Wikipedia talk:Manual of Style/Accessibility/Archive 4

How I navigate pages with JAWS and my thoughts on navboxes
I can't find a good place to fit this reply, so I've created a new section for it. My screen reader JAWS uses quick navigation keys that can jump to the next HTML element of a specific type, get out of the current HTML element or go to the next different HTML element. Other modern Windows screen readers, including the free NonVisual Desktop Access, use a similar scheme. In JAWS, to get out of infoboxes, I press the ">" key to move to the end of the table, and to move to the TOC from the lead section, I press "H" to move to the next heading (as the TOC is a level 2 heading). My problem with navboxes is that there's no way to reliably get out of them by jumping past HTML elements. I usually use the "n" key (move to the next non-linked text block), but that usually gets me to the wrong place. Otherwise, it helps if navboxes are either at the end of an article or close to another HTML element (a heading, a table, etc). Here's a full list of JAWS keystrokes including Internet keystrokes. —Preceding unsigned comment added by Graham87 (talk • contribs) 14:38, 25 August 2008 (UTC)


 * And what about infoboxes? Do you usally skip them, or are a good summary of the article? &mdash;surue&ntilde;a 14:51, 25 August 2008 (UTC)


 * I think we're trying to solve a non-existent problem. Graham, could you take a look at Suicide note and tell me whether you can skip the first template there?  WhatamIdoing (talk) 21:44, 25 August 2008 (UTC)


 * I can get close to the end of it with one keystroke. Moving to the first non-linked text moves me to the last item which happens to be a self-link for "suicide note", and moving to the next div tag moves me to the text "This box:", which is close enough to the end of the navbox. I can't move past the navbox with the table navigation keys because it seems to be a one-column table which JAWS considers to be a layout table, and by default, JAWS does not count those. As for Suruena's question, I use infoboxes to get a piece of information quickly, but otherwise I just skip past them. Graham 87 06:09, 26 August 2008 (UTC)
 * All the navboxes I found today (well, yesterday) are fundamentally tables. Can you give me an example of a page that you can't get past (or at least very close to the end of) the navbox?  WhatamIdoing (talk) 06:31, 26 August 2008 (UTC)
 * No, not the ones I've checked through. It seems that hitting "z" to go to the next division gets me out of the navbox most of the time, and if not, hitting "n" to get to the next non-link text will also work. I didn't think of moving to the next division to get out of navboxes before ... I'll keep it in mind. Graham 87 11:16, 26 August 2008 (UTC)
 * So it shouldn't actually matter where the navboxes are placed then?
 * Placing a vertical navbox at the end of the lead is a layout problem. It looks a bit strange to have the navbox start at the top of the table of contents instead of at the top of the page (when there are no other images and no other infoboxes present), and a long navbox (longer than the table of contents) can interfere with the text of the next section.  If it really didn't cause any more trouble for you than the typical infobox, then I'd rather put the navboxes at the top of a section instead of the end.  WhatamIdoing (talk) 19:56, 26 August 2008 (UTC)
 * No it doesn't really matter where navboxes are placed. The trick for mooving past them I found isn't intuitive, but it works. Graham 87 00:49, 27 August 2008 (UTC)
 * Thanks for your response. Since it doesn't actually matter for accessibility, and it does matter for visual layout, then I'll change the guideline to list them after the infoboxes instead of after the lead's text.  WhatamIdoing (talk) 02:58, 30 August 2008 (UTC)

