Get people used to the ubiquity interface regardless whether they begin the process with a mouse or keyboard. This maintains consistency and you only have one interface to receive feedback and work on.

So for mouse navigation, this is how I’d see it working : when you select text, the selected text flies (think maximizing window like effect) and becomes part of the normal ubiquity interface (just like you had made the selection and pressed the activation keys).

The difference is that now, a user only has to move the cursor over the action (say “email”) and rotate the scroll wheel for it to change to other alternatives. Maybe have an animation here too…like an old petrol pump…or something that makes it obvious that it is the motion of the scroll wheel that’s causing the change.

In addition, since it is mouse only, words like “to” will have to appear automatically (for example, in “email to”.

Finally, using the same example of emailing text to a contact, the user will be able to use the scroll wheel to move over the alternatives. If there are a lot of alternatives, like in email contacts, group them (there are several possibilities, but one obvious one would be by first inital…an index tab of sorts above the current contact.

This lets everyone get used to and improve on one interface, which is a good thing from the start. This will also be good from a touch screen point of view - using ubiquity on a smartphone or internet tablet, for example.

From my experience, people who use a mouse a lot, like to know what effect their clicks and scrolls are having. This is why it is important to show them whats going on.

How about launching the screen you see when pressing ctrl+space if you click on the icon next to your selection?

How about a menu similar to the one on DeviantArt? It’s “deep”, functional and pretty easy to use.

Regards,

Roberto

Copy (with formatting, as text)
Search (in background tab, foreground tab)
Open as link (in background tab, foreground tab)
Translate

And since highlighting normally performed “using the mouse”, then I don’t want to switch to the keyboard to initiate any of those commands tens of times daily.

All current extensions that provide similar mouse-interface to the one you suggested lack the ability of either choosing what to appear in the list or the ability to create personal commands.

Ubiquity can do this much better by giving the user deep customization and the ability to disable the interface he/she doesn’t like. Some people might even add a command at the end of the pop-up list that opens the command-line interface and disable accessing it through the keyboard in use the key combination for something else.

Also, it would be more productive if the user is allowed to set some rules that are automatically performed. For example, if the selection is a regular text autocopy, if it looks like a link in a text format then open in new background tab, if it is a single word then translate into Xyz language, and if it is non-English text then translate to English.
In addition, some people don’t like having an icon appearing when selecting text. This can be dealt with through a gesture that tells Ubiquity when to engage (Double click on any empty space -> Ubiquity Activated -> make the selection -> Ubiquity shows the commands list).

Hope to see this happening in the near future.

Wouldn’t a mouse driven UI could be much closer to Automator, or Yahoo pipes or even Lego bricks?

Simple scenarios like maps and Twitter are fine with menus but longer term vision is to move complex data (not just text selections) between sites that don’t know about each other.

Some more complex scenarios to consider:
Flickr Photo Set -> Ebay Posting -> Twitter when done
Craigslist search -> Google Map Results -> Send as Email
For each stock in my Etrade account -> Search CNN

These types of things won’t be possible with a traditional menu based approach.

A possible idea is to replace the context menu entirely. They’re all actions. I understand this may be a bit annoying for some people but I think that can be mitigated by pinning the common actions at the top of the list, and allowing people to drag/sort the items.

I’ve been wanting all the firefox actions to be made part of ubiquity since first trying it. Like opening dialogs(history,options,addons), bookmarking pages,print, etc.
So this may be a good way to work on both.

However, having a mouse-based interface for Ubiquity is not a bad idea as some users are not so used to a common-line based interface especially average users and beginners.

Although this makes for a large parameter space for the user to navigate, the radial nesting of the menus means that for a particular action the user quickly learns, via motor memory, the correct sequence of mouse motions. It essentially becomes like mouse gestures (you can quickly activate a feature by drawing out a particular shape), but is discoverable because the radial menu shows the categories/options at each step (if you are navigating slowly…).

I would think that having the user drag the Ubiquity arrow through a series of radial menus might work. Downsides are that perhaps not all users are sufficiently dexterous for mouse gestures to feel natural, it’s sufficiently different that it might confuse some users, and it may interfere with mouse-gesture add-ons people already use.