king.p65
Library Home Page Design 457
Library Home Page Design:
A Comparison of Page Layout for
Front Ends to ARL Library Web Sites
David L. King
The author examined the home pages of all 120 libraries in the Associa
tion of Research Libraries (ARL) in order to compare design similarities
and differences. This was accomplished by first filling out a question
naire on each home page and then tabulating the findings using simple
averages and medians. Areas examined were: backgrounds, document
headers, document footers, document body, page length, number of steps
to library home page from parent institution Web site, and domain name
servers. A typical home page is discussed using the averages of these
results.
he World Wide Web, once used
by a handful of researchers,
now is used by everyone from
university faculty members to
their thirteen-year-old children. Home
pages abound, with content ranging from
how to cook a Butterball turkey (http://
www.butterball.com/) to information on
President Clinton’s cat, Socks (http://
www.whitehouse.gov/WH/kids/ html/
pets.html). And libraries are no exception.
Many libraries have a Web site; the ones
that do not probably have one in the plan
ning stages.
What do these library home pages look
like? More important, when a visitor finds
a library’s main, front-end home page for
its Web site, what does he or she see? The
purpose of this project is to examine the main,
front-end home pages of libraries in the As
sociation of Research Libraries (ARL) in or
der to note current trends in library Web
page front-end design, and to create a
“typical” library home page design based
on this project’s findings.
Literature Review
Mark Stover and Steven D. Zink have
written an article comparing the design
of forty higher education library home
pages, chosen through Yahoo!, a popular
search directory.1 Their research “was
designed to measure the relative quality
of a particular genre of home pages (uni
versity and college library home pages) in
terms of design and organization.”2 To ac
complish this goal, they assigned a twelve-
point scale to various features found on li
brary home pages, such as the number of
links found or if the author/designer of
the home page was identified. Their con
clusion reflected that many library home
pages are badly designed, possibly because
of a “lack of knowledge of hypermedia and
a sketchy, emerging literature pertaining
to Web page design principles.”3
David L. King is the Electronic Services Librarian in Cook Memorial Library at the University of South
ern Mississippi; e-mail: d.king@usm.edu.
457
mailto:d.king@usm.edu
www.whitehouse.gov/WH/kids
http:www.butterball.com
458 College & Research Libraries September 1998
In 1995, Pieter A. van Brakel, Cerina
Roeloffze, and Amanda van Heerden said
that “not much has been published which
could be of use to guide a person when
creating a home page for a specific pur
pose.”4 In just over one year, the field of
home page design books and articles has
taken off. Now, one can easily find articles
and books that describe Web page design
in general. Although these books and ar
ticles do not focus specifically on library
Web pages, they are still useful for un
derstanding design basics.
The goal of this project was to
examine the main, front-end home
page of all 120 ARL libraries in order
to compare design similarities and
differences.
A very helpful article on Web design
is by Kathleen Falcigno and Tim Green.5
Both Lincoln D. Stein 6 and Paul
McFedries7 have written useful books on
the subject, Stein focusing on information
providers. Some Web sources that are
useful are the Web Style Manual 8 and the
Guide to Web Style.9 In addition, some
library sites have begun to publish
their own Web page standards. One of
these can be found under the library Web
site at the University of Wisconsin-Madi
son.10
Methodology
According to Stover and Zink, “elements
common to many home pages, whether
by design, imitation, or default, are: a de
scriptive title, a header that names the
organization or sponsoring body, a short
paragraph describing the purpose of the
Web site, link choices that often are com
bined with descriptive icons (roughly
equivalent to a table of contents), and con
tact information identifying the author of
the Web site.”11 Most Web home pages
consist of four main parts:
1. The header describes what can be
found at the Web site.
2. The body provides most of the in
formation found on each Web page.
3. The footer gives contact informa
tion, copyright information, and tells
when the page was last revised.
4. The design concept of the home
page.
The goal of this project was to exam
ine the main, front-end home page of all
120 ARL libraries in order to compare
design similarities and differences. This
was done by first creating a survey of
forty-four questions pertaining to home
page design. Next, each ARL library home
page was examined and the survey filled
out for each. Then, the data were tabu
lated using simple averages (and medi
ans in two cases). This was done to estab
lish a current state of the art in ARL li
brary home pages. A “typical” home page
is discussed using the averages of these
results.
