W3C home > Mailing lists > Public > www-html@w3.org > January 1999

Re: an official W3C browser test

From: Gordon Worley <redbird@orlando.crosswinds.net>
Date: Sat, 9 Jan 1999 17:08:37 -0500 (EST)
Message-Id: <l03130300a9f099829e6b@[10.50.227.4]>
To: www-html <www-html@w3.org>
>I don't think programmers will cheat so they pass the test.

	I agree with you.  I simply want to make sure that we have covered
all possible events that may occur.  Realizing that the bad press will hurt
the browser makers business, I doubt that any of them would want to cheat.
Still, it's good to cover all the bases.

>BUT when the w3c issues an official browser test,
>programmers will test their beta-browser on it,
>and tweaking the code until it passes the test.
>Programmers may incorrectly think that passing
>the test means a 100% working browser.
>
>This means that issuing an official browser test
>may even ENCOURAGE BUGGY CODE.

	What I think your talking about is cheating by accident.  The
browser passes the test suit, but only works on a limited basis in the real
world.  These types of bugs already exist in most browsers, and are refined
through the sub-version releases that come out about once every two months
or so.  Therefore, the test needs to consist of a mini-Web site (or
something of this like) with several pages that test different technologies
in different situations.  While it will be impossible to cover every
combination, the widest range possible should be.  Not knowing exactly how
the test suits work, this idea could probably be better refined by the W3C
or someone else on the list.

________________________________________
Red Bird Island Productions
Gordon Worley
http://www.crosswinds.net/orlando/~redbird/
mailto:redbird@orlando.crosswinds.net
Received on Saturday, 9 January 1999 17:06:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 March 2012 18:15:38 GMT