I've thought of a specific function option for the tool shelf but can't determine whether it is a good idea or not. Thus this discussion.
I'll provide a scenario:
I'm going to do some modelling and I'm going to start with a sphere (just for arguements sake).
I call the sphere in from the tool shelf. I'll use the Ctrl key to place it at world center. I turn it into a polygon sphere.
I want to polyextrude some of the polygons.
I am still at Object level, and with H9's bias towards “selection-action”, if I wish to pre-select some polygons I have to know that I can't do that at Object level. I need to switch to sop level.
If however, I chose the Polyextrude from the shelf, and if the sphere is pre-selected, the tool sends me to sop space where I the need to select the polygons I wish to work on, then hit Enter to initiate the tool.
If the sphere is not selected at Object level, I need to click the sphere once to send me to sop space and then select the polygons that I wish to work on.
Each of these cases is an “action-selection” process.
So here's the question:
At Object level, should there be a hot key combination( Ctrl-Alt?), or some other option, so that when I call in a primitive I'm automatically sent to that objects sop space?
What I'm thinking is that this would better maintain a “selection-action” bias, since I would be able to pre-select polygons and then call the polyextrude.
I want to be clear here, the question isn't an invitation to complain about H9 workflow.
I'd just like to know if the option I've described would be a useful addition.
Full Version: Shelf tool functionality discussion,
In HV8 there was really just one way to get at various pieces of the extremely organized network and node based approach, which was to jump into that network. Of course it's very easy to conceptualize as well since its how folder structures and hierarchies etc work. The only functions that could be done would be the ones that are allowed in that context.
The only inefficient and hard part with that approach is un-learning the other ways.
So, personally I'd prefer to keep a very razor-sharp distinction between the tools and operations which are performed at the object level, and those which are really operating or wil take the user to the subobject level, or within whatever sub-context might exist.
If I need to see the end result of how the operation I'm doing is working I'll just observe in object level while adjusting a parameter in a sub-context.
So, to sum up that UI babble, the most simple approach is usually the best approach, and even though it might be fewer clicks to do something from an object context instead of a subcontext, the concept is actually more straitforeward.
The only inefficient and hard part with that approach is un-learning the other ways.
So, personally I'd prefer to keep a very razor-sharp distinction between the tools and operations which are performed at the object level, and those which are really operating or wil take the user to the subobject level, or within whatever sub-context might exist.
If I need to see the end result of how the operation I'm doing is working I'll just observe in object level while adjusting a parameter in a sub-context.
So, to sum up that UI babble, the most simple approach is usually the best approach, and even though it might be fewer clicks to do something from an object context instead of a subcontext, the concept is actually more straitforeward.
I'm with Andrew - simpler is better, and for me simpler is understanding exactly what's going on(objects containing SOPs) and setup the GUI to let me get at what I need the easiest possible way. It's for that reason I find doing “everything in the scene pane” to be rather complicated and convoluted. I find it akin to looking at a city through a telescope rather than just stepping back and analyzing the structure. For a long time I've always had a desktop with shows objects with their params in the same network('p'), and SOPs of the selected object in their own network with params displayed at the same time. I don't so much “dive in and out” of SOPs and OBJs, I simply click where I need to do something. I seriously don't understand why everyone doesn't work more or less this way(once they've grokked Houdini, of course). It's for that reason I tend to use the shelf for things like fast prototyping of DOPs, adding the odd quickie object or light, and custom things like switching desktops. And I love the shelf, too. 
Again, I agree, not to argue about H9's “workflow”, just what works faster for me.
Cheers,
J.C.
Again, I agree, not to argue about H9's “workflow”, just what works faster for me.
Cheers,
J.C.
At Object level, should there be a hot key combination( Ctrl-Alt?), or some other option, so that when I call in a primitive I'm automatically sent to that objects sop space?
“Create In Context” if enabled, does exactly that. So, my answer is no.
Shelf.
I think (may be i wrong ), shelf idea, come from “M” (SESI sorry please).
I never use shelves in “M”. Just in the first steps when start learning. Then I delete all shelves. For shelf-working i must do too many “clicks” and mouse moves.
1. I understand what tool i need.
2. I move mouse and select from shelf menus shelf to needed.
3. i search the icon of the procedure\tool and find it.
4. I click it (of course without any options)
5. I see result and result may* be* success.
When i work with shelf,i can`t understand the process, the network and other.
But. “M” have other* pipeline philosophy.
Why Houdini need “M” way? What next step? Tonnes of slow scripts without optimization, like thousands “M” scripts?
