|
Links Want To Be Links
You are here: irt.org | Articles | World Wide Web (WWW) | Links Want To Be Links Published on: Sunday 9th January 2000 By: Jukka Korpela This document explains why you, as a Web page author, should not try to prevent Web browsers from displaying textual links underlined and image links with borders around them. It also discusses common mistakes in setting (suggesting) colors for textual links. In special cases, it might actually help users if you "tailor" link presentation; there are both HTML and CSS methods for that. The beauty of textual links is explained and compared with the drawbacks of non-textual links and pseudo-links like buttons. Finally we present various ways of "animating" or otherwise enhancing links without destroying their natural beauty. Content
Links are our friends - but what are they, really?Linking turns texts to hypertext, and hypertextuality is among the key factors that make the World-Wide Web a web of interconnected things. Using links on your Web pages, you can conveniently let your readers find background information, technical details, definitions of terms, alternative presentations, etc., on your pages or somewhere else. You can also set up "hypertextual tables of content", and you can chain pages together using links to "previous" and "next" pages. That way you can keep the size of individual pages reasonable, which is essential for fast accessibility. By itself, a link is a relation between things which can be documents or parts of documents or locations in documents. In other words, a link is a pointer or reference. A link does not "do" anything, just as text like "see p. 42" in a book does not do anything. It is something that can be followed, by a browser on user command, or by an indexing robot, or by some other software.
Indexing robots (aka "spiders" or "crawlers") are essential for
search engines,
which in turn are crucial for users' ability to find Web pages.
There's a lot of
talk about being "search engine friendly",
and there are widespread myths about the importance of
The icon at the end of the preceding paragraph exemplifies a method of
attaching "flags" to links
Even if we don't count indexing robots, which follow links their own ways, following a link doesn't mean just clicking on a link and getting the linked page onto screen. Authors often ask how to open a link in a new window or how to "download" a file without "opening it on a browser". There questions reflect ignorance of what links are and how the Web works with different media types. But the situation also illustrates how misleading it is to identify a link with one common way of using links. In particular, on popular browsers, the user can use the right button of the mouse to get a menu of alternatives like "open in current window", "open in new window", and "save onto local disk". Usually authors' attempts to dictate the ways in which links can be followed will just make things more difficult. Links want to get noticedIt belongs to the very basic functionality of a Web browser to indicate links as links. How they do that may vary greatly, depending on the browser and its settings. And it belongs to the very basic skills needed in Web browsing to be able to recognize a link as displayed on one's browser and to follow a link. There is very little you can do as an ordinary Web author to improve the browsers or educate the users. What you can do is that you don't prevent browsers from doing their job or users from making use of that. Recognizability isn't enough. In visual browsing, links should be recognizable on quick scanning. Web users' typical reading habits involve scanning for interesting headings, emphasized texts, images, and other visually distinctive material. Sometimes they (we) read texts in detail; it is even more rare to move the cursor over a particular piece of text or other material; and this answers the suggestion that users can recognize links as links from "tooltips" or from text color (and other properties) changing on "mouseover". The preceding sentence contains one word which is a link, just for illustration; it actually leads to a somewhat one-sided definition. You should have no difficulty in noticing which word is a link, just because I did nothing to suggest any particular presentation of links. Links want to look normalThis document mainly discusses the appearance of links on visual browsers, such as Internet Explorer, Netscape Navigator, Opera, Lynx, Mosaic, and many others. They have different ways of displaying links by default, and there are several ways to change the defaults. However there are many similarities, and some browser features in this area can be regarded as so common that a Web author should usually avoid trying anything that might cause confusion in a typical browsing environment. Whether you like the colors commonly used as link indication by browser defaults is up to you - if you don't, you can change things in your browsing environment. But it is important to realize that if you use such blue colors for some texts which are not links, a very large percentage of your readers will assume that they are links and get frustrated and confused. Similarly, if you manage to prevent browsers from using their normal presentation of links, it's usually a Pyrrhic victory. How underlining may help
In printed matter, underlining usually indicates
emphasis. (Sometimes it could make some other distinction,
e.g. in linguistics one might underline expressions in the object
language.) For good or bad, Web browsers generally don't use such a
convention; rather, many of them by default use underlining for
links. Someone might say that this is one form of
emphasis. But indicating something as more important than normal text
is logically different from setting up a link. Some text could be
emphasized, using e.g. the
To many people, underlining means a link. Simplistic as such
thinking can be, it's a reality. As a consequence, you should not try to use underlining for anything but
links. In particular, you should not use the
