; An Evaluation of the Tabbed Interface in Web Browsers
Learning Center
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

An Evaluation of the Tabbed Interface in Web Browsers


  • pg 1
									       An Evaluation of the Tabbed Interface in Web Browsers

Tabs first introduced in HyperTIES, a browser for one of Ben Shneiderman’s projects for an
interactive encyclopedia system. It was later popularized by Opera in 2000 before being adapted by
other popular like Firefox and Internet Explorer. The paper attempts to understand the reasons for
the popularity of the tabbed interface and if it still holds strong as the best means of browsing across
multiple websites. The paper will underline some of the disadvantages of tabbed browsing given that
browsers have now become the most widely used computer application and are now used for a wide
variety of web applications. If there are drawbacks then are there any alternate solutions? Are there
any current research work going on in this area? If no then what can be done about it? These are
some the questions the paper will attempt to answer.


The tabbed document interface (TDI) allows multiple documents to be opened within a single
window. It was brought about as an alternative to multiple document interfaces (MDI) which would
open a new window for every document causing a clutter. Since the scope of this paper is web
browsers, I will talk about TDI and MDI completely within this context.

I always thought that Opera was the first ever browser to implement tabs in a web browser. But
after some research I have come to know that it was in fact first implemented in Ben
Schneiderman’s HyperTIES Browser [1]. The paper on HyperTIEs reveals a detailed study carried
out on windows selection strategies. The use of multiple tabs, sorting tabs based on usage and time,
giving control of tab placement to users and other strategies have already been tried and tested in
this system which has been in the works since 1983. It is amazing that it took more than a decade
for it be fully used to popular web browsers. Opera successfully popularized the tab implementation
in 2000. Firefox and Internet Explorer followed suit and now all versions of all popular web
browsers use a TDI implementation for browsing across different pages.

Success of Tabbed Browsing:

The fact the TDI has comes default with all browsers is a proof of its success. People don’t want
multiple windows open in the taskbar and also don’t like to see them cluttered under a single panel
forcing the user to click on the panel, waiting for the list of windows to show up and then selecting
the window. Tabs allows the user to view all the documents/web pages open, read the title of the
web page and easily click on the tab using the mouse or through the keyboard navigation using
specific hot keys.

Implementing multiple tabs under a single window allows the operating system to allocate a process
for each tab or have a single process for all tabs. This helps in reducing the CPU usage and the
memory load.
Problems with Tabbed Browsing:

Some of the problems with tabbed browsing are old, well-known ever since it was implemented and
some problems are new with more and more sophisticated usage of web browsers. I will try to list
most of these problems in the next few paragraphs.

The first obvious problem is the inability to view multiple web pages at the same time. Tabs are
maximized to the full size of the windows making it difficult to view other tabs at the same time and
making it difficult to compare the content. But with evolution of open source web browsers like
Firefox it is now easy to design plug-ins which allow tab views to be split into different parts,
making it easy to view 2-3 tabs at the same time.

Web browsers are now used for a variety of purposes like checking email, reading the news, settling
your utility bills, etc. This means users now have a more tabs open than ever before. This has caused
a clutter of tabs at the top bar. The current browsers handle this by reducing the width of the tabs
and trying to adjust as many tabs as possible within the limitations of the display screen. When the
number of open tabs goes beyond a certain number, some browsers open up either a scroll button
or a drop-down menu option to traverse across the different tabs. This means that not all the tabs
are visible to the user at all times. Scalability is now a major issue. This will be discussed in more
detail in the next section.

Another problem having many tabs open is the issue of arrangement and context. Suppose if a user
is using Google Reader to access their daily feeds. The user might want to read about a feed article
and click on its URL to open it in a new tab. He might have other tabs open from different feed
items. Many a times it happens that a user might have skipped one of the open tabs and completely
miss out on reading the contents of a tab. Another problem is in understanding the context of the
tabs. If suppose the user traverses to another link through one such feed item’s article. Later, after a
long time of use, if the user would like to know how he/she reached upon this article, it is not easy
to find out at once. These are issues that have come up recently due to more complex use of the
browser. There is thus a growing need for sorting the tabs and arranging them in “groups” based on
certain user-specified criteria.

What is being done to solve it?