I learn Houdini short time, but i see excellent speed. Super speed. In H8. In H9 i see “running python script” again and again. I see shelf. (about GUI-s - i wait new graphics cards.)
I think, the power Houdini in fast,detailed understanding and working with geometry data from point level to HI-poly animated models, without ideology “one click results”.
Shelf is need, may be, only for starting learn and not more.
When the “one click” men go inside, he see the “real face of package” again
In “M”, for first* steps only, and some custom tools putting. Not more.
In Houdini i disable all shelves, when i work with shelves, i don`t understand anything what i do.
Sad. When i want take the detailed help about new H9 tools, the “help” show me the Shelf using. And i go to shelf and search and press and see the “running python scripts” and see result and go inside and See The Real Face Of Houdini Without detailed help
H8. Rise
May be i need time for relearning.
I think (may be i wrong ), shelf idea, come from “M” (SESI sorry please).
I never use shelves in “M”. Just in the first steps when start learning. Then I delete all shelves. For shelf-working i must do too many “clicks” and mouse moves.
1. I understand what tool i need.
2. I move mouse and select from shelf menus shelf to needed.
3. i search the icon of the procedure\tool and find it.
4. I click it (of course without any options)
5. I see result and result may* be* success.
When i work with shelf,i can`t understand the process, the network and other.
But. “M” have other* pipeline philosophy.
Why Houdini need “M” way? What next step? Tonnes of slow scripts without optimization, like thousands “M” scripts?
I learn Houdini short time, but i see excellent speed. Super speed. In H8. In H9 i see “running python script” again and again. I see shelf. (about GUI-s - i wait new graphics cards.)
I think, the power Houdini in fast,detailed understanding and working with geometry data from point level to HI-poly animated models, without ideology “one click results”.
Shelf is need, may be, only for starting learn and not more.
When the “one click” men go inside, he see the “real face of package” again
In “M”, for first* steps only, and some custom tools putting. Not more.
In Houdini i disable all shelves, when i work with shelves, i don`t understand anything what i do.
Sad. When i want take the detailed help about new H9 tools, the “help” show me the Shelf using. And i go to shelf and search and press and see the “running python scripts” and see result and go inside and See The Real Face Of Houdini
Without detailed help H8. Rise
May be i need time for relearning.
When i see words “running python scripts” - i afraid. Really. I remember MEL. Many many many MEL scripts without optimizations. I really afraid..
SESI. please. Don`t turn Houdini to “M”. It wrong way.
I go to search other work, if this happen, really.
Fast results without brain work is a “bad black magic”.
Enemies not sleeping, but wait and smile.
Peoples without mind attack the earth.
~ ~ ~
H8 rise!
SESI. please. Don`t turn Houdini to “M”. It wrong way.
I go to search other work, if this happen, really.
Fast results without brain work is a “bad black magic”.
Enemies not sleeping, but wait and smile.
Peoples without mind attack the earth.
~ ~ ~
H8 rise!
probbinsMaybe I miss something but being at the Object level you can press “4” (or choose “Primitives” from the selection mode on the left toolbar), select your polygons and choose PolyExtrude from the shelf. This actually places you at the SOP level but you stay in the viewport while this happens.
I am still at Object level, and with H9's bias towards “selection-action”, if I wish to pre-select some polygons I have to know that I can't do that at Object level. I need to switch to sop level.
Dragos
digitallysaneprobbinsMaybe I miss something but being at the Object level you can press “4” (or choose “Primitives” from the selection mode on the left toolbar), select your polygons and choose PolyExtrude from the shelf. This actually places you at the SOP level but you stay in the viewport while this happens.
I am still at Object level, and with H9's bias towards “selection-action”, if I wish to pre-select some polygons I have to know that I can't do that at Object level. I need to switch to sop level.
Dragos
No you're not missing anything. The point here is that you have to know about this step ahead of time. You also have to be aware of the distinction between objects and sops.
live_fx
When i see words “running python scripts” - i afraid. Really. I remember MEL. Many many many MEL scripts without optimizations. I really afraid..
H8 rise!
There are actually two different types of scripts (or are there more?) which I never quite understood until I started using Houdini. I used maxscript, not sure how mel is but I think it's basically the same in this regard.
Basically, there are the type of scripts “interface level scripts” I would call it that anyway, where you are making Python order the interface around just as if you were to click on those buttons. Variables can be used of course. Mainly objects can be created copied, renamed, keyframes can be set etc.
