W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2001

RE: title inclusion

From: Mirabella, Mathew J <Mathew.Mirabella@team.telstra.com>
Date: Thu, 6 Dec 2001 11:13:26 +1100
Message-ID: <73388857A695D31197EF00508B08F298061E2F56@ntmsg0131.corpmail.telstra.com.au>
To: "'W3C WAI Web Content Accessibility Guidelines list'" <w3c-wai-gl@w3.org>
Joe/All.

What I discussed in my email was an outline of how I have used alt and
title.  I was not putting up my practices as guidelines at all.  I was just
wanting to give my reasons for using certain HTML features the way I do.

If my way of using them is, in part, disagreeable to some people, then all
the more reason for version 2.0 of the guidelines to contain detailed and
specific information on how these attributes should be used.  Developing
this detailed information is something that we should work on, and I think
we are.

I agree that we need to write to spec, and that "faulty" devices need to be
fixed.  I too am not happy with hacking around incompatibilities.  However,
sometimes I do hack around these incompatibilities so that users can receive
accessible content "NOW".  But I do this only if it does not degrade the
level of accessibility.

Throughout my examples, I have not used the title= for graphics at all.  My
discussion was based around the use of alt= for graphics and title= for
links, not title= for graphics.  When I said "Generally use", I really meant
"always" and should have stated as such.

So I always use alt= for graphics, never use title= for graphics, and
sometimes use title= for links when the link element content is not
descriptive of the function of the link.

I too feel that consistency is important.  Link element content should
therefore consistently be descriptive of the link so that the title= for
links is not required.  But when I have to retro-fit a site where the
designers simply will not agree to change the link element content because
of changes that would occur to the appearance on the screen, I find the
title= for links very useful.

As for re-coding all the links to contain redundant title= attributes just
so that I can say "all the links have title's" seems to be a waste of time.
However, I guess it is not unreasonable when you have tools which can do the
job quickly.  Certainly, when starting out from scratch developing a site,
there is no reason for not using title= for all links.

Also, the style sheet example you gave works well;
ul { list-style-image: url(image-URL); list-style-type: disc }

But if you want to make the graphics clickable as part of the link, you have
to place the <a> before the <li> for each link, closing the </li> first then
the </a>.  HTML tidy suggests that this is not valid HTML.  I wonder if this
is just an HTML Tidy warning, and not something to worry about.

So as you can appreciate, the issues are complex and not so black and white.

Cheers.

Mat.


-----Original Message-----
From: Joe Clark [mailto:joeclark@contenu.nu]
Sent: Thursday, 6 December 2001 9:25 AM
To: WAI-GL
Subject: Re: title inclusion


>I think that the title= attribute is quite important, but I also 
>think that there are some guidelines on how it should be used for 
>images and links.

And these are not really very good guidelines.

>I will try to outline the ways in which I use the alt and title 
>attributes. My reasons for using alt and title in this way are 
>chiefly because of the issues with screen readers and browsers 
>already discussed in previous emails to the list on this topic.

No, we need to write to spec. Faulty devices, whether they be 
Netscape 4 or some version of Jaws, must be fixed by the makers. For 
this same reason, I no longer recommend that people separate links by 
a printable character to make up for the deficiencies of old screen 
readers. I know perfectly well that not everyone can upgrade-- I ran 
entire Web sites on an old PowerMac 7100 for years. But at some point 
we have to accept that the benefits of standards-compliant handling 
outweigh all the effort expended to kludge around the bugs of 
noncompliant devices.


>For images that are not used as link element content: 1. Use alt to 
>describe the image.
>2. Use longdesc with a "d" link to provide a longer description if 
>required. 3. Generally do not use the title attribute for images.

On any given page, either all images have titles or none should. 
Consistency is important. (I give the same advice for links-- all 
have title or none do.)

In some cases, alt and title will be the same, which is not so bad. 
In other cases, titles can be different.

This, in effect, is my advice.

>For images that are used as part content for a link (e.g.. a link 
>that contains an image used as a bullet with text to describe the 
>link).

Very unusual!

And best repaired by using a stylesheet:

ul { list-style-image: url(image-URL); list-style-type: disc }

>1. Use the alt attribute to describe the image (e.g.. alt="bullet")

No, alt="*" should be the only alt text ever used for bullets. I 
suppose that title="" would be appropriate here. (Blank title is as 
valid as blank alt. Absent title is slightly less deplorable than 
absent alt.)

>  2. Use the title attribute for the link to describe the link. This 
>is often the same text as the non-graphical link element content. 
>Screen readers will then read the title, allowing the user to skip 
>over links without hearing "graphic, bullet..." all the time.

I cannot really figure out the HTML examples given. I think it has 
something to do with

<a title="LINKTITLE"><img title="IMAGETITLE"></a>

where the question becomes "Which title prevails? Inner or outer?" 
Ideally, we should have access to both, but the inner one is clearly 
the most specific and should prevail. I have a test file that I 
experimented with, and I found that generally the innermost title 
*does* prevail.

>For images that are used as the sole content for link elements: 1. 
>Generally use the alt attribute of the image to describe the 
>function of the link.

Always use an alt!

>2. Generally do not use the title attribute for the image 3. May use 
>the title attribute for the link, but generally would use the same 
>text as that used as the alt text for the image.

Aha!

>However, if a picture is being used as a link, and it is desirable 
>to describe the picture to the user through the alternative text, 
>appropriate alternative text that describes the image may not also 
>describe the function of the link. In these cases, I use the title 
>attribute in the link element to describe the function of the link, 
>which then allows me to use the alt attribute in the image element 
>to actually describe the image.

This is fair.

>For me, the use of title= for links is all dependent on looking at 
>the link element content and deciding whether or not that content 
>(text or image with alt=) is descriptive of the link and appropriate 
>as a link name or link content. If it is, then the title is not 
>required, if it is not, then the title= for the link is required.

This is also fair. I would nonetheless be consistent within a page, 
and duplicate alt text into title if necessary.

>For example, if the text that is used as link element content 
>cannot, for design reasons, be descriptive of the link, then you 
>have a case where you should use the title attribute to describe the 
>link. One example of this is the "d" link that is used with longdesc.

Can we kill off D-links forever, please? Like longdesc on the frame 
element, they never made sense. Description of a whole page or a 
frame is pretty much irrelevant to blind users. Devices must directly 
support longdesc anyway; I tire of hacking around incompatibilities.

>2. When title= is used, screen readers often ignore the contents of alt=.

Such screen readers need to be bitch-slapped until they shape up.
-- 
   Joe Clark | joeclark@joeclark.org | <http://joeclark.org/access/>
   Accessibility articles, resources, and critiques
Received on Wednesday, 5 December 2001 19:14:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:16 GMT