I could not find any technical papers looking into resolving these issues. Nor does it seem to be any
work going on in finding an alternative to tabbed browsing. Given that tabs have been used for a
long time now it would require a huge effort to change an already well established interface. But
there are some interesting developments taking place which build on the existing TDI
                                   TabViz: [2]

                                  This project is handled by a group of 3 students from University of
                                  Michigan School of Information. They have some interesting
                                  browsing strategies. One of the ideas is a "Fan-Shaped view" of the
                                  tabs. They have developed a prototype which shows the tabs in a
                                  radial menu. It is a beautiful visualization that sits at the bottom left
                                  corner. It quiet easily solves the problem of understanding the
                                  context of the tabs. The radial menu is divided into different sectors
                                  each sector representing an open tab. A "child tab" which is opened
                                  through a "parent tab" is represented by hierarchal placements of
                                  the sectors. They even suggest coloring the sectors based on usage
or importance. This would also solve the problem of user-based arrangements of tabs and also to
easily identify tabs which haven't been read yet.

                                                  Favitabs: [3] FaviTabs aims to save some of the
                                                  screen space by replacing the tabs with the favicons.
                                                  The developers assume that users are familiar with
                                                  the favicons that they regularly use and shouldn't
                                                  have a hard time finding the require tab. This design
                                                  allows for 30 tabs to be opened and visible at the
                                                  same time. Another feature which allows for
                                                  grouping of tabs is the concept of "drawers". The
                                                  favitabs can be placed in different drawers which are
                                                  defined by the users themselves. The space below
                                                  the address bar which is usually reserved for
                                                  bookmarks is being used to places these drawers.
The favitabs of the selected drawer is viewable below the drawer space area. So this allows the tabs
to be grouped based on user preferences. The prototype also has a settings page which has a search
filter to search for a web page, sort the tabs and also a timeline view of the tabs.

                                           Collapsible Tab groups: [4] This is a not a radical re-design
                                           in the TDI implementation but still very useful. This more
                                           of a concept than an actual implementation and calls for a
                                           vertical arrangement of the tabs either at a left or right
                                           sidebar. This gives two advantages: more tabs are visible in
                                           a single view and the sidebar width can be adjusted so that
                                           the page title is clearly visible in all of the tabs. It solves the
grouping of tabs by placing the tabs open off another tab under its original parent tab in a tree like
structure. It’s called collapsible because all the tabs grouped under a single tab can be easily collapsed
to a single tab using a click of a button. This also allows for more tab space. It has a search bar at the
top of vertical stack to search for the tabs by URLs or keywords. It also has the ability to highlight
the tabs which was recently updated or color tabs that haven't been read yet.
How good are these solutions?

Almost all the solutions provided seem to solve the arrangement of tabs quiet conveniently. TabViz
takes up too much screen space whereas FaviTabs requires the user to be familiar with the favicons.
Collapsible tabs have a feature where the user has the option to either view it in favicon mode or in
tab mode. Maybe this can be altered to have the regularly used web pages have favicons and the
non-regular ones as normal tabs. This would especially be helpful for collapsible tabs since placing
the favicons next to each other can save lot of space and provide more viewable area for tabs. The
notion of using colors to identify status of tabs seems to work intuitively and is also easy to

One area where all of these solutions fail is in terms of scalability. Opening a lot of tabs still require
scrolling and search filter for browsing across the tabs. A normal user is said to have usually about
10-12 tabs open. These solutions would work very well for such users. But there are power users
who have 30-50 tabs open. A simple poll conducted by Slashdot [5] reveals about 20% of the users
have more than 10 tabs open. Using the scroll feature is no different from the current
implementation and using the search filter requires additional steps. Though favitabs support around
30 tabs, it still increases the cognitive load on the users by having to identify the favicons. Also, it is
not always possible to identify all favicons easily especially when you visit new pages. So there
doesn't seem to be any solution yet for scalability.

It also looks like there may have to be different solutions implemented for different users. For users
having less than 10 tabs open, the above solutions would work just fine. But with having tabs more
than 30-50 we would require additional interface ideas, like maybe using, keyboard navigation to
easily browser across tabs and have easy access to search filters. The problem with scalability is in
using the mouse for browsing across the scroll options and for clicking on the search filters located
at the corners. Navigating through keyboard is much faster and given that power users usually spend
lot of time on the internet, they are used to hotkeys for navigation. None of the above solutions
talks about keyboard navigation.


There is a definitive need to improve the current TDI implementation. There seems to be no
research work or actual design work that suggests a radical shift from the TDI implementation. Most
of them improve upon the current solutions with distinct features that illustrate the need to group
tabs and categorize them based on context and usage. Scalability, though, still seems an issue and
requires more research work.


[1] Designing to Facilitate Browsing: A Look Back at the Hyperties Workstation Browser

[2] TabViz: Visualizing browser tabs in a useful way

[3] FaviTabs
[4] Collapsible Tab Groups

[5] Slashdot Poll | How many browser tabs do you

To top