A more common problem than attempts to underline non-link texts is that authors often want to prevent browsers from underlining links. Suppressing link underlines seems to be one of the most frequently asked trivial questions on Web authoring, and it seems to be one of the most popular uses of stylesheets. It's mentioned as number 1 in the (assumably somewhat sarcastic) document 5 cheap CSS tricks at the Website Abstraction site. Many people would agree that underlining is per se the wrong way to indicate texts as links. But it's one thing to have an opinion about browser behavior and another thing to try to force other people's browsers to behave in a certain way. If you have a strong opinion about it, think what I would feel, as your potential reader, if I have an equally strong but opposite opinion and you manage to force me to view your page your way? (Right, I'd most probably leave your page at once.) But there is also a very objective reason not to fight: underlining can be the only way in which a browser indicates a link as a link. Authors often think that colors do the job, but that's not something you can really count on. It is a basic accessibility requirement that one should avoid relying on colors as the only method of conveying some information. See Web Content Accessibility Guidelines, principle 2: Don't rely on color alone. The following simple examples illustrate this:
It is fairly popular among Web authors to try to suppress underlining
of links and try to re-establish it on
"mouseover" It has been argued that users can set their browsers to override authors' presentational "commands" if they really need to, e.g. force the browser use underlining of links if links would otherwise be indistinguishable from normal text. That's true in a sense, in most cases, but users need to know that and take the extra trouble. But generally, robust Web authoring should be prepared to having stylistic suggestions overridden by user settings, not rely on that taking place! Naturally, you should not try to force other people's browsers to underline links on your pages either. There can be very good reasons why people have configured their browsers not to use underlines. What about "the ugly borders"?
One of the most common questions on HTML
authoring is how to get rid of the "ugly border", and
the answer
Removing the borders is even more harmful than removing
link underlines It has been said that users can move the cursor around to see which images are links, but that's really not a very brilliant idea. It turns immediate visual perception (via just taking a glance over the page) to a game which is not fun at all. Generally the desire to remove borders relies on the assumption that users behave or should behave in the "click on everything, it might be a link" fashion, or at least in the "don't stop waving your mouse hand, you just might see a kewl tooptip or a status line message which tells you something" fashion.
Normally IE 4+ displays the value of a
A border around an image can surely spoil the
esthetic appearance of an image to some extent; the Dilbert image
above is an example of this, since the image itself is
non-rectangular. Sometimes the effect is really bad indeed. However,
removing a vital usability feature is not a solution except
in special cases
How recognized link colors helpJakob Nielsen, the usability expert, wrote the alertbox Top Ten Mistakes of Web Management in 1996, making a very important point on link colors:
In fact, red color is more typically used for active links, and due to its usual connotations it wouldn't be particularly suitable for visited links. Thus, one can say even stronger than Nielsen did: some flavor of purple is the de facto standard for visited links. This serves the additional purpose of making them less prominent than unvisited links, yet different from normal text. In a review of the mistakes in 1999, Nielsen classifies this error as "severe" and explains it more:
Playing with link colors just because of the author's own esthetic preferences, or ideas of what's "cool", is undoubtedly one of most obvious and most harmful symptoms of bad page design. It is often based on a misguided attempt to design a page or site which "stands on its own". But as Keith Instone has put it in his Usability Heuristics for the Web:
There are two distinct points actually:
The best an author can do about this is to do nothing with
link colors. If however an author suggests link colors, the
suggestions should correspond to typical browser defaults:
blue ( Reversely, if you have decided to suggest any colors or backgrounds on your page, you should normally suggest all the link colors too. (However, a suggestion like a colored border around an element could be made as a "standalone" suggestion.) The discussion above suggests that you shouldn't deviate too much from "normal", which means rather light background and dark text. For example, suggesting black or dark blue background would imply that you can't use normal link colors. See also Colors in Dan's Web Tips. But if you decide to use dark background, make sure the text and link colors are distinguishable against it, and consider suggesting an increased font size and a sans-serif font. The reason is that text which is legible when black on white tends to become rather hard to read when white on black unless you try to help by making it sans-serif and larger than normal. The impact of browsers' ideas of "visited"The preceding discussion may have given a bit overoptimistic idea of browsers displaying visited and unvisited links differently. The problem is that browsers have different ideas of what it means to be visited. (They have different and usually user-configurable rules for reverting a visited page to unvisited state after some period of time, but here we discuss a different issue.) Consider a page containing links to two locations on another page, e.g. to the definition of "glyph" and description of Unicode" in my tutorial on character code issues. If you follow the first link and return, then Netscape shows the other link as unvisited but IE turns it too to visited. Which is better? Hard to tell. But it's surely sometimes inconvenient that you cannot see which locations of a page you have checked by following direct links to them. There's nothing you can do about it in any direct way. But in site design, this is an issue to be remembered when considering whether to write one long page or a set of interlinked pages and which format is the one presented as primary to users if you have both formats. If you have relatively small interlinked pages, then the odds are that even IE's concept of "visited" corresponds to what the user has actually looked at. Could "tailored" link presentation ever help the user?