The “fast” scripts would be vex/vops/expressions/pythonexpressions. Where the computer is using pure math instead of referencing some other part of the program. Those would be typed within a certain part of a “native” operator which would just cook until it figured out whatever value it needed.
So, I think with python it would be more how python was used vs how fast and slow it is. I think python is a really great thing for Houdini since it is object oriented and very robust scripting language etc. For the super performance stuff I'm guessing VOPs would be the way to go? Seemed to be really fast with particles anyway.
probbins
No you're not missing anything. The point here is that you have to know about this step ahead of time. You also have to be aware of the distinction between objects and sops.
But aren't the other 3D apps the same in this regard? In Maya you have to switch to component mode first before you extrude the faces. The only real difference is that Houdini lets you have a better interface to component mode (SOPs).
yeah thats the super hard thing to teach students with max is the difference between component/subobject/subcontext mode, and object level.
It only becomes clear when things mess up and then ok time to learn transforms and vector math etc. In Houdini it's nice colors and things go to their native transform, instantly understandable I think.
It only becomes clear when things mess up and then ok time to learn transforms and vector math etc. In Houdini it's nice colors and things go to their native transform, instantly understandable I think.
andrewlowell:
i`m about “running python scripts” on fluid operations and some others called from shelf and mantra in some cases.
First alert in mind “wow. scripts in fluid solvers? and render ?”. Of course, i sure, this will be optimized in future I just afraid
I`m a maya user. Maya from release to release have changes in shelves (new primitives, for example), new UI-scripts, new panels “web” buttons, but “instances” is simple and stupid, as in earlier versions. Sprites too.
- - -
Shelves ? Ok. Python? Ok. New GUI ? Ok.
But when i can see in Houdini, for example, SOP color ramps* and splines*? This excellent implemented in Maya and used in shading, animations, particles operations.
I understand this is a hard code parts, but it really* needed.
I don`t wait the shelves and new beauty gui-s, scripts. I wait other features.
i`m about “running python scripts” on fluid operations and some others called from shelf and mantra in some cases.
First alert in mind “wow. scripts in fluid solvers? and render ?”. Of course, i sure, this will be optimized in future
I just afraid I`m a maya user. Maya from release to release have changes in shelves (new primitives, for example), new UI-scripts, new panels “web” buttons, but “instances” is simple and stupid, as in earlier versions. Sprites too.
- - -
Shelves ? Ok. Python? Ok. New GUI ? Ok.
But when i can see in Houdini, for example, SOP color ramps* and splines*? This excellent implemented in Maya and used in shading, animations, particles operations.
I understand this is a hard code parts, but it really* needed.
I don`t wait the shelves and new beauty gui-s, scripts. I wait other features.
yeah I want a Python CHOP definition, I just see Python as actually adding new possibilities just because it's a robust and OO language
live_fx
First alert in mind “wow. scripts in fluid solvers? and render ?”. Of course, i sure, this will be optimized in future
All the Fluid code along with the rest of DOPs is all coded in C++. That “Execute Python Scripts” that you see when you use the shelf is nothing more than Houdini wiring up nodes for you and setting some parameters. You can view the shelf as being a set of sophisticated presets. If you ignore the Shelf you can still setup networks that produce the same results.
“If you ignore the Shelf you can still setup networks that produce the same results.”
Yes, and i use it for learning H9 fluids and other.
And so
“Shelves”
“it is a good idea or not.”
For learning steps, for understanding new features, with massive help docs support - very good idea ( but help docs show me creation process through shelf - after H8 docs explanations, this is inconveniently)
For work - may be (i never use shelves in Maya, just custom scripts with hot keys).
In all cases, if someone don`t like use the shelf - close it, if like - use it. Only problem with workflow through shelf, explained in new docs. Not more.
Yes, and i use it for learning H9 fluids and other.
And so
“Shelves”
“it is a good idea or not.”
For learning steps, for understanding new features, with massive help docs support - very good idea ( but help docs show me creation process through shelf - after H8 docs explanations, this is inconveniently)
For work - may be (i never use shelves in Maya, just custom scripts with hot keys).
In all cases, if someone don`t like use the shelf - close it, if like - use it. Only problem with workflow through shelf, explained in new docs. Not more.
live_fx
In all cases, if someone don`t like use the shelf - close it, if like - use it. Only problem with workflow through shelf, explained in new docs. Not more.
A very good point. While documentation showing off the shelf is great, traditional methods should be taught too.
ah ha, thats the issue being discussed here, that sums it up I think.
