opinions about when null ALT text is appropriate

[brothercake]

This is maybe a small thing ... but it's been doing my head in for ages ...


The spec clearly allows for images which are irrelevant, like spacer gifs or trivial decoration, having null alt text

By a similar token, for images where whatever would be appropriate as ALT text is written next to the image anyway, like a photo caption, then it's pointless to add that ALT text to the image - all a blind user gets is the same phrase read twice. In such cases, a null value can also be used.

But I have what seems to me a grey area - check out this news page (http://www.mori.com/index-news.phtml)

I reckon those blue images down the middle column shouldn't have ALT text, because:

1 - they just decorate the headline, and although the imagery is relevant, the choice if image is still essentially arbitrary

2 - a description of what's in the picture might be counter-productive; it might detract attention from the headline and summary text

3 - any ALT text which served as a caption would be duplication, since the headline is the logical choice of caption text

But I just can't decide; is it better to provide exact equivalance and describe the image in ALT text, or am I right in thinking that the news item as a whole is more readily understandable if the picture is not described

[Jerome]

Just my opinion,

1. I don't like the alt flags

2. Alt flags are necessary when slow loading images/sites are used (to determine what is in the square or where it links to) like an intro page-image of >50k

Your site loads quick (36k modem!) and when loaded the layout is almost complete. Before people have time to see everything the images are loaded as well.

I should not use them,
Jerome

[Jerome]

Btw, well constructed, nice layout,
Jerome

[brothercake]

Thanks :)

NB - by alt 'flags' do you mean the tooltips that appear when you mouseover an image, that contain the ALT text? Because that is *incorrect behavior*, a *direct contradiction* of the spec by MS - alt text should not be displayed as a tooltip, because the purpose of alt is to *replace* the image when the image is *not displayed*.

All images *must* have ALT text; it's not optional; but you can override IE's stupidness like this:

<img src="me.jpg" alt="A photograph of me" title="" />

And of course in some cases (like I mentioned) null alt text is allowed:

<img src="spacer.gif" width="100" height="1" alt="" />

[Catman]

I'd err on the side of including alt description for the images -- in each case, it does not seem to me that the description would duplicate the headline. Moreover, since the image is a link, you need something to indicate what the image links to.

To make it stand out from the regular text, you might consider doing somthing like this: alt="Image Description : A blue-tinted photograph of London Bridge linking to the article 'Current Affairs, Londers Views': End Description"

[brothercake]

Originally posted by Catman
To make it stand out from the regular text, you might consider doing somthing like this: alt="Image Description : A blue-tinted photograph of London Bridge linking to the article 'Current Affairs, Londers Views': End Description"
Hey I like that - I hadn't considered such a complete explanation. I don't think it's necessary to say that it's an image, because afaik screenreaders do that anyway, but otherwise thanks; that's the perfect solution :thumbsup:

[meow]

Uhm, but that would be confusing. alt shouldn't describe the picture but rather it's function. You would have to more or less duplicate the content of your text links, I think.

[joh6nn]

i agree with you, that there should be null alt tags there. a description of the picture would serve no purpose.

by the way, what does everyone else use, to evaluate usability like this? i have a freeware program called speakonia, that reads webpages aloud. what do the rest of you use?

[brothercake]

Originally posted by joh6nn
what does everyone else use, to evaluate usability like this?
I've got iCab for OSX which has a screenreader built in. I was gonna get JAWS ... but it costs $1000 ...

I don't know about this ... I was under the impression that ALT text should not describe the function of the image, it should desccribe what the image is; function is what the TITLE attribute is for.

But my confusion arose because what the image looks like is essentially irrelevant to the story. But does it add value to the page content nonetheless, enough that it deserves describing?

Dang this is so confusing :eek:

Hang on .. let's see what the WAI says ....

[brothercake]

When a text equivalent is presented to the user, it fulfills essentially the same function (to the extent possible) as the original content. For simple content, a text equivalent may need only describe the function or purpose of content. For complex content (charts, graphs, etc.), the text equivalent may be longer and include descriptive information.

Text equivalents must be provided for logos, photos, submit buttons, applets, bullets in lists, ASCII art, and all of the links within an image map as well as invisible images used to lay out a page.

Quicktest! A good test to determine if a text equivalent is useful is to imagine reading the document aloud over the telephone. What would you say upon encountering this image to make the page comprehensible to the listener?


Hmm .. well that makes sense; if I described that image over the phone, well then yeah, I'd describe it as a blue tinted photo of the tower of london.