Let us begin with a simple case where it might make sense to
modify link presentation, not so much for usefulness but for esthetic
reasons on which most people could agree. Assume that our page is a
thumbnail
gallery, i.e. a collection of reduced-size images which are
links to full-size versions. Colored borders might slightly distract
the user's attention from the visual content itself. And since the
distinction between visited and unvisited links isn't usually
essential in such cases, we might explicitly tell the
user that all pictures are links to full-size images and suggest
border removal. This could be done either
using the As regards to removing underlines, consider the rather special case of my page containing a table of characters so that each character is a link, referring to a description of that character. It is clearly better if the characters are not underlined - especially since the table contains the low line (underline) character and some other characters for which underlining would be confusing. Naturally there is an explicit statement about the characters being links. A different special case, somewhat analogous to a thumbnail page (where all images are links) would be a page where all words (except perhaps some headings) are links. It might be better to explain this and try to suppress the underlining, since it is inconvenient to read text where all words are underlined or otherwise different from normal text. This could be useful for language teaching and study material where every word is e.g. a link to a glossary entry. But note that e.g. the texts of the Perseus project do not apply such techniques, although most words in the texts are links. (You can read, say, Caesar's Gallic War in Latin so that almost every word is a link to information about the meaning and grammatical form of the word. Very nice if you know some Latin and wish to learn more.) In fact, the word "most" gives a hint why the technique wouldn't necessary be such a great idea. What would you do if you need to introduce words which are not links, for some reason? Moreover, if the underlining seriously bothers the user, he can probably turn it off in his browser's setting. But this example illustrates what kinds of possible reasons related to the content and structure of the page there might be to affect link presentation.
Finally, probably the most common reason for
affecting link presentation is to make different links look
different. As an example, consider
my attempt to create a hypertext version of an RFC.
It contains both "official" links, corresponding to references in the
RFC itself, and "unofficial" links, which I wrote to refer to
information which might be useful to readers. I'm not that happy with
the results. Basically, I need six different colors (or
actually more - we should consider "hover" colors too): two sets
of unvisited, visited, and active links. It would clearly be harmful
to remove one essential distinction when introducing another one. It
seems acceptable to use the same "active" color for both links, but
this still means five colors which must be clearly
distinguishable from each other and from other colors used on the
page. And the color scheme then becomes rather unintuitive and
difficult to remember. (An author might consider using frames to make
a legend visible permanently, but this would suffer from the usual
drawbacks of frames.)
Especially if you have created a downloadable set of pages so that people can study them offline, too, it would be very useful to distinguish between "internal" and "external" links. Internal would mean a link which refers to something within the downloadable set, and external would thus in practice mean a link which works in online use only.
Yet another distinction one might wish to make is based on the
language of the referred document. For example, a
document in Spanish and mainly intended for Spanish audience might
present links to documents in English (or generally languages other
than Spanish) differently from links to Spanish documents. This would
save those users' time who can't read English. In principle, one
could use the
So how could it be done?Most methods of affecting link presentation are based on stylesheets, or CSS. To learn to use stylesheets, you need to study a good tutorial, and using them you need to consult CSS specifications and documents describing actual browser behavior. For some essential links, see my document How to use stylesheets?. Stylesheets are, by design and as actually implemented, suggestions only, and they can be ignored by browsers for various reasons, some of which have been explained in this document. The same applies to presentational HTML markup, too. Text emphasis
First let's mention some simple but useful HTML
methods. You can put
text-level markup
around No underline
You can suggest suppression of underlining by
specifying
which can be read as follows: if there is an element belonging to
class No border, or a modified border
Suppression of borders can be suggested in a
stylesheet
into a stylesheet. Normal CSS techniques
(selectors)
can be used to limit the
suppression of borders to some
Link colors
Suggesting link colors looks deceptively easy, if you
are familiar with the basics of
RGB colors.
