Hello Houdini users,

I would like to ask other users what quality of life improvements would like to see in future Houdini versions.

For me it implies features that would improve my day to day workflow. The more often a feature gets used the biggest the value it has. So for example I love the new kinefx but a change in kinefx might not be a strong quality of life improvement since is simply not an everyday task.

With this in mind I made a list of areas that I use the most and therefore would have a really strong quality of life value in my opinion. Thinking about the low level architecture and workflow that we deal with in every project while creating the higher level tools for FX or worldcreation.

• Network editor and nodes
  
• Wrangles
  
• Vops
  
• HDA creation and management
  
• Parameter UI creation
  
• Geometry attribute management
  

The thing is that this areas seem to have very little improvements since long ago and that is mostly the reason of writing this post. I would love to know your opinion about this topic guys.

What areas do you think they would change your day to day workflow the most? What things would you improve on those areas?

I have some ideas about what things could be improved on these areas but it would be to long for this first post.

Can't wait to read your opinions.

Multi-threaded GUI like Nuke would be the highest productivity booster IMO. So as you are working, creating/editing the network of nodes, the viewport would process the queue of operations in the background without locking the entire app.

I submitted this as an RFE but don't have the ID handy.

animatrix_
Multi-threaded GUI like Nuke would be the highest productivity booster IMO. So as you are working, creating/editing the network of nodes, the viewport would process the queue of operations in the background without locking the entire app.

I submitted this as an RFE but don't have the ID handy.

That is great example of a brilliant quality of life improvement that would impact pretty much every single time we open Houdini! I really hope they consider this one.

The only "good" thing about the whole app being frozen is that it makes it pretty clear when all network has been cook. Perhaps it would be good to add some kind of visual aid on the node to indicate when is done cooking kind of similar to tops but simpler like a single little dot or something.

Now the ideas I had for the network editor and the nodes are not as gamechangers as yours but perhaps they trigger other ideas so I'll mention them anyway.

• First idea is to be able to change the size of a node. Colors and shapes are kind of ok but being able to have some nodes bigger than others would be quite helpful in terms of reading a network.
  
  
• Then if some nodes can be bigger in the network I was thinking they could have a modern configurable widgets so that they could do thing by pressing buttons on the node itsef without even having to go to parameters panel. The simplest widget could be simply adding and image as a background for the node instead of the icon. There could be buttons and even sliders. All of this would be posible to add optionally to any node although they would be as they are now by default.
  
  
• Node info visualisations. Something like the attribute visualisations on the viewport but on the network editor and with data that makes more sense per node like the thing we can find in node info panel. It could be of different types like changing the color of the nodes or little tooltips on every node with text and numbers. For example each node having the polygon count or attributes being created or modified, groups etc... It would be like looking at the info panel but for all nodes at the same time to easy compare and read how specific data is flowing through the network without having to select every node one by one.
  

I probably haven't explained all that well so I will try to make some screenshots that better explain what I mean.

As a relatively new user - 7 months or so, I don't have the depth of experience to prioritise specific areas for attention, but I do want to add my backing for 'quality of life' improvements in the areas listed by Andreu.

I can totally understand the big-ticket technical challenges and new architectures being the attractive challenges to work on, but for the user, fixing the minor inconveniences that need to be overcome 10, 20 or 50 times a day - every day - can be even more important.

Here are a couple of simple examples:

Parameter UI: Two days ago I was making a HDA. The main network was relatively simple to create. Creating a parameter point-group drop down menu was unreasonably difficult. I shouldn't need to have to learn to script in Python to do that.

Snapping: I'm told there is no way to set preferences / defaults for the snapping tools and I use them frequently. I have to set the same eight sliders in the Multi-snap setup to my preferred setting. Every. Single. Project. That’s annoying.

and a really big one:
The Viewport needs to be 100% reliable. No professional software should ever need to have a 'Reset Viewport' menu option.

I'll stop typing before I end up with a longer post than I intended : )

On other big area I think could really do with attention: Redeveloping the UI slider set-up / interface / operation. Discussed in detail here:
https://www.sidefx.com/forum/topic/72398/ [www.sidefx.com]

animatrix_
Multi-threaded GUI like Nuke would be the highest productivity booster IMO. So as you are working, creating/editing the network of nodes, the viewport would process the queue of operations in the background without locking the entire app.

I submitted this as an RFE but don't have the ID handy.

I've read somewhere that they are reworking the viewport for vulkan, which would prevent the UI to freeze while cooking.
Is this what you mean?

The number one QoL improvement I want is for SideFX to bring in a UI designer to upgrade the look and feel of many parameter interfaces.

In my view most of them are guilty of being nothing but giant lists. The new Guide Groom is a prime example: this is not artist friendly at all since nothing is visually organized and frequently used tasks are not prioritized in the interface. In fact, probably the least used button is the very first one: "Clear."

I think some of the new Solaris nodes show the path forward. I really like the Material Linker and Light Mixer. They are visually-oriented, prioritize common tasks and are just generally much easier and pleasant to use than the typical Houdini big ol' list o' stuff.

Mike_A
and a really big one:
The Viewport needs to be 100% reliable. No professional software should ever need to have a 'Reset Viewport' menu option.

+1

Had so many times already that it feels like I messed something up... but no it was the viewport haha.

