- From: Eric A. Meyer <eric@meyerweb.com>
- Date: Wed, 18 Sep 2002 11:15:04 -0400
- To: public-evangelist <public-evangelist@w3.org>
At 7:28 -0400 9/18/02, Jonas Jørgensen wrote:
>Actually it's more like 96%, according to my server logs.
Other sites may tell a different story:
<http://www.hacksrus.com/reports/log.agents.html>.
>Many sites do something like
>
>if (document.all)
> doStuffSpecificallyTaileredToMsiesCssBugs;
>else
> doStuffWhichWillWorkInW3cCompliantBrowsers;
In the group to which I belong at Netscape, where we're involved
in standards support, we recommend that site authors who need object
detection do the following:
if (document.getElementById)
doStuffThatWillWorkInW3CCompliantBrowsers;
else if (document.all)
doStuffTailoredForIE4.x;
else if (document.layers)
doStuffTailoredForNN4.x;
else
warningMessageAboutLostFunctionality;
...or something like that. The point it that we tell them to detect
for the W3C DOM detection first, because more browsers support it.
Then the fallbacks come. This assumes that they feel a need to even
have the fallbacks, which most commercial site developers do. It's a
small performance win in standards-compliant browsers, too, since the
first test is to recognize them. You can shortcut the rest of the
testing branches that way. In some cases that can actually have a
perceivable effect on performance.
>If other browsers were to implement document.all, they would also
>have to emulate all of the weird bugs of IE5/Win. Given the enormous
>amount of bugs we're talking about here, that would be an impossible
>goal.
See for example the rather lengthy discussion contained in
<http://bugzilla.mozilla.org/show_bug.cgi?id=154589>.
>The only way out of this nightmare is W3C standards.
Indeed so.
--
Eric A. Meyer (eric@meyerweb.com) http://www.meyerweb.com/eric/
Author, "Cascading Style Sheets: The Definitive Guide,"
"Eric Meyer on CSS," "CSS 2.0 Programmer's Reference," and more
http://www.meyerweb.com/eric/books/
Received on Wednesday, 18 September 2002 11:15:55 UTC