If you use
"browser-safe" colors
only, you can use simple color notation in
CSS: a:link { color : #33f; }
into a
stylesheet
However, suggesting link colors the right way takes more. It's a matter of knowing the correct approach and applying it consistently. The motivation for the rules can be found in material on CSS; here we just list the rules:
As regards to rules 1 and 2, compare them with the reasons for the
good old recommendations to set all
a:link { color : #009; background : #fff none; }
a:visited { color : #609; background : #fff none; }
a:hover { color : #f66; background : #fff none; }
a:active { color : #f00; background : #fff none; }
Setting the background image explicitly to
For an explanation of what such selectors mean in detail, see the
CSS2 Specification,
sections
The link pseudo-classes:
The example above more or less corresponds to typical browser
defaults. Changing the color codes you can modify the presentation,
but think twice before making them very different from the defaults.
Note:
Rule 3 means, for example, that in order to strongly emphasize a link,
you put the
It is sometimes suggested that
The technique discussed above can also be used to highlight
links when they are selected by the user, i.e. to achieve
"mouseover"
effects more easily than using JavaScript. You simply specify
suitable rules for a:hover { font-weight : bold; }
This could be used without any other
stylesheet
There are some relatively complicated, mixed-technique ("DHTML") methods which work on Netscape 4, too; see Martin Webb's What is So Dynamic About Dynamic HTML?, especially the example Altering Content for both Internet Explorer and Netscape 4.0. The technique works, under some conditions, even for making normal text (not link text) change properties and content on mouseover. Perhaps this applies to the heading text "Mouseover effects" above, on your browser.
Note that the CSS approach of using
<script type="text/javascript"><!--
function popup() {
window.open('popup.html','','width=300,height=150,'+
'resizable=1,screenX=300,screenY=300,top=300,left=300');}
//--></script>
and the body part might then contain <a href="http://www.irt.org/articles/js058/" title="Events and Event Handlers" onmouseover="popup()" >event attribute</a>
But beware that such popup windows are generally regarded as a
nuisance. At least make sure you don't leave the unexperienced user
with a window he can't close or enlargen or otherwise manipulate! For
technical information, see
JavaScript window FAQ
and
WebReference's
Window Spawning and Remotes.
If you really have meaningful use for such techniques, consider using
frames instead of popup windows. You could divide the
window into a main frame and a small "footnote frame" under it, and
use e.g. Making different links look different
If you wish to suggest link colors in a manner which
distinguishes between different types of links
a:link { color : #009; background : #fff none; }
a:visited { color : #609; background : #fff none; }
a:hover { color : #f66; background : #fff none; }
a:active { color : #f00; background : #fff none; }
a.ext:link { color : #036; background : #fff none; }
a.ext:visited { color : #336; background : #fff none; }
a.ext:hover { color : #f66; background : #fff none; }
a.ext:active { color : #f00; background : #fff none; }
Note that the
Can you make multidimensional classifications of
links? Say, can you use different colors for internal and external
links and make a visual distinction between links to
tutorials and links to technical references using some other method,
say different fonts? In principle yes. But it's not easy. In
particular, it is possible in HTML to specify that an element belongs
to several classes, say As an alternative approach to making distinctions between different types of links, consider the following methods which let you avoid interfering link colors at all (i.e. you'd let colors default to what the user is accustomed to, which is almost always the optimal option):
Attaching "flags" to links
The most reliable method of indicating distinctions
between different types of links is to add textual
explanations. They need not be long. You can use, say, the
string "[X]" after each external link, and put somewhere near the
beginning an explanation (legend) of the notation(s). You
might provide image alternatives for them, too (using markup
like Since such "flags" are part of the normal textual content, they work on all visual browsers as well as on non-visual user agents. Note that although a legend is needed, just as you'd need one if you used a color scheme, the notation could be relatively intuitive and easy to remember, as opposite to color assignments which would be rather arbitrary.
In the HTML markup, it's probably best to separate such a
notation from the preceding text by a
no-break space,
These notations would help in avoiding disappointments and waste of time. I suppose you've seen and followed links like ISO 8879 just to see that they lead basically to ordering information only, not to an online version of the document you're interested in. So [B] might help. Similarly, [<-] may help in realizing that a link leads to something you've already read. To someone who has jumped into the middle of a document, the same notation would be informative, too.