To make sure I follow. Vertical navboxes are now are OK after the infobox in the lead (although still preferred at the bottom of the article)? And if a vertical navbox is placed within another section, does it still need to be at the bottom of the section? Also, on updating this, the text is also at WP:LEAD and WP:LAYOUT, I think. Sandy Georgia (Talk) 03:05, 30 August 2008 (UTC)
 * Also, WhatamIdoing, I'm going to be traveling, so if this changes, if you don't mind, please make sure knows so he can add it to the August monthly updates for the WP:SIGNPOST.  Sandy Georgia  (Talk) 03:07, 30 August 2008 (UTC)
 * eeek, a third question. WhatamIdoing, you put navboxes *before* images; are you sure about that?  Sandy Georgia  (Talk) 03:09, 30 August 2008 (UTC)
 * Sandy, I'd rather have your questions now, no matter how many, than a mess later. My intention was to put navboxes after infoboxes, because when both are present, the infobox is more important by far -- being, as it were, an actual part of the article instead of merely part of the navigation infrastructure -- and should appear at the top of the article (instead of much further down the page or in a later section).
 * However, since the infobox is actually part of the article, then (per other comments on general principles on this page) we should probably not have first some article (infobox) — then some not-article (navbox) — and then go back for the rest of the article. But I'm (a) open to other arrangements and (b) no longer certain that it makes sense to have infoboxes listed separately from navboxes at all.  I invite other opinions.  WhatamIdoing (talk) 04:53, 30 August 2008 (UTC)
 * The problem is many articles have all three: an infobox (with an image), followed by a separate image, followed by yet another navbox.  We used to have them in that order.  I no longer know what matters; would need to hear from Graham, but I'll be traveling and will probably lose track of this.  And I can't think of a sample off the top of my head, but I've seen all three crammed into one lead.  Sandy Georgia  (Talk) 04:59, 30 August 2008 (UTC)
 * I don't thinkthis is as big a deal as spelling errors, broken tables, or things like that. I'd slighgtly prefer the image being before the navbox because a well-written image description will give important information for the article and I think the image is more important than a navbox. But this shouldn't be a hard and fast rule ... if it looks better for the image to be below the navbox, then put it there. Graham 87 10:27, 30 August 2008 (UTC)

Update, navboxes before text

 * OK, so the preferred (not required) order is images before navboxes, and navboxes are now before the text, not after. I've updated this page as well as WP:LEAD and WP:LAYOUT; if anything changes, please sync those pages as well.  Sandy Georgia  (Talk) 13:11, 30 August 2008 (UTC)

Sidebars
Sandy,

Can you give me an example of a horizontal navbox used in the lead? I've never encountered it, and I have looked at some thousands of articles. (Many of which, I admit, display properly standardized navboxes created by Arcadian.) WhatamIdoing (talk) 22:48, 25 August 2008 (UTC)


 * The AMX-30E article raised earlier in this discussion (the nominator has now moved that to the bottom, but lots of editors do what he had done.) I can't think of any others right now, but I come across exactly that a lot.  When the infofox isn't so long, or when there isn't an infobox, the horizontal nav works in the lead.  The problem at AMX-30E was that, because of the long infobox, moving the horizontal navbox to after the text left a huge white space.  Sandy Georgia  (Talk) 22:53, 25 August 2008 (UTC)
 * Ah, I see. So it's not so much that it's a horizontal navbox as it's a tiny navbox.  WhatamIdoing (talk) 00:41, 26 August 2008 (UTC)
 * By complete coincidence I saw the bit where it says navboxes in the lead and came to question it here. I've never come across this. Sandy, are there any current FAs where it can be seen? Matthewedwards (talk • contribs • email) 06:44, 26 August 2008 (UTC)

Tournament bracket templates
Please see 16TeamBracket-Compact-Tennis5 and other members of Category:Tournament bracket templates which contain tables which seem to me to be very inaccessible. Thank you. Andy Mabbett | Talk to Andy Mabbett 19:30, 27 August 2008 (UTC)


 * Like at 2005 U.S. Open - Men's Singles ... those tables are very hard for me to read, and I can only make some sense out of them because I know a bit about tennis and the fact that a set usually has six games. At the very least they need to have column and row headers. Am I right in my reading that Roger Federer beat David Nalbandian in straight sets 6-2, 6-4, 6-1, in the first quarter-final? I'm not even going to try to read the rest of the table. Graham 87 03:09, 28 August 2008 (UTC)
 * Yes, you are correct. It seems to me that it might be better to flip the chart so the winners in the final round would appear/be read first? Of course, that would be counterintuitive for the average English reader who is used to reading left to right... L'Aquatique [talk  ]  03:10, 29 August 2008 (UTC)


 * I'd expect the quarter-finals to be listed first in a chart like that. It's best to have them in chronological order. It's sort of interesting to find out if any of the top contenders were eliminated, or if there were any surprises. Graham 87 05:50, 29 August 2008 (UTC)


 * "best" in what way? It's not accessible to display them as the are at present. Andy Mabbett | Talk to Andy Mabbett 08:17, 29 August 2008 (UTC)


