The first thing that I do when I get a script written by someone else is to test the supplied code on a number of different browsers just the same as I would if I had written the code. I usually do this because I am intending to rewrite the code to work the way that I want it to but I still suggest that you do the same because what is claimed and what is true are not necessarily the same thing as turned out to be the case in this instance. I deliberately inserted the code into a table because this was back when many of the CSS commands didn't work properly and so sites (including this one) used tables for layout. By doing this I was hoping to be able to find and fix the problem that stopped the scroller from scrolling in Mozilla.
The results from my tests were surprising given the claims on the site where I got the script from. As expected, the script worked properly in Internet Explorer and didn't work in Mozilla. The surprising thing was that the script also didn't scroll in Opera 5 and Netscape 6. In Opera 6 the scroller scrolled part way then stopped and started from the top again meaning that only half the message was visible. Netscape 4 produced a blurred mess of text and background colour across the screen even though the script contained code specific to the Netscape DOM. Opera 7 was scrolling (I think) but everything on the page was vibrating sideways so rapidly that nothing was readable on the page.
So much for the usability of that script. I had tested on seven browsers and the code only worked on one even though the documentation claimed that it would work on most of the browsers that I tried. Perhaps it does work on most of those browsers when you don't include the script inside a table but that wasn't what I was looking for so I didn't test it that way.
I would suggest that the only time that it is not necessary to test your script in at least these browsers is if you are writing for an intranet where you have control of the browsers that will be used to access the site. When ayou control the browser then you only need to include code that works with that browser.
This article written by Stephen Chapman, Felgall Pty Ltd.