The notations [->] and [<-] don't look nice, but
they are probably the best we can achieve using short sequences of
ASCII characters. You might wish to consider using small images,
naturally with
Sometimes you might make such a string itself a link, especially if
you don't want to cause a long text such as the name of a publication
to be presented in "link style". Thus, in particular, you could use
The beauty of textual links - and the drawbacks of the surrogatesWe have presented some reasons to avoid using images as links. Here we take a more positive look: it's not basically a matter of what's wrong with image links but what's so nice about text links. How they workConsider the process of detecting and using a normal, textual link:
We have stated the obvious, namely the typical properties of textual links, to emphasize that other types of links or their surrogates usually lack one or more of those well-known properties. Consequently, they imply deviation from the users' normal mode of surfing. They are more restrictive, too; consider, when reading about the surrogates in the sequel, for example what you would do if you wanted to open the "linked" resource in a new window.
To make the best possible use of the beauty of
textual links,
write descriptive but concise link texts.
Obviously, texts like "Click here" don't satisfy this
requirement - and they look really stupid on paper, and
sound stupid when read aloud! There are also several other
dangerous words
which you should avoid in link texts and around them. Ideally, the
link construct
For further illustration, let us consider some alternatives. In all
cases, we wish to link to irt.org main page, i.e.
Grouping textual linksYou can organize sets of links in various ways, e.g. as lists or tables. To take a simple example, the following small table contains links to irt.org main page and to top level subpages as a nice, compact set:
One problem with such a table is that it may force horizontal scrolling if the browser window is narrow. So unless the set of links is rather small, a simple sequence of links might be better. You would just put the links into a paragraph and separate them with spaces or, preferably, with spaces and some separator characters like vertical line "|". Since it is preferable not to break a line before the separator, it's best to use a no-break space before it; and after the separator, you can use a space or a newline (which are equivalent). So the markup would be something like
resulting in a presentation which automatically adapts to the window width: Articles | FAQs | Games | BBS | Resources | Software | Books | Downloads | About | Beta | Find
For large collections of links, there are
several ways of organizing them e.g. into tables.
You could also use nested lists (typically, a
Adorned textual linksAn unadorned textual link like irt.org is extremely suitable for normal use. It is particularly convenient when it is to appear in running text, since it has the same text size as normal text.
Text emphasis
One could use increased font size like irt.org or even irt.org in some cases. In running text, this is usually acceptable only if the link is probably displayed at the first line.
Using a colored presentation would be more risky, for
reasons explained above,
but a careful slight change in background color might be
acceptable and useful. One might use a very light gray background like
Fancy colors like those used on the irt.org main page itself might work in some cases, but only if it is evident that the texts are links and if it does not matter too much that the unvisited/visited distinction is lost. It does matter a lot e.g. on a page which contains a list of links leading to separate documents, like a FAQ index page, say JavaScript VFAQ page. How else could the user keep track of which answers he has consulted when trying to educate himself in general or to find an answer to a particular question? And even on basic navigational pages, such colors, suppressing the distinction between unvisited and visited links, might make life harder to first-time visitors. They have problems in figuring out how the classification of things really works; the link names often don't say much. Perhaps one could suggest that unvisited links be underlined and visited links be not underlined?
There are many other variations one could apply to
textual links, like suggesting a particular font face or putting them
into boxes, either using
stylesheets
Images as linksLet us then consider the use of an image, like the irt.org logo, as a link:
You should write no space or line break between the ">"
which terminates the
This works reasonably well, if it has been designed robustly
in the sense of including a
useful
and just replace the content If you have a company logo on all pages, you might make it a link to the company's main page. People might guess it could be a link and get accustomed to using it; but then you would need to make sure every occurrence of that logo is a link! Moreover, don't rely on a logo as a link. Most probably you can find a natural way of including a textual link too, e.g. by making the first occurrence of the company's name a link. Image mapsImage maps are surprisingly often used for "navigational" links, despite serious accessibility problems. In the most tragicomic cases, all information at a site is purely textual, but the only way of accessing it is via an imagemap menu, which does not work or works very clumsily in text-only mode. Image maps can be great for using geographic maps and other things which are inherently graphic and two-dimensional by their nature. But turning simple menus to image maps is generally counter-productive. The following image map has been built from a screen capture of the irt.org main page, which contains a menu which is colored but textual. ![[irt.org links]](irtmap.gif)
If you use an imagemap, remember to include an JavaScript pseudo-links, and links with target
Client-side scripting
can be used in many ways in conjunction with links. In fact, if you
