Personal tools

Labs/Test Pilot/Test Proposals/Tabbed Browsing

From MozillaWiki

Jump to: navigation, search

Contents

Tab Open/Close Study

  • Champion: Limi (limi@mozilla.com)
  • Status: Draft
  • Duration: 1 week
  • Implementation: Not started yet. Will be implemented with Test Pilot extension.
  • Data to collect:

Introduction

Tabbed browsing is now the standard interface offered by all major web browsers, and its been argued that the browsing experience becomes difficult to manage when more than a handful of tabs are opened.

In conjunction with our recent focus on tabs and their possible reinvention (see the Summer 2009 Design Challenge: Reinventing Tabs, as well as in numerous blog posts) this study will seek to provide designers and developers additional information and insights to help inform the design of future tabbed interfaces.

Goals

This will be the very first study of tabbed browsing through Test Pilot. Our goal for this study is to understand and design:

  • Tab behaviors after people close a tab (approved)
  • Ways to manage and navigate among a large number of tabs
  • What the distribution of tabs are, wrt. how many people have open, max/min.
  • How many people open new windows to group tabs
    • How widely spread are sessions? Ex. Google search and new results in tabs
    • How long do split sessions survive?
  • Behavior around tab dragging and docking (approved)
  • Tab selection errors
    • T2(tab select) - T1(tab select) < 1.5s

Instrumentation

  • Tab open/close hooks are easy
  • Tracking "open in new window"(OiNW) and subsequent metrics is not easy.
    • Domain overlap
    • Lifetime variation in OiNW tabs versus other tabs (expectation is shorter if these are truly speculative opens, preserving the original)


Details

Design question 1: What would be the best default tab behavior after users close a tab?

In the past, we have observed that people close a tab in different contexts, for example:

  1. When people finish using the tab: e.g. users close the tab when finish watching a video.
  2. When people organize browser tabs: e.g. users close some less important or less used tabs when there are too many tabs open in a browser.
  3. When people temporarily hide a tab: e. g. users close a tab when he detects someone is passing by; he would like to open it again when nobody is watching.

For different contexts, people have different intentions after close a tab, for example:

  1. open another interesting tab
  2. close another less important or less used tab
  3. open whatever tab
  4. open the most often used tab

For this test, we would like to to explore what would be the best default browser behavior after users close a tab. We will simply log what users would do after he/she close a tab, and through these log data, we may be able to detect certain patterns that help us design the default browser behavior that meaningful to most users. 

Annotations:

  • T1: the tab user just closed it
  • T2: the tab shows automatically after user closed T1
  • T3: the tab user selects after T2


Metrics to consider:

  • User actions after close T1:  T3 - T2=?
  • Relations between tabs:  relations between T1&T2? relations between T1&T3?  relations between T2&T3? (e.g. opened through the same tab? within the same domain? opened at the same time? no relation at all?)
  • Notable charactor for T1,T2 and T3: e.g. most often used? open the longest? previously opened?


Generic parameters to log:

  • How many tabs are open in total when users close T1?
  • Any add-ons or settings may influence this user behavior?

         * so far 273 add-on is related to tab behaviors based on AMO.

  • Considering both mouse action and keyboard action

         * close the current tab: command+w (Mac); contrl+w (Win)

Design question 2:  What should be the best default placment when users open a new tab?

We have observed that people open a tab for different reasons, for example:

  1. Open a new blank tab to type URL or to perform search
  2. Open a new tab through links within current tab

We also notice that people doing following actions after they open a new tab:

  1. drag the newly opened tab next to the previous tab
  2. drag the newly opened tab next to a particular tab
  3. drag the newly opened tab to a new window


For this test, we would like to explore what we can do to support people open a new tab in a meaningful way.


Annotations:

Tc: the opening tab before user open Tn (could be Tl or Tr)

Tn: the newly opened tab

Tl: the tab left to Tn after Tn is dragged

Tr: the tab right to Tn after Tn is dragged


Metrics to consider:

  • Type of Tn:  with content (opened through a link), or blank?
  • User actions after opens Tn when it is blank:  how users load content for the newly open blank tab? e.g. type URL, perform search, through other means? After the loading the content, do user perform drag and drop to group Tn with other tabs?
  • User actions after open Tn when it is with content: activate Tn immediately? Time between open Tn and activate Tn? do user group Tn with other tabs?
  • Relations between tabs: relations between Tc and Tn? relations between Tn and Tl/Tr?

Generic parameters to log:

  • How many tabs are open when users open and drag Tn?
  • Any add-ons or settings may influence this user behavior?

         * so far 273 add-on is related to tab behaviors based on AMO.

  • Considering both mouse action and keyboard action

         * open a new blank tab: command+t (Mac); control+t (Win); right click on a tab, through menu button

         * open a new tab through a link: command+click (Mac); control+click (Win); right click, setting in Firefox


Basic info for this test:

  • How long the browser is opening for?
  • Tab/window amount per person during the browser opening
  • How many times Firefox becomes unfocused during the time opening?
  • Permanent tab amounts open during the broswer opening.