 * The current ordering is the best way because it's how people would expect to find the results. The way they are displayed, however, is hideous for accessibility. Maybe it'd be better to have one table for the quarter-finals, another table for the semi-finals and so on. Either that or the table rows and columns need captions, and there needs to be a uniform number of rows and columns. How I imagine it is a row containing something like "first quarter-final | Roger Federer v. David Nalbandian | 6-2, 6-1, 6-4|Roger Federer" with column headings of "match title | players | score | winner", where vertical bars separate the elements in my examples. Note that there is no problem with large tables if they're needed: this list of Sydney weather observations from the Bureau of Meteorology is completely accessible. Graham 87 03:47, 30 August 2008 (UTC)

Military history campaignboxes
The Military History Wikiproject has been using navigation campaignboxes placed under the infobox of articles, for a long time. They are often used singly, as in Battle of Moscow, and sometimes in multiples, for example, there are three in Operation Barbarossa.

Are these an accessibility problem? Can they be improved by providing an anchor and/or skip link in the box, or by including it within the infobox table? Or is there a problem simply with having navigation near the beginning of an article?

We're currently discussing refurbishing a body of navigation templates for tanks and armoured fighting vehicles at WT:AFV, and I'd like to incorporate input from here.

Thanks. —Michael Z. 2008-08-28 15:38 z 


 * They're not much of a problem, because as discussed above, they're easily skippable. They would e even easier to skip past if they were part of the infobox table. However the fact that they're collapsed by default makes them fairly easy to skip past with the arrow keys in screen readers that support these hide/show buttons (and most modern ones do). . Graham 87 01:49, 29 August 2008 (UTC)

Relevance (lists)
We state here a requirement that unordered lists (those beginning with a bullet instead of a number) not have blank lines between them. There are good reasons for this that have absolutely nothing to do with access for disabled readers, and it's mentioned in other guidelines.

But is this actually an issue of accessibility? Does a screen reader signal the beginning of a list? For example, does this:


 * First item
 * Second item

get read differently from this:


 * First item


 * Second item

If there's no actual difference for disabled users, this should not be mentioned in this guideline. WhatamIdoing (talk) 22:47, 31 August 2008 (UTC)
 * Agree. - Dan Dank55 (send/receive) 02:25, 1 September 2008 (UTC)


 * Yes, by default, JAWS and other screen readers signal the beginning and end of each list with "list of x items" and "list end". Therefore the lists in this version of Antonio Vivaldi sound horrible. Graham 87 02:33, 1 September 2008 (UTC)

Thanks, Graham. Then we should definitely keep this. However, the current text is not IMO ideal. It says, ''Do not separate list items by more than one line-break. If list items are separated by more than one line break, the HTML list tags will be ended''.

This is presumably just fine if you know exactly what a line-break is and believe that invisible HTML list tags are important. However, "no more than one line-break" may be misinterpreted as "no more than one blank line" by some readers, and the relevance of HTML list tags is non-obvious.

Could we rephrase this to something like what WP:MOS says, which is "Do not leave blank lines between items in a bulleted or numbered list unless there is a reason to do so, since this causes the Wiki software to interpret each list item as an individual list." -- except perhaps adding that it causes screen readers to announce each item as belonging to a separate, one-item list? WhatamIdoing (talk) 06:22, 1 September 2008 (UTC)


 * Agreed with all of that. - Dan Dank55 (send/receive) 13:30, 1 September 2008 (UTC)
 * P.S. But let's make it clear we're talking about article-space. On talk pages, it's common for people to leave a blank line between someone else's bullet point and their bullet point to enhance readability.  If screen readers do a bad job reading this, then we should complain to the people who make the screen readers. - Dan Dank55 (send/receive) 14:04, 1 September 2008 (UTC)
 * It's not the case that screen readers do a bad job of reading such things; they read what they're given, and if we give them garbage, that's what they'll read. Why should talk pages be any less accessible to people who cannot see, than article pages? In fact, I see that the use of colon characters to indent discussions like this one actually generates definition list mark-up. I wonder if that's a good habit to perpetuate? Andy Mabbett (User:Pigsonthewing); Talk to Andy Mabbett; Andy Mabbett's contributions 11:41, 6 September 2008 (UTC)