Results
The results can be categorized in seven
sections:
1. Backgrounds
2. Document headers
3. Document footers
4. Document body
5. Page length
6. Number of steps to library home
page from parent institution Web site
7. Domain name server
Backgrounds
In HTML, the background for a home
page is “blank” by default and usually is
rendered gray in popular graphical Web
browsers such as Netscape. However,
Netscape has popularized the idea of us
ing other textures, colors, and designs for
the background, which can “provide a
strong thematic design element for a
page.”12 Backgrounds in HTML currently
consist of two primary types: a back
ground image and a background color. A
background image is an in-line graphic file,
in either .jpg or .gif formats, that has been
designated as a background to the home
Library Home Page Design 459
page within the
tag. It can add
color and/or texture to a home page. A
background color consists of an extension
to the tag that specifies a cer
tain color scheme and replaces the default
color with another color chosen by the
HTML author.
Two items were examined in relation
to background images/colors:
1. Was a body background used?
2. When backgrounds were used, was
a background image or a background
color used?
Fifty-five percent of the test group did
not use a body background, opting, in
stead, for the default background. Of the
45 percent using body backgrounds, 19
percent used a background image in ei
ther a .gif or a .jpg extension, 24 percent
used a solid background color, and 2 per
cent used a combination of the two. Of
the 2 percent using combinations, one li
brary used a solid white background im
age (rather than a white background color
setting), and another used a combination
of an image of a green bar, running verti
cally down the left margin, and a solid
white background color.
Document Headers
“A consistent titling design at the top of
WWW pages allows the reader to imme
diately know what the main point of the
document is, and what (if any) relation
ship the page may have to other pages in
a related group.”13 This titling concept
is also known as a header. “Nobody
wants to scour a large chunk of a page
to determine what it is all about. In
stead, include a descriptive, large head
ing ( or ) at the top of the page
to give your readers the instant feedback
they need.”14
A document header is placed at the top
of a home page and gives the home page
reader information about that page’s con
tent, similar to the function of a book or
article title. Sometimes headers consist of
in-line images displaying the name of the
Web site, or they may simply be text.
Headers also may include graphic images
or hypertext links.
Five items were examined in the docu
ment header:
1. Was the library’s name included?
2. Did the name consist of text or an
in-line image?
3. Did the header mention the univer
sity?
4. Was there any linked information
included in the header?
5. Were any in-line images included
in the header?
All 120 home pages included a title of
some sort. Of these, 26 percent were text-
based headers, 71 percent used some type
of digital image, and 3 percent used a
combination of the two.
The library’s name was included in the
header in 96 percent of the test group. Of
the 4 percent that did not include a name,
usually the word library was mentioned
in a general sense, but the specific proper
name of the library itself (i.e., University
Libraries of the University of Southern
Mississippi, Cook Memorial Library, etc.)
was not included in the header informa
tion. Also, 83 percent of the document
headers that were examined mentioned
the library’s parent institution (usually in
this format—the University Libraries of
the University of Southern Mississippi).
Because some ARL libraries are not uni
versity libraries, this criterion did not
apply to 9 percent of the test group.
Next, the inclusion of in-line images
was noted. Graphical in-line images usu
ally consisted of a drawing or a photo
graph. Sixty-one percent included a
drawing with their name, usually an em
blem (such as the university seal or the
graphic images found on official library
publications). Also included were deco
rative icons and bullets. Photographs
were included in 12 percent of the test
group, usually of the library building or
portrayals of library life (i.e., students
studying).
Hypertext links were included in 34
percent of the headers. Thirteen percent
460 College & Research Libraries September 1998
of the headers included various other
items, such as: two Common Gateway
Interface (CGI) program links, informa
tion about Netscape, comments and dis
claimer information, contact information,
city and state of the library, an image map,
a paragraph of text about the library, Web
page version release notes, a counter pro
gram, home page creation comments, a
photo credit, the modification date, and
the abbreviation for the university.
