W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > October to December 2003

Re: Breadcrumb dividers (semantics and accessibility)

From: David Woolley <david@djwhome.demon.co.uk>
Date: Wed, 19 Nov 2003 07:51:27 +0000 (GMT)
Message-Id: <200311190751.hAJ7pSj03313@djwhome.demon.co.uk>
To: w3c-wai-ig@w3.org

> Where is your evidence that the great 
> preponderance of people with relevant 
> disabilities are *not* using "recent desktop 
> graphical browsers"?

Basically from observing what non-disabled people in a relatively
affluent country do.

It actually surprised me how recent a browser you need to avoid this
problem.  I checked the browsers on my office system.  This is next
machine in line on the replacement cycle, but whilst not replacing
as often as financial institutions, we do replace more frequently
than some public service organisations.  The last time I checked,
my office was in South London, not Outer Mongolia.

The OS is NT 4 workstation, which is past end of life for Microsft (but
Microsoft has had to revise up lifetimes for current products because
the end of life it used for NT 4 doesn't reflect when people stop using
it in practice).  (NT 4 Server is still supported.) I think I've done a
browser update on this.  I've certainly added all service packs and hot
fixes except for the attempt to push IE 6 onto me as a critical update.
The primary browser is IE 5.5, which is still receiving hot fixes.
I have also updated at least some of the standard fonts.

Other developers are likely to have IE 6, but operations people,
depending on when their systems were updated, are likely to be on
the original browser from that time.

The test was to send the superset character as both &#x and &# forms.
Superset as it was one of the semantically valid characters assuming
that the detail was on the right (many suggest detail on left is
better, for left to right reading order).

IE 5.5 failed the test.  It displayed two missing glyph box

The other browsers I checked (these are from when I was involved with
code that generated HTML, and are mainly considered "don't care" by
the marketing people) were:

NS 4.75 failed - two question marks.

WebTVViewer 2.0 failed - uninterpreted entity for &#x and question mark
for the decimal one

Mozilla 0.9.3 (the Netscape 6 I have is older) displayed both correctly.

Amaya V5.0 worked (this is quite old).

Lynx 2.8.1rel1 made an attempt although the charecter wasn't particularly
recognizable.

Note that, as well as the mainstream browser failing:

- it is actually all the obscure browsers that got things right, even though
  they were quite old;
- the target market for Web TV is unlikely to upgrade for something like
  ten years (my home TV is about 17 years old) and is likely to include
  the elderly, who are more likely to be disabled.

NT 4 is Unicode aware.  I haven't tried these in the Windows 98 partition at
home; Windows 98 is not Unicode aware although is still supported by Microsoft;
that might mean that IE 5.5 tries to compensate more.

The Symbol font was not Unicode encoded and the superset character is not
in the updated core fonts, so the browsers that worked, could
only have been doing so by using a hack which meant it had specific
knowledge of Symbol (there were some residual Star Office fonts, so just
maybe these were properly encoded and found).  Explicitly hinting to
IE that it should use Symbol didn't work.  I'm assuming you are not
proposing the misrepresentation of a character below U+0100 and a 
specific font selection!

This is the test code:

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <title>d:\superset2.html</title>
</head>
  <body>
 &#8835; &#8835;
Received on Wednesday, 19 November 2003 03:15:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:13 GMT