Insert as a universal modifier

[indolering]

A tester used yelp insert, should insert be a universal modifier?

It would take the ambiguity out of what happens when one hits enter.
As it stands there is a mode, some commands it inserts the preview,
others it opens a new tab, and in others it performs a transformation
of the selected text...

http://tinyurl.com/cyuf46

-Zach

[Aza]

The ambiguity of the return key was entirely a by-product of my laziness when I was sprinting to do a demo. We should standardize those actions, as well as make it easy to follow those guidelines.

-- aza | ɐzɐ --

[Gray Norton]

Hmm...it's clearly best to avoid arbitrary inconsistencies, and I think it would be great to have a set of conventions, design guidelines and best practices. But it seems to me that having an ironclad 1:1 correspondence between specific keys and specific types of actions (opening a tab/window, inserting in the the current doc, transforming the current selection) is not necessarily the answer.

Specifically, it seems most important to me that <return>/<enter> consistently mean "execute this command." And since some commands are inherently "go-somewhere" types of commands and others are "insert-something" types of commands and others are "modify-this" types of commands, that would imply different results when hitting <enter> for different commands.

Such a system does make each command responsible for setting user expectations, via an adequately descriptive name (e.g., "go to yelp" vs. just "yelp") and an informative preview, but I think that's a good thing.

For cases where the primary function of the command is to display something in the Ubiquity preview pane, or where a secondary function exists (e.g., insert the Yelp summary in the current page instead of going to the full Yelp page), then I agree that adopting and promoting some conventions (both for specific keys and for visual presentation/selection of available functions) would be very helpful.

- Gray

[mitcho (Michael 芳貴 Erlewine)]

Gray et al,

On Mon, Feb 2, 2009 at 6:06 AM, Gray Norton <gno...@gmail.com> wrote:

Specifically, it seems most important to me that <return>/<enter> consistently mean "execute this command." And since some commands are inherently "go-somewhere" types of commands and others are "insert-something" types of commands and others are "modify-this" types of commands, that would imply different results when hitting <enter> for different commands.

I agree with you here. However, I also believe it is important to think about a typology of the possible enter behavior and build specific API-hooks for each of these behaviors. This would particularly become important down the line if we get serious about stringing multiple commands together (a la http://getsatisfaction.com/mozilla/topics/piping_the_output_of_one_command_to_another_command), so that Ubiq can identify the type of the output of one command to feed into the next. If we know that a certain command opens a tab, say, and has no textual result, then it would be illogical to let the user "pipe" that output to the next command, and so Ubiquity could disallow that.

Let me know if that isn't clear... ^^

mitcho
 

On Sat, Jan 31, 2009 at 5:06 PM, Aza <a...@mozilla.com> wrote:

The ambiguity of the return key was entirely a by-product of my laziness when I was sprinting to do a demo. We should standardize those actions, as well as make it easy to follow those guidelines.

-- aza | ɐzɐ --

On Sat, Jan 31, 2009 at 2:50 PM, indolering <indol...@gmail.com> wrote:

A tester used yelp insert, should insert be a universal modifier?

It would take the ambiguity out of what happens when one hits enter.
As it stands there is a mode, some commands it inserts the preview,
others it opens a new tab, and in others it performs a transformation
of the selected text...

http://tinyurl.com/cyuf46

-Zach

 

--
mitcho (Michael 芳貴 Erlewine)
mit...@mitcho.com
Linguist, coder, and teacher