Interface Draft: Breadcrumb Concept

Definition

A breadcrumb should provide a trail for users navigating within a page structure. It provides an easy way to access previously seen pages within the page structure. Users sometimes do not have in mind how to navigate to a certain page via multi level menu(s).

The breadcrumb is not intended to replace the browser back button. The browser back button is the most used element in web applications and must behave as expected.

Problem

For web applications - where items are irrevocably changed - there is no way back (except a certain undo functionality is implemented). Once an item has been changed the browser should warn. And if so, the user needs to know the difference between 'back' and 'undo'.

Overview on 'Click Levels'

The rough overview shows main functional areas and click levels (this is a diagram only, not a specification of the page flow R4).

Diagram of functional areas

Why the following breadcrumbs do not make sense

Path

Path breadcrumbs are dynamic and show the path that the user has taken to arrive at a page Wikipedia on Breadcrumbs.

• PubMan does not provide a strict hierarchical structure. In many cases there is no single path to a page.

History of pages/Action history

• This is true for static content. If items change at one point in time it does not make sense to jump to a page reflecting a state which is different from the current/real one.

Attribute

Attribute breadcrumbs give information that categorizes the current page.

• Would make sense, but as pages are role dependent the attribute might reflect a wrong categorization

Why a breadcrumb makes no sense for the following functionality

Most pages are anyway redundantly displayed in a breadcrumb or inconsistent in use, because

Simple Search

• is anyway provided persistently inside header
• is not clear where breadcrumb leads to (revise search via advanced search?)

Advanced Search

• is anyway provided persistently inside header

Organisation Search

• is anyway provided persistently inside header

Export

• will mostly appear at the end of the breadcrumb
• will automatically be available after any search

Item Details (Revisions, Attachements, ...)

• will mostly appear at the end of the breadcrumb

Submission(methods)

• is inconsistent because item changed in status

Actions

• will be a mixture of structural breadcrumb and action history
• as actions lead to modifications the breadcrumb is rather a undo history (not possible!!)

Assumption

There is a certain functional area, where items do not change and where a breadcrumb could provide help (see Diagram). PubMan should provide a breadcrumb only for item lists, view item version, item details.

This additional navigation should not be called breadcrumb or be offered in a breadcrumb manner/stype in the interface because it is different.

Examples

In most cases the labels are not so long. The trail can have up to 3 entries. Entries are cropped on the left side. Example:

Public Repository

Click Level	Label
1-2	Results (10)
3	Results (10) > View Item (escidoc:1707)
4	Results (10) > View Item (escidoc:1707) > Rleased Items (10)
5	View Item (escidoc:1707) > Rleased Items (10) > View Item (escidoc:1707:3)

Depositor Workspace

Click Level	Label
0	Pending Items (4)
3	Pending Items (4) > View Item (escidoc:1707)
4	Pending Items (4) > View Item (escidoc:1707) > Submitted Items (2)

Mixed

Click Level	Label
0	Pending Items (4)
x	Pending Items (4) > Results (10) >
x	Pending Items (4) > Results (10) > Results (99)
x	Results (10) > Results (99) > View Item (escidoc:1707)

Each entry representing a search result e.g. "Results (10)" should have a link applied to revise the search and to show the search string.

Publication Management