Document Footers
“Footers, the elements at the bottom of
WWW pages, are extremely important in
identifying the origin, authorship, author
information, legal status, and last revision
date of the page . . . .”15 Footers usually
include many types of information, includ
ing author ’s name, author’s institutional
affiliation, hypertext links to related local
pages, copyright dates, revision dates, con
tact e-mail addresses, official marks or
logos, and institutional affiliations.16
Fifty-eight percent of the document
footers examined included an author or
author ’s name in a variety of formats. Of
these, there were three varieties:
individual’s names, general names, and
committee names. Of the individual’s
names, 22 percent listed a full name and
2 percent gave an abbreviation of the
individual’s name. Of the general names,
Web master was used most frequently
(12%); there were also variations on Web
master (Web manager, Web administra
tor, Web editor) and original, cute names
(Web housekeeper). Also, ten percent
gave committee or departmental names
such as the Reference Department, the
Web Services Team, or the INFOMINE
Development Team.
Some type of contact information was
given in 70 percent of the document
footers. Usually, this was an e-mail ad
dress (67.5%) and was linked to the
hypertext tag. Three percent of
the libraries included mailing addresses.
In 43 percent of the document footers, the
author ’s institutional affiliation was in
cluded, usually by mentioning either the
library or the university.
Only 33 percent of library footers did
not include hypertext links. Of those that
did include some type of link, ten percent
were links to the university home page
alone. Twenty-seven percent included a
variety of links to things such as Library
Hours, a University Welcome, or a What’s
New type of link, but not the university
home page. Thirty percent of footers in
cluded both the university home page
link and other local links.
A copyright statement was included in
17.5 percent of footers. Sixty-three percent
included a revision date for the home
page. Some type of official mark, seal, or
logo graphic was included in 12.5 percent
of the footers, and 19 percent of library
home pages included the Uniform Re
source Locator (URL) of the page in the
footer.
Other things that appeared included:
Only 33 percent of library footers
did not include hypertext links.
CGI counter programs (six libraries),
copyright statement for a photograph
(one library), the library’s telephone num
ber (four libraries), the author ’s job title
(two libraries), JavaScript at the bottom
of the screen (two libraries), and an “Un
der Construction” statement (one library).
Document Body
The body of a Web page is similar to the
body of a letter; it contains the primary
content of the home page. This section can
be divided into three areas: graphics,
links, and text.
Graphics
Graphic in-line images were included on
98 percent of the home pages. Buttons and
icons were found at 53 percent of the sites.
Only three library home pages included
photographs of the library, and one in
cluded a photograph of a book and a sign
http:affiliations.16
Library Home Page Design 461
TABLE 1
Number of Links per Home Page
No. of
Links Libraries Percentage
0-9 18 15%
10-19 45 37
20-29 36 30
30-39 10 8
40-49 4 3.5
50-59 4 3.5
60-69 1 1
70-79 1 1
80-89 1 1
Average 21.62 links
Median 19 links
post. Tables, which are sometimes digital
images, were included on 24 percent of
the home pages. Drawings/graphics that
were not used as clickable buttons or
icons were included in 34 percent of home
pages. Image maps were found on 15 per
cent of the home pages.
Different types of divider lines also
were used. Mostly, they were achieved by
the
(horizontal rule) tag, which
produces a line across the page. It was
used on 67.5 percent of the pages. A digi
tal in-line image of a divider line, in .gif
or .jpg formats, was used on 16 percent
of the pages, and 7.5 percent of the pages
used a combination of
tag lines and
.gif/.jpg lines. Only 9 percent of the test
group chose not to use any type of divider
line within the home page.
Other miscellaneous graphics found
include: in-line images that say “New!”
(used in 7.5% of the test group), blinking
text (one library), HTML default bullets
rather than bullet images (7.5%), and one
Beta HTML 3.0 .gif in-line image.
Links
The average number of hypertext links
per home page was about twenty-two,
and the median was nineteen. Two pages
had only five links, whereas one site in
cluded eighty-one links on its home page
(see table 1). Ninety-eight percent of li
brary home pages used typical, text-based
links, whereas one library used only but
ton-type links and one library used only
an image map to link to information
within the Web site. Of the text-based
links, 26 percent provided only the link,
with no description other than that found
within it (i.e., Library Catalog), compared
to 16.5 percent that provided a link and
an explanation of where it went other
than within the link itself, usually in a
summary or paragraph format. Fifty-six
percent used a combination of links with
descriptive text and links with no descrip
tive text other than that found within the
link text itself.
A great majority of ARL library Web
sites (91%) used the default color (usu
ally blue) as the text link color. Other col
ors found were: red (one library), dark red/
maroon (one library), green (three librar
ies), dark blue (two libraries), yellow/tan
(one library), dark purple (one library).
