Different UX for multi-edit vs. single record edit

Question

I have a user datagrid with inline editing capability. In some cases there is additional information that appears in a bottom pane (permissions details etc). The user also has the ability to bulk edit. As an example, they may want to change permissions across a number of users. I feel this needs to be in a dialog so the user can make changes and only save once they are sure of the changes. This creates, on occasion, two different experiences for the same information depending on whether the use edited one. Has anyone experienced this dilemma?

Edit: I cannot give show an example, as this is an internal project, but perhaps I can clarify.

To edit a single user, there are two actions:

1. Edit fields inline on the datagrid
2. Edit permissions in the panel that slides up from the bottom of the screen

To edit multiple users at once the app user must select multiple rows, they can then:

1. Edit fields in a form (these are the same fields that are in the datagrid)
2. Edit permissions from a form (these are the same permissions that you would edit for a single user but applied across all selected users)

I am proposing to put all multi-edit actions in a dialog as the need to be carefully considered and activley saved, whereas single user edits are auto saved.

Here is a rough sketch:

Option A represents what I am planning for single user edit.

Option B is a modal for multi-suer edit.

Answer 1

The convention for this type of functionality is achieved using property editors, although nowadays many people will be more accustomed with the term inspector.

These are available on a multitude of software products like word processors, vector graphics editors and other graphical editing software like XMind (the property editor is on the bottom right, which affects single and multiple selection):

A similar concept can be seen in Balsamiq, which is integrated into this forum:

Property editors allow the editing of all fields that are common to the selected object/records.

Notice that a (non-modal) panel editor is preferred over (modal) dialogs, as dialogs often hide the selected object/records, which could be important for the user (also sometimes as part of a feedback loop). But the complexity with non-modal editors is that changes has to be applied immediately (no cancel/save), which could be a good 2-week headache if the information is to be persisted to a server.

I recommend you adopt this strategy - a panel editor to edit all selected records, whether a single one, or multiple ones. Such consistency may come with editing-speed penalty, but it seems odd that a single edit and a multiple edit get different implementations. You can always keep in-line editors for speed if you wish (possibly as part of accidental discovery).

Answer 2

After some time trying out multiple options. I settled on the simple idea of inserting a new row after the last selected row with the form in that row. Example below.

^{download bmml source – Wireframes created with Balsamiq Mockups}