 * Sounds good. I didn't realise that separating lists was mentioned in the Manual of Style when I wrote that section, but it needs to be mentioned in the accessibility guidelines as well. Graham 87 14:05, 1 September 2008 (UTC)

Note: &lt;tt&gt; is not &lt;code&gt;
Someone or other seems to be confused about the meanings of some &#91;X&#93;HTML entities. To give examples of source code snippets (like wiki markup, template code, etc.), use. The  element is nothing but a font-styling tag and has no semantic function at all. I just checked the current version of the W3C HTML specs, and for some reason it has not been deprecated according to the Index of Elements, but this is surely either a typographical error or other oversight, as all other purely-presentational markup has been deprecated in favor of CSS since 1998.

A lot of other semantic (or "logical" as someone wrote in the guideline page – where'd that come from?) elements get confused with, as well, including  ,   and   (documentation). These all have distinguishable semantic purposes.

Presently, MediaWiki (or at least Wikipedia's installation of it) actually does not accept  or , only   and   (and  , which we should ignore as being no different from deprecated obsolete junk like  ,  , etc.). This should probably be undone by the developers, as there are often good semantic reasons to use these elements, especially in technical articles and in template documentation. Something for WP:VPT. I have proposed this already, actually.

That fact that most browsers (we cannot be certain that it is all of them, since I doubt anyone, including me, has possibly ever used every browser in existence, and I'm a professional web developer) display  and   visually as monospaced, non-proportional text, at the same size and otherwise styled identically, is irrelevant. They have different meanings (or rather one has a meaning and the other does not), and it is likely that some screen readers and other software (XML applications, and such) are already distinguishing between them, and have for some time. Inline microformats like hCard also sometimes depend heavily on the semantics of HTML elements not being violated. (I doubt that hCard in particular makes any use of these specific elements, but that's beside the point; who knows how many microformats their are or will be?)

If for some reason monospaced type is needed but it is not code, it is really more appropriate to use (a template for  ) instead of the   version, just to better keep the content and the presentation separate.

PS: We should not use quotation marks with, etc. (other than for purposes of actual quotations, of course), since "scare-quoting" them is simply redundant; the markup itself already sets them apart as non-regular segment of the running prose. I.e., write "use &lt;code&gt;reason=&lt;code&gt; to provide a reason for applying this template", not "use '&lt;code&gt;reason=&lt;code&gt;' to provide a reason for applying this template" (and note that "use &lt;code&gt;'reason='&lt;code&gt; to provide a reason for applying this template" would be severely wrong unless the single-quotes were part of the actual literal string required in the source code!).

—  SMcCandlish  &#91;talk&#93; &#91;cont&#93; ‹(-¿-)› 23:25, 31 August 2008 (UTC)


 * I've been guilty of misusing ; and you're right, so I'll try not to do so again. But microformats do not make use of HTML elements, other than generically, with a few specific exceptions  and none do or are likely to make use of   or   other than as interchangeable elements. Andy Mabbett | Talk to Andy Mabbett 00:28, 1 September 2008 (UTC)

Moving images
I'm no expert, but I think that moving images like a recent "picture of the day": Image:Tablet press animation.gif and especially flashing images like Image:Strobe.gif could cause problems for people with some photo-sensitive conditions like epilepsy - I don't have that, but I find them distracting. My preference would be to use a static version, and link to the moving image, with a caution in the link in place. Andy Mabbett | Talk to Andy Mabbett 00:54, 1 September 2008 (UTC)
 * I'd have no objection to tagging moving images as moving and letting users set a preference that they only want to see the first frames of animation; that might have value to people sensitive in one way or another, and it would also help people who prefer to avoid animation because of bandwidth. (I don't keep up with image issues; maybe this is possible already?)  Tagging would be much preferable to discarding; animations are very important in sci/tech articles. - Dan Dank55 (send/receive) 02:24, 1 September 2008 (UTC)