Two libraries used no text-based links.
Graphic-based links were used at 63
percent of the Web sites. Button-style links
were used in 55.5 percent of the test
group. Of those, 4.5 percent used a but-
ton/icon with no text to describe what
that button linked to, whereas 50 percent
used a combination of button/icon and
descriptive text. A combination of buttons
with text and buttons with no text was
used on one percent of the test group.
Image map links were found at 16 per
cent of the libraries. All these used a com
bination of graphics and descriptive text
that described the image map links.
Text
In this section, text refers to unlinked text.
Text was included on 117 ARL library
home pages. The font size of the main
body of text was usually the default set
ting (91%). Three libraries used the
tag for all the text, two used the tag,
and four used Netscape font extensions
(+1, -1, +4, 2). Of course, the two libraries
that included no text-based links also
used no unlinked text.
462 College & Research Libraries September 1998
TABLE 2
Number of Lines of Text
per Home Page
No. of
Lines of Text Libraries Percentage
0 13 11%
1-9 60 50
10-19 30 25
20-29 4 3
30-39 4 3
40-49 2 2
50-59 1 1
Average 10.31
Median 8
The amount of text on a home page
varied from one line to fifty-one lines. The
average number of lines was ten, and the
median number of lines was eight (see
table 2).
Text headers were used frequently
within the body of the home page. Linked
text headers were used in 27.5 percent of
the test group, 27.5 percent used unlinked
headers, 16 percent used a combination
of linked and unlinked headers, one
page used a graphic image header
rather than a text-based header, and 28
percent used no headers on the home
page.
Page Length
There are differing views on optimal page
length for individual home pages. Ac
cording to Stein, “A good rule of thumb
is to make a page at least as long as a
screen, and not longer than ten screens,”17
whereas McFedries says “nobody likes
scrolling through endless screens of text.
Pages with more than three or four
screenfuls of text are hard to navigate and
tend to be confusing to the reader.”18
For this study, page length was deter
mined by using the Print Preview com
mand in Netscape. Sixty-eight percent of
home pages were one page in length, and
29 percent were two pages. Two sites were
three pages in length, and one site was
four pages long.
Number of Steps to Library Home Page
from Parent Institution Web Site
When browsing the Web for library sites,
one notices differences in placement of the
library site on the parent institution’s Web
site. Sometimes the library home page can
be found easily, and other times one must
literally hunt for it for five to ten minutes.
Forty-four percent of library home
pages could be found in one step, and 37.5
percent could be found in two steps. Ten
percent could be found in three steps, and
one was buried under four steps. Nine
library home pages had no parent insti
tution and were thus excluded from this
criterion.
Domain Name Server
Whether the library home page is located
on the library’s domain name server was
determined by the URL address of the
site. Usually, if the library has its own sys
tem, the URL follows the style of “http://
www.lib.name of university.edu/.” If the
library has an address such as “http://
www.university of blah.edu/library.html”
or “http://www.university of blah.edu/
academic/services/site_1/buildings/
Smith library/library.html,” this usually
indicates that its home page is stored some
where other than at the library. At the
least, the tunnel to access the library’s
page is supplied by someone else. Sixty-
four percent used the format that indi
cated their own server, wheras 36 percent
used another server to store and/or pro
vide access to the library’s home page.
The Typical ARL Library Home Page
Using the statistics taken from this sur
vey, a clear picture of a typical ARL library
home page emerges:
� Backgrounds: The typical ARL li
brary home page has a 50/50 chance of
some type of background image or color
because nearly 50 percent of libraries used
this design format. If a background is
used, it is most likely a solid color, prob
ably white or tan, but it is equally pos
sible to find a page using default colors.
http:blah.edu
http://www.university
www.university
http:university.edu
www.lib.name
Library Home Page Design 463
� Document headers: The name of the
library is displayed, most likely in a
graphical, in-line image file format (.gif
or .jpg). The image mentions the names
of the library and the university, probably
using this format—University Libraries
at the University of Michigan. The header
probably does not include any hypertext
links, but there is possibly some type of
graphic image along with the name of the
library, such as a symbol used on the
library’s stationary or a university seal.
The findings presented here are not
meant to show what is being done
right in library Web pages, but
simply what is being done.