wish to have a scripted event on your page, the way which works most
widely is to add the
A typical example is the most common (?) way of setting up a
"link" which opens a document in a new window,
a "popup" window. There is a robust way
too, so there is no reason to do it so that when considered at the
HTML level - which is what e.g. indexing robots do - it's a
dysfunctional link with
One can also specify, or rather suggest, that a link
be opened in a new window, or in a particular named window or
frame, by using the
Using a normal link "enhanced" with both Java pseudo-links
The
Java language
(which is so often
confused with JavaScript)
can be used to write programs which might get executed by the
user's browser. And such a program, or applet, could be
written so that it provides a "navigational menu". It can be written
so that works in a manner similar to
image maps
A Java applet which creates pseudo-links that way can be written in a
manner which degrades gracefully on browsers which do not support Java
or have Java disabled. Just write the alternative menu into the
content of the
"Form-based" methods: buttons and pulldown menus
For some reason, authors often wish to use buttons as
links. That's technically rather simple: just write a
This is confusing, since a button suggests form submission, which typically means at least a dynamic query, perhaps something more serious like ordering a product. If "buttonistic" approach is combined with relying on client-side scripting, i.e. if the form is non-functional when such scripting is disabled, you've gone pretty far - from usefulness and accessibility. Another popular way to build "navigation" is to use "pulldown menus". They normally involve more than one alternative. Example: See the document Navigational pulldown menus HTML for a discussion of both the techniques (client-side and/or server-side scripting) as well as serious arguments against the approach. It also lists reasons why normal links are superior to such menus, and most of the reasons apply to any comparison between links and pseudo-links.
There is a special reason not to use forms in vain:
form submissions can be regarded as potentially dangerous. This
normally applies to forms using the
"Animating" linksLink texts could be made grow big or turn green on mouseover, in some browsing situations, as discussed above. Here we shall discuss other methods which might serve the purpose of providing additional advisory information related to links. The basic idea is to let some users get extra info without following the link. You might view this as an opportunity to "advertize" the linked resources and to allure people into following the links, or you might regard it as a method to prevent people from wasting their time (and network resources) by following links which really won't give anything to them. There need not be a contradiction: by helping people to avoid links which are useless to them, you help them to make better use of those which are useful.
The basic tool for giving such extra information
is the
One interesting technique is the one used at the lists of articles pages at irt.org, such as the page which lists all articles. The page contains, in addition to article names which are links to the articles themselves, a "Synopsis" (summary) and a "Techniques" (keywords) text for each article, below the article name. This of course means that the page is rather long. But using certain techniques one can make several browsers hide those texts and display them only on mouseover so that they appear in a box which covers part of the normal document content. So one gets a "tooltip" effect, which is rather flexible but requires somewhat complicated "DHTML" techniques. Discussion of "DHTML", which is actually a mixture of various techniques, is beyond the scope of this article. But if you wish to study the techniques, essential ingredients include
This approach means that no new window is created, so some of the
problems of
popup windows
Another, more common way to create a "tooltip-like" effect in some
browsing situations is to use the
event attributes like <a href="http://nmnhwww.si.edu/msw/" title= "Scientific information about mammal species, organized taxonomically (by order, family, subfamily, genus), by Smithsonian Institution" onmouseover="window.status= 'Taxonomically organized info about mammal species by Smithsonian Institution'" onmouseout="window.status=' ';return true" ><cite>Mammal Species of the World</cite></a> On a JavaScript-enabled browser, the text "Taxonomically organized info about mammal species by Smithsonian Institution" will then appear at the status line when the cursor is on the link. Independently of this, IE 4+ and Opera display the other advisory text as a "tooltip". Note that the text on the status line is relatively unnoticeable, and often displayed as truncated. If you wish to make the advisory text appear more strikingly, you need to use other techniques discussed. There is also a way to use images as tooltips, if JavaScript is enabled:
The "tooltip images" might be useful e.g. to a reader who does not
quite know the English words for the breeds but recognizes breed types
from images. The example above uses just basic JavaScript techniques:
simple code for Further readingFor other important aspects related to links in Web authoring, see especially
Jukka Korpela, Jukka.Korpela@hut.fi Feedback on 'Links Want To Be Links'
View the author profiles.
| |
.
![[Information about dog breeds]](1.gif "Information about dog breeds")
.
since with textual links, there's a considerable chance that they will
be seen in a color which tells the user they're links (or perhaps with
a frame around them, or something). For image links, it seems that
the border is the only way for browsers to indicate them as
links, so does it make sense to try to prevent that?
![[an image which symbolizes the AcroRead program and the PDF format]](pdf.gif)
irt.org liability, trademark, document use, privacy statement and software licensing rules apply.