes, Partitions, Groups, Overrides and Layers in > Softimage? > we build a consensus on how to use it (everything on the BG partition > hidden > for example) and even tools to move things to the right partitions based > on > one acting as template, etc.. > Not quite the same thing. > > With passes, partitions, groups, etc.. Softimage defines the structure and > users merely label the parts in some way that is intuitive to them. > Partitions can only appear inside of passes, overrides always appear > immediately below the partition or object which it overrides, and so forth. > You are defining the structure, not how we use it which relies on a gentlemen’s agreement as it can end up in a mess, which is what I intended to highlight. > In Houdini, the networks are much more arbitrary. The user does more than > label things. They also define structure of the assets. The user can impose > self restraint and stick to a naming scheme, template for arranging elements > in the network view, etc.., but there is no consistent structure which all > users will see uniformly imposed by Houdini in the manner you see with > Softimage. This can be disorienting to the non-technical user as the data and > presentation can be radically different. > It was not my intention to compare a pass system with a scene organisation… but it is worth talking about it too as there are insights worth sharing. In Maya you tend to use groups to literally give you structure in your outliner (like dividers), in Softimage I was using empty groups to be able to separate those that did material overrides from those that managed visibility… and layers were forbidden because it was just too much to follow. That is far from great. > It is strange because it is precisely the very sophisticated HDAs system that > allows Houdini to scale teams massively while keeping complexity under > control. > > A good example; > > You're comparing apples to oranges here. The point is to get an intuitive > understanding of the data you're working with. You're talking about something > completely different. > You raised the scalability issue but probably my answer didn’t explain well enough my thoughts… allow me another try; - the solution to scalability in softimage is reference models and relaying on the traditional outliner and schematic interfaces to dive into the jungle of nodes you have now in your scene. This is very limited and we all know that. - the solution to scalability in Houdini is abstraction, how are you going to visualise the data in an asset if the logic is controlled by the asset itself? Or a packed primitive whose attributes result in geometry shaders doing stuff like creating geometry at render time? Or a node that is dynamically evaluated under a very particular set of conditions? There is no point of even trying because it is pretty much an impossible or absurdly costly task. I hope this clarifies why I don’t see this schematic view in Houdini, it would not be meaningful. Great thread though… Thanks Matt. jb > Matt > > Message: 1 Date: Mon, 14 May 2018 10:16:35 +0100 From: Jordi Bares > <[email protected]> Subject: Re: Houdini : non VFX jobs? To: "Official > Softimage Users Mailing List. > > On 14 May 2018, at 00:01, Matt Lind <[email protected]> wrote: you're > dissecting things at a more granular level than is intended, and as a result > you're losing sight of the overall discussion. > > a new user coming into Houdini doesn't have that historical background, nor > does he/she care. He only sees a lot of special case tools that require > inside knowledge to understand and use. That is the immediate point of > frustration that isn't resolved well with documentation, and in many cases, > not even discussed at all. This is one deterrent from adopting Houdini from > the generalist's perspective. > > You are right this could bring a lot of entry level comfort and easier > transition. May comment it with the guys at SideFX. > > Houdini doesn't have good tools for dealing with the macro view of a scene > for the generalist. When you open a scene you're not familiar with, or one > you haven't opened in a very long time, you want to get a general overview of > it's structure in a few seconds. That is the purpose of mentioning the > schematic view as it provides that overview at a glance. Does it tell you > everything? No, of course not, but it doesn't have to either. It does tell > you the links between nodes such as who is constrained to whom, where the > envelopes reside, which nodes have shapes/lattices/etc. and very importantly > ? hierarchical relationships to understand how rigs are put together. Again, > we're talking about the big picture. Explorer??? that?s for micro-level work > when you want the dirty details on an object. > > On this I believe you are way too close to Softimage because it is not > trivial either to follow a complex scene, or a character? not saying it is > not easier (it is) but it is not trivial either. > > It's not good for the broader picture as you have to spending a lot of time > clicking on nested node after node until you find what you're looking for, > and even then there's often a lot more information displayed than you need > leading to excessive noise. That's exactly the same problem with ICE > compounds as digging into nested compound after nested compound you begin to > lose sight over the bigger picture you're trying to grasp. This isn't a > discussion about which is more powerful, it's about presenting information > that is better suited for high level working for the non-technical user. > > Indeed this is a byproduct of a node approach, hence my personal preference > for VEX Wrangles instead of VOPs (no wire, fully defined in one single node > under the SOPs roof) > > As for networks and subnetworks. Great, you have a system. Most people do > not, or if they do, it will not be the same system as yours. THAT is the > point. > > Same as with Passes, Partitions, Groups, Overrides and Layers in Softimage? > we build a consensus on how to use it (everything on the BG partition hidden > for example) and even tools to move things to the right partitions based on > one acting as template, etc.. > > I'm not suggesting Houdini be rebuilt from the ground up. I'm highlighting > sticking points between it's current state and why more generalists don't > adopt it. When you get into a larger production pipeline, as much as you need > the low level power Houdini provides with assets and such, there is just as > much need at the opposite end of the spectrum with getting users into the > pipeline to do work. > > It is strange because it is precisely the very sophisticated HDAs system that > allows Houdini to scale teams massively while keeping complexity under > control. > > A good example; > > I am developing a character, export the asset to disk and animators start to > use it. > > They discover a problem with one control? > > I pick the asset, fix it and export the same version > > These users (let?s say rather than 1 there are 20 animators) get the asset > WHILE THEY ARE WORKING, without interruption. > > No scripts, not nothing.. bang. > > Imagine the change is enormous, just add a version and they can choose the > version they want to use? again, all dynamically. > > Now scale this to everything is an asset where the city buildings are all > being modelled live, the cars rigged, the characters updated? and you have to > do NOTHING to get the latest and greatest version. > > And now go further assets contain assets that contain assets, all versioned > based where. > > City v1.0 contains BuildingA v1, BuildingB v1 and BuildingC v1 > > City v2.0 contains BuildingA v2 and , BuildingB v1 and BuildingC v1 > > And those buildings indeed contain the windows as assets, the doors, the roof > furniture? all versioned of course > > You get it? no pipeline required, no scripts, no nothing. > > Very very quickly you can see that may be, just may be, having the best > f-curve editor is not even important in the big scheme of things. > > cheers Jb > > ------ Softimage Mailing List. To unsubscribe, send a mail to > [email protected] with “unsubscribe” in the subject, > and reply to confirm. > >
------ Softimage Mailing List. To unsubscribe, send a mail to [email protected] with "unsubscribe" in the subject, and reply to confirm.