But it does specifically say "fulfills essentially the same function" and the function of the image is, well, nothing excepts it's a link which is right next to another, identical link.

[meow]

Describing the picture makes sense only if he image IS the content, I think. Like if you have pictures of art. Then you would maybe want to have Artist + Name of the painting in alt and a full description in londesc or maybe a short description in alt.

I looked around the JAWS site and maybe they have taken it away, but there used to be a demo for developers that ran something like 30 minutes and then you had to reboot to start it again. I never tried it though because I've read many warnings about how it can mess your system up. It obviously needs to make a lot of changes.

[liorean]

The alt text should be equivalent in function to the image for a non-image user agent. Spacers or eye-candy should be empty. Navigation should be clearly written, preferably short. Pictures should be explained for their content, in a way that seems to make sense in context in a non-image browser. Maps, schedules, hierarchy trees or other graphically descriptive material that simply can't be explained by short text entries should be referenced in a way that makes sense, and should also be linked to.

Never should you use alt for tooltips - that's what title is for. An alt attribute that doesn't make sense in non-image browsers is worse than an alt attribute left out or empty when it shouldn't be.

I recommend using links or lynx for testing whether your pages makes sense with alt displayed instead of images. You can also try using a user stylesheet in moz/opera, or behaviors in ie, to replace images with their alt attribute. You can even try one of the bookmarklets that replaces images with their alt tags.

[brothercake]

Originally posted by liorean
I recommend using links or lynx for testing whether your pages makes sense with alt displayed instead of images.
Indeed; but while Lynx is a useful analog for screenreading, it's not quite the same - with Lynx you still have random access to the content, courtesy of your eyes; with a screenreader you must sit there and listen to the whole thing linearly.

The judgement call here is whether a user who doesn't get that image or equivalent is actually missing anything - whether those pictures can be described as "pointless eye candy" - and more specifically whether the overall comprehension of what the screenreader says is increased or decreased by the image description.

Thanks for all these responses - I'm pretty sure now that my original assessment was right, ie, null alt text :)

btw - yeah JAWS does do funny things to your video settings ... I wouldn't recommend trying to install it on your main machine, but it'd be allright on a second machine or in a VM. Remember that JAWS is designed for people who need it, rather than for developers, so I don't think its fair to describe this as a problem with the software.

I know a couple of others that might be worth checking out

Windows Eyes - http://www.gwmicro.com/

IBM Home Page Reader - http://www-3.ibm.com/able/hpr.html

[Catman]

I see I'm in the minority, but I can't resist adding in this thought: If the image doesn't warrant alt text and merely duplicates a link, should it be there in the first place?

I'd also suggest that when thinking about this issue, the idea of functionality should include not merely mechanical operations (e.g., it's a link) but also rhetorical effects. Granted, the images on the page in question do not convey information in the same sense as a diagram of a heart or a bar chart, but there is certainly something added by the image used: Why London Bridge rather than the Tower of London or Big Ben? Why a jet rather than a tank? Why a lab technican injecting a mouse rather than an animal alone in a cage?

Just some food for thought.

[brothercake]

Originally posted by Catman
Why London Bridge rather than the Tower of London or Big Ben? Why a jet rather than a tank? Why a lab technican injecting a mouse rather than an animal alone in a cage?
Specifically in this case, no reason - the images are arbitrary; any of those suggestions would be equally effective.

As to why it's there - equally, no real reason, except it looks nice; and looks, surely, are irrelevant to a blind user.

Or are they? Not being blind, I don't know, but a blind person may still have a concept of "blue" even if they've never seen blue, by association with other things they've been told in the past are blue?

[Borgtex]

That will sound horrible, I know, but with the small percentage of people not being able to see the images, does it pays to waste time adding a comment for every image and the translations in multilanguage sites? For image buttons it's ok, however, as it can benefit both "blind" & "no blind" users

[joh6nn]

i've been using this: Speakonia (http://www.cfs-technologies.com/home/?id=1.4) it's free, and easy to use.

i think we've pretty much covered the question of whether or not these images need alt text. now we seem to be into whether they need longdesc text.

[brothercake]

Originally posted by joh6nn
i think we've pretty much covered the question of whether or not these images need alt text.
We've certainly completed the debate, although I wouldn't say we have a definitive answer ;)

I guess, as with so many things, there no definitive answers; it's very much a judgement call.

[brothercake]

Originally posted by Borgtex
does it pays to waste time adding a comment for every image and the translations in multilanguage sites?
If the images need it, yes it does. IMO the fundamental need is not in question; the questions surround individual cases - the value of each description to the users who 'see' it.