W3C home > Mailing lists > Public > html-tidy@w3.org > April to June 2000

Re: TIDY - Objects to euro symbol (?) & very pedestrian about ALT for gifs

From: <html-tidy@war-of-the-worlds.org>
Date: Sat, 29 Apr 2000 22:20:25 -0500
Message-Id: <p04310100b53153fa1612@[216.229.13.10]>
To: html-tidy@w3.org
WCD <WCD@iebsl.co.uk> wrote:

>IMAGE & ALT
>===========
>I can fully accept the rational of using ALT with images that are associated
>with hyperlinks or cause actions for forms, etc.
>However I and thousands of other HTML authors use single transparant pixel
>GIF images (typically named 1.gif) to control table layouts. The width and
>height attributes of the image are used to force a fixed amount of
>horizontal or vertical space when the overall table is working in
>percentages as most browsers will not correctly render a mix of percentages
>and pixel dimensions.
>
>Whilst I will grant you there may be better and perhaps more modern
>approaches to controlling table layouts there are literally millions of HTML
>pages that use this technique. It will be far too costly for everyone to be
>forced to edit them over such an innane issue. It seems pedestrian in the
>extreme that such invisible graphic images should be required to have an
>alt="something" attribute. HTML is verbose enough without adding this sort
>of overhead which will only add to the world-wide-wait.

It doesn't have to be alt="something".  It can be alt="" (nothing).

Tidy could possibly use a heuristic that keeps track of img srcs and
dimensions and note images where the same image is given, say, 3 or more
different sizes, and presuming they are resized transparent GIF references
and provide them with alt="" attributes automatically.  Especially if
height or width is 1.  I have an implementation of a compilation of such
data for a page, but alas that code isn't mine to share.

It is generally safe to assume that single-pixel height or width images are
either transparent or purely decorative and thus can be provided with
alt="" if an alt attribute is absent.  A boolean setting of
presume-resized-transparent=yes (or shorter setting) could theoretically be
used to control this presumption.

Anyone like this idea?
-- 
         ,=<#)-=#  <http://www.war-of-the-worlds.org/>
    ,_--//--_,
 _-~_-(####)-_~-_  "Did you see that Parkins boy's body in the tunnels?" "Just
(#>_--'~--~`--_<#)  the photos.  Worst thing I've ever seen; kid had no face."
Received on Saturday, 29 April 2000 23:20:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 April 2012 06:13:43 GMT