On the plus side, knowing this actually does happen sometimes, the first thing to do when some weird stuff happens is reset the viewport, and that fixes the problem :-D

Andr

animatrix_
Multi-threaded GUI like Nuke would be the highest productivity booster IMO. So as you are working, creating/editing the network of nodes, the viewport would process the queue of operations in the background without locking the entire app.

I submitted this as an RFE but don't have the ID handy.

I've read somewhere that they are reworking the viewport for vulkan, which would prevent the UI to freeze while cooking.
Is this what you mean?

I guess it could help but I don't think that change will make Houdini truly multi-threaded like Nuke. If I add a Subdivide node to a geo and it has to cook for 10 seconds, I don't want the whole app to be locked, I should be able to continue working, creating new nodes, editing existing ones, etc just like in Nuke.

animatrix_

Andr

animatrix_
Multi-threaded GUI like Nuke would be the highest productivity booster IMO. So as you are working, creating/editing the network of nodes, the viewport would process the queue of operations in the background without locking the entire app.

I submitted this as an RFE but don't have the ID handy.

I've read somewhere that they are reworking the viewport for vulkan, which would prevent the UI to freeze while cooking.
Is this what you mean?

I guess it could help but I don't think that change will make Houdini truly multi-threaded like Nuke. If I add a Subdivide node to a geo and it has to cook for 10 seconds, I don't want the whole app to be locked, I should be able to continue working, creating new nodes, editing existing ones, etc just like in Nuke.

One thing I think could be an upgrade if it ever becomes multi-threaded is if each node has an area like in tops that shows it cooking while you continue to work. Like a red light for uncooked, yellow for cooking, and green for cooked. Then maybe an option to pause the cook.

Perhaps it would be good to add some kind of visual aid on the node to indicate when is done cooking kind of similar to tops but simpler like a single little dot or something.

Nice one. That's pretty much what I was trying to say in my earlier post but you just explained it better JalexM.

Another little node network improvement I would love to have would be expandable subnetworks.

Subnetworks to have and alternative expanded mode where they look similar to a for loop with all the internal nodes accesible without having to dive inside. And with a click of a button you could toggle between the current folder view and the forlooplike expanded version.

Parameters promoted on the subnetwork would be kept correctly in the expanded version as well.

This way you could have the advantage of both things, having nodes grouped in folders but also been able to access contained nodes without having to dive in.

I was thinking about the expandable forloop like view for subnetworks and it would be as well super interesting when the subnetwork is actually a diferent context.

For example you could have a cop2net expanded in sops and see the nodes from sops and even edit or create new cop nodes from sops.

A Better Vex code window

Redo the whole Houdini U.I from scratch.

BrianHanke
In my view most of them are guilty of being nothing but giant lists. The new Guide Groom is a prime example: this is not artist friendly at all since nothing is visually organized and frequently used tasks are not prioritized in the interface. In fact, probably the least used button is the very first one: "Clear."

There's a radial menu which has most of those groom functions in that drop-down list, but for some reason they're no longer present in the hotkey editor. Switching between say Sculpt and Cut was fast and precise with hotkeys, but now one has to crapshoot around the radial menu which is neither fast nor precise.

A few simple, random, and silly things I do a lot of every day:

- object merge nodes with a hotkey. Everybody has they own script and shortcut but it's so common that a default way would be good.

- reshuffling the parms on vops and wrangles. The default Vex Setup and Attribute Bindings should not be exposed like this. It takes space and I can't remember the last time I needed them. Put them in a folder and automatically promote every parm to a dedicated folder right at the top.

- saving whole vops (or even subnets?) as presets, including the internal nodes. Not everything needs to be an hda. Sometimes you have just a lot of similar things in one shot or sequence and don't want to go through the whole hda pipeline.

- Dealing with large strings in the Geometry Spreadsheet. Asset names can be very long and parsing/checking them is tedious and common. Have them right-aligned, abbreviated, or something along those lines. Same with expressions in parms.

- ability to visually identify Houdini instances with a colored border. It's silly but I often have multiple Houdini instances open at the same time, jumping from shots to shots, or older versions, and it would help to instantly see where I'm at.

There are some UX ideas I showcased here recently (the last 3rd of the presentation)


It might help inspire some ideas though it's barely the tip of the iceberg in the whole UX design I use.

Soothsayer
A few simple, random, and silly things I do a lot of every day:

- object merge nodes with a hotkey. Everybody has they own script and shortcut but it's so common that a default way would be good.

- reshuffling the parms on vops and wrangles. The default Vex Setup and Attribute Bindings should not be exposed like this. It takes space and I can't remember the last time I needed them. Put them in a folder and automatically promote every parm to a dedicated folder right at the top.

- saving whole vops (or even subnets?) as presets, including the internal nodes. Not everything needs to be an hda. Sometimes you have just a lot of similar things in one shot or sequence and don't want to go through the whole hda pipeline.

- Dealing with large strings in the Geometry Spreadsheet. Asset names can be very long and parsing/checking them is tedious and common. Have them right-aligned, abbreviated, or something along those lines. Same with expressions in parms.

- ability to visually identify Houdini instances with a colored border. It's silly but I often have multiple Houdini instances open at the same time, jumping from shots to shots, or older versions, and it would help to instantly see where I'm at.

Great list!