Improve Connection Editor!

[g2m.agent]

1. The attribute list should show the data type. int, float, string, bool, ...

2. The data types are best distinguished by color.

3. There should be a set of buttons to quickly filter attribute types. for example, you can quickly show or hide all "ai_______" Arnold attributes.

This is the picture I drew (*please ignore the layout):

[NateLikesTea]

I agree that there should be a better way to organize all that information.  Having quick-button filters would be nice.

[cattermole91]

Hello,

I have also thought about this for some time, and I like the UI mock-up, however I have some slightly different ideas.

Rather than a toggle button for specific groups, how about some "category folders"? For example all the attributes related to Arnold could be in "folder" named "Arnold Attributes". Rather than buttons, a user can easily expand /collapse the folders. Perhaps some categories are collapsed by default so only common attributes are displayed.

I'm not exactly sure what a "quick button filters" would do exactly, how would it work?

The idea of having attribute types visible and colour coded is great!

I would also recommend that the attribute name and type should be columns that can each be sorted. The standard way of changing the sorting is to click on the column header, and the sorting inverts.

Here is a basic mock-up. The "header" for each column (in the screenshot titled "Attribute Name" and "Type") can be left-clicked to change the sorting order.

--------------------------------------

The easiest feature to add, and probably the most simple for users to grasp and use is a search bar or filter text field (2 fields, one for each source and destination sides of the UI).

Using the Node Editor can help with this (because you can filter by attribute names), however realistically I find the Node Editor hides attributes that I want. The Connection Editor is a fall-back for making connections, but it's really painful to use.

Often the user knows the name of the attribute, but they need to dig though all the attributes to find it which is annoying. I've seen people (including myself), jump to the script editor to write a quick Python script to connect the attributes rather than use the Connection Editor and spend the time trying to find the attributes needed for connection.

With a filter bar, the user can type "trans" and you'll filter only "translate" attributes. For Arnold attributes, you could type "ai". The UI should assume the user wants wild-cards ("*") on either side of the text.

The search/filter should also search the attribute's "short name", the "long name" and the "nice name", with a priority.

I would also recommend changing the default sort mode to "alphabetical (A-Z)" by default.

I know the current default sort mode is based on the declared attribute initialization order, however for transform nodes. Maybe you'd need to support a "legacy" sorting mode, but I'm not sure how many users will be annoyed if the sorting changed - I personally would not be.

I hope my idea doesn't get lost in the sea of ideas, thanks for reading.

David

[NateLikesTea]

Since we can't vote on comments, just wanted to say that I agree with David.  I also resort to quick scripting to avoid using the Connection Editor as it currently is.  His layout ideas make a lot of sense (as with the OP), and a quick search filter bar would really speed up workflow, as long as you don't have to type wild cards (*) as David suggested.