W3C home > Mailing lists > Public > public-html@w3.org > August 2007

Re: Add Example Explaining that Alt Text is Dependent Upon the Context

From: Thomas Higginbotham <thomas@thomashigginbotham.com>
Date: Thu, 23 Aug 2007 12:14:36 -0400
Message-ID: <46CDB26C.2000202@thomashigginbotham.com>
To: public-html@w3.org

I believe it to be very relevant. For example, if my sighted friend 
tells me to click the image of the flag with stars and a large, red 
asterisk in the left corner, and the alt text simply says, "Australia", 
I might not have a clue what my friend is talking about. Alt text should 
accurately describe the image to non-sighted users unless the image has 
no purpose other than decoration. Using "Australia" as a tool-tip with 
the title attribute is perfectly fine since that is the purpose of the 
title attribute.

It seems that many people are confusing alt and title.

James Graham wrote:
> Philip Taylor (Webmaster) wrote:
>> None the less, I prefer to defer to Gregory who
>> can provide first-hand insight into whether
>> "Australia" as ALT text might be sufficient in
>> certain contexts, and -- if so -- how we might
>> go about codifying those contexts so as to ensure
>> that the specification leaves no doubt as to
>> the context(s) in which such terse ALT text
>> could justifiably be used.
> Imagine context like this:
> <p>Visit The XYZ Company's site for your country:</p>
> <ul>
> <li><a href="uk.xyz.com"><img src="uk.png" alt="United Kingdom"></a>
> <li><a href="us.xyz.com"><img src="us.png" alt="United States"></a>
> <li><a href="au.xyz.com"><img src="au.png" alt="Australia"></a>
> </ul>
> In this case it is strictly irrelevant what the image au.png depicts; 
> it could be the word "Australia" written in the XYZ company's 
> corporate typeface, it could be the Australian flag, it could be 
> anything; the only important fact is that a sighted user with access 
> to the images would identify it as the image corresponding to the 
> choice for Australia. Therefore the role of the alt text, as an 
> alternative to the image, is to allow a user without access to the 
> image to make the same identification. The single word "Australia" is 
> enough to fulfill this task; any more detail about the exact content 
> of the image is unnecessary.

Thomas Higginbotham
Received on Thursday, 23 August 2007 16:17:26 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:15:25 UTC