The problem with user-friendly methods comes about when the easy and less abstracted/flexible approach becomes the approach most people use, this area gets more support and therefore even more use, and eventually becomes the standard way of doing things and the more flexible/powerful workflow is lost and sometimes doesn't play as well with newer features. I see remenance of this with 3ds everywhere. There are even features built around the shortcomings of other features!
On a lot of older max features like volume select / xform / mesher etc. A lot of max users don't even know what these are for.
Houdini has very good, current, and bug free “old school” features which make it powerful and these features live in a very non-presetable enviornment. Could there be a better way to get in and do a point SOP for instance.
To my knowledge there still isn't a good doc on how to bounce point attributes around from different contexts, took me a while to really adopt this mentality coming from other mentalities etc where attributes don't really exist, I had to figure it out with the help of forum questions etc.
How bout a build and improve your favorite Maya Deformer / Max Modifier with a SOP DA competition to show that even if it takes more time, the features are more flexible, here's mine.
http://www.andrew-lowell-productions.com/andrew-lowell-productions/download/temp_skinMorphAssetVidTutorial.mov [andrew-lowell-productions.com]
I do like shelves in the sense that I can add my own stuff to them, It would be cool to have a lot of preset compositing templates in COPs that do standard video post effects like glows, color correction, keying etc.
The problem with user-friendly methods comes about when the easy and less abstracted/flexible approach becomes the approach most people use, this area gets more support and therefore even more use, and eventually becomes the standard way of doing things and the more flexible/powerful workflow is lost and sometimes doesn't play as well with newer features. I see remenance of this with 3ds everywhere. There are even features built around the shortcomings of other features!
On a lot of older max features like volume select / xform / mesher etc. A lot of max users don't even know what these are for.
Houdini has very good, current, and bug free “old school” features which make it powerful and these features live in a very non-presetable enviornment. Could there be a better way to get in and do a point SOP for instance.
To my knowledge there still isn't a good doc on how to bounce point attributes around from different contexts, took me a while to really adopt this mentality coming from other mentalities etc where attributes don't really exist, I had to figure it out with the help of forum questions etc.
How bout a build and improve your favorite Maya Deformer / Max Modifier with a SOP DA competition to show that even if it takes more time, the features are more flexible, here's mine.
http://www.andrew-lowell-productions.com/andrew-lowell-productions/download/temp_skinMorphAssetVidTutorial.mov [andrew-lowell-productions.com]
I do like shelves in the sense that I can add my own stuff to them, It would be cool to have a lot of preset compositing templates in COPs that do standard video post effects like glows, color correction, keying etc.
andrewlowell
How bout a build and improve your favorite Maya Deformer / Max Modifier with a SOP DA competition to show that even if it takes more time, the features are more flexible, here's mine.
http://www.andrew-lowell-productions.com/andrew-lowell-productions/download/temp_skinMorphAssetVidTutorial.mov [andrew-lowell-productions.com]
You should show this on CGchannel.com, Andrew!
cheers,
sy.
I must say I am new to Houdini and the shelf is simply giving me the oportunity to learn fast the relationships between all the modules, of course i can see already in a near future I will either creat my own shelf tools or hide it altogether but for me has been extremely good to learn Houdini.
I believe SideFX has done the right moves in an incredibly clever way as this plus documentation/tutorials and the huge effort in training is already paying (at least for us, we had 1 licence of 8 for more than 2 years, now have started using it and have 2 plus 4 batch renderes)
Thanks
I believe SideFX has done the right moves in an incredibly clever way as this plus documentation/tutorials and the huge effort in training is already paying (at least for us, we had 1 licence of 8 for more than 2 years, now have started using it and have 2 plus 4 batch renderes)
Thanks
probbins
At Object level, should there be a hot key combination( Ctrl-Alt?), or some other option, so that when I call in a primitive I'm automatically sent to that objects sop space?
There are actually two hotkey options. You can either press 2, 3, or 4 to go to the geometry level and begin selecting respectively points, edges, or faces. Another option is to press F8 which toggles between object and geometry levels. Your last geometry selection mask will be there when you toggle with F8.
This hotkey is designed to help “M” users transition because it is the hotkey for toggling between object/component levels in that app. The 2, 3, 4 hotkeys offer more choices.
Houdini 9 was designed so that if you make a selection first (go to sop level and select faces) then polyextrude then the extrude will happen right away. If you have an object selected at the object level and click polyextrude first then Houdini puts you into sop mode and prompts you for a selection. M just polyextrudes every face on the object but we thought that that was a little presumptuous. Therefore we revert to an action/selection method which then requires that you confirm the selection.
Robert