 * Tagging would be a good idea ... complex animations make JAWS unresponsive. I have images turned off in my web browser to save bandwidth, but I realise that isn't an option for everyone using shared computers. I always thought that only still images were displayed on the Main Page with a link to the animation, but that might have changed. Graham 87 03:06, 1 September 2008 (UTC)
 * Notes: This was last discussed at length at Wikipedia:Village pump (proposals)/Archive#Avoid movement in pages (May 2007 archived diff). I won't attempt to summarize it all (though I do apologize to Andy for my tone and goodfaith in some of the comments there), but there are numerous pros and cons to both displaying or not-displaying animations directly within articles. And numerous editors/admins supporting each position, hence no consensus for a change seems to have been reached there.
 * The relevant policy section is at Image use policy, which has stated since late 2005 that "Inline animations should be used sparingly; a static image with a link to the animation is preferred unless the animation has a very small file size." (Though I haven't seen any articles that actually use this idea...)
 * Discussing this at Village pump (technical) and/or Wikipedia talk:Image use policy might be preferable(?), but it'd be good to have accessibility experts in on the discussion wherever it is. Hope that helps. -- Quiddity (talk) 19:48, 1 September 2008 (UTC)
 * It helps quite a lot. I didn't know that policy prefers static images with links to animations in most cases; I'll mention that in my reviews when it comes up. - Dan Dank55 (send/receive) 03:54, 2 September 2008 (UTC)
 * Or, Tell us if you ever see an article that is already using the static-link method! (I've never seen one.) Perhaps the policy should rather be changed to reflect actual practices/usage...
 * Personally, I strongly support including the majority of animations directly within the article. The animation at Engine, or the three used in Horse gait, for example, are immensely informative/educational, and I'd be sorry to see them get removed by one level - where many readers will never discover them. -- Quiddity (talk) 03:51, 6 September 2008 (UTC)

Scope
A lot of people browse Wikipedia on cellphones and PDAs these days. Some browse on Kindle, which has gray-scale only, no color. I imagine some people like to listen to articles from Wikipedia while they're driving. People who use small screens and screen readers would agree with the main thrust of this page, which seems to me to be, put things where people are expecting them to be, to aid navigation. Should we mention that the issues on this page have wide applicability? I think some people think of this page as having more limited scope. - Dan Dank55 (send/receive) 02:41, 1 September 2008 (UTC)


 * Agreed. The more accessible a page is to "the disabled" (&trade;), the more accessible it is to everyone, not least the people you list - and search engines. Andy Mabbett | Talk to Andy Mabbett 08:22, 1 September 2008 (UTC)


 * And a large number of people over 40 have a larger default font size on their computers. - Dan Dank55 (send/receive) 13:58, 1 September 2008 (UTC)

Template: Geographic Location
How does Geographic Location render in assistive browsers? Could it be improved? (For an example, see Bergman, Edmonton). Andy Mabbett | Talk to Andy Mabbett 19:01, 1 September 2008 (UTC)


 * See also GeoCompass (used on Chelmsford); and Compass-table (used on Marylebone). It is proposed that these two should be merged into the above. To my mind, the latter looks the most accessible, in that it includes compass-directions as text, and linearises properly. Andy Mabbett | Talk to Andy Mabbett 19:12, 1 September 2008 (UTC)


 * Geographic Location doesn't render well with JAWS. The images with no captions sound like "link graphic Compass_rose_pale.svg/50px-Compass_rose_pale.svg", the links for viewing, discussing, and editing the template get in the way, and it's hard to tell what direction you're reading about. GeoCompass reads OK linearly but it is still hard to figure out what direction everything is. Compass-table is the most accessible, but the table would be easier to use with row and column titles. When I navigate from east to southeast, I also hear the "southwest" part because JAWS thinks that's the title of the row. Speaking of directional templates, the Infobox Australian Place shows the suburbs or towns surrounding a certain place, but it's hard to tell the directions of the towns with a screen reader. An example is at Nollamara, Western Australia. Graham 87 01:18, 2 September 2008 (UTC)

Use of colour to convey information
WCAG 1.0 says:

2.1 Ensure that all information conveyed with color is also available without color, for example from context or markup. - http://www.w3.org/TR/WCAG10/wai-pageauth.html#tech-color-convey

We should therefore discourage the use of colour (or shading-pattern) alone to convey information, as seen in List of Oasis band members. Andy Mabbett (aka Pigsonthewing); Talk to Andy Mabbett; Andy Mabbett's contributions 21:48, 2 September 2008 (UTC)