� Document footer: Probably some
type of author/maintainer of the Web site
is listed. This most likely appears as a gen
eral name, such as Web master, and this
name is linked to a tag, which
acts as an e-mail link. The author/main
tainer lists his or her institutional affilia
tion. Some hypertext links are included
in the footer area, such as a link to the
mother institution’s home page. A revi
sion date for the home page is given.
� Graphics: In the body of the Web
page, some type of graphics are em
ployed, mainly
lines used for di
viders between sections of the document
and buttons used to link to important
places, such as the library catalog. Image
maps and tables probably are not used.
� Hypertext links: Links included are
mostly text based, and the link text uses
the default color (the default in Netscape
and other popular graphical Web brows
ers is blue). Some links include a brief
description of content for that link, and
others provide no explanation other than
that found within the link text, such as
Library Catalog. The graphical links
found most likely appear as clickable
buttons, which have some type of de
scription located on them, usually a word
or phrase. The average home page has
approximately twenty-two hypertext
links.
� Unlinked text: Unlinked text is in
cluded on the typical ARL library home
page. There are approximately ten lines
of unlinked text, and the default size is
used for the text. The text consists of brief
welcome messages and/or descriptions
of hypertext links. Some type of text
header is used to separate different sec
tions of the home page, and these head
ers can be linked text, as well.
� Additional highlights: The typical
ARL library home page is approximately
one printed page long, and can be found
from the parent institution’s main home
page in one or two steps. The library most
likely houses the server for the home page
within the library.
Conclusion
The purpose of this study was to exam
ine current trends in library Web page
front-end design and to describe a typi
cal library home page design based on
that examination. The findings presented
here are not meant to show what is being
done right in library Web pages, but sim
ply what is being done.
Through close examination of the li
brary home page standard and consider
ation of deviations from that standard,
Web masters can create or improve their
own library home pages, combining uni
formity with originality so that visitors
to library home pages can quickly find
the information they seek.
464 College & Research Libraries September 1998
Notes
1. Mark Stove and Steven D. Zink, “World Wide Web Home Page Design: Patterns and
Anomalies of Higher Education Library Home Pages,” Reference Services Review 24 (fall 1996): 7–
20.
2. Ibid., 12.
3. Ibid., 15.
4. Pieter A. van Brakel, Cerina Roeloffze, and Amanda van Heerden, “Some Guidelines for
Creating World Wide Web Home Page Files,” Electronic Library 13.4 (Aug. 1995): 383.
5. Kathleen Falcigno and Tim Green, “Home Page, Sweet Home Page: Creating a Web Pres
ence,” DATABASE 18 (Apr./May 1995): 20–28.
6. Lincoln D. Stein, How to Set Up and Maintain a World Wide Web Site: The Guide for Informa
tion Providers (Reading, Mass.: Addison-Wesley, 1995).
7. Paul McFedries, The Complete Idiot’s Guide to Creating an HTML Web Page (Indianapolis:
Que, 1996).
8. Patrick J. Lynch, Web Style Manual. Yale Center for Advanced Instructional Media. [Online]
1996. Available at .
9. Rick Levine, Guide to Web Style. Sun Microsystems, Inc. [Online] 1995. Available at .
10. Nancy McClements and Cheryl Becker, “Writing Web Page Standards,” College & Research
Libraries News 1 (Jan. 1996): 16–17. Also found in Electronic Library Access Committee (ELAC),
User Documentation Working Group, UW-Madison Campus Libraries Web Page Standards and Guide
lines. [Online] June 19, 1995. Available at .
11. Stove and Zink, “World Wide Web Home Page Design,” 9–10.
12. Levine, Guide to Web Style, available at .
13. Lynch, Web Style Manual, available at .
14. McFedries, The Complege Idiot’s Guide to Creating an HTML Web Page, 167.
15. Lynch, Web Style Manual, available at .
16. Ibid.
17. Stein, How to Set Up and Maintain a World Wide Web Site, 267.
18. McFedries, The Complete Idiot’s Guide to Creating an HTML Web Page, 164.
http://info.med.yale.edu/caim/sylemanual
http://info.med.yale.edu/caim/stylemanual
http://www.sun.com/styleguide/tables
http://www.library.wisc.edu/help/tech
www.sun.com/styleguide
http://info.med.yale.edu/caim/StyleManual_Top.HTML