The Browser Test Limbo: How low can you go?

I test in old browsers. Browsers so old there’s no expectation I’ll ever receive a visit from one.

At present, here’s a sampling of popular browser statistics:

IE 6 65.5%
IE 5 4.4%
O 7 1.9%
Ffox 19.2%
Moz 4.0%
NN 4 0.3%
NN 7 1.1%

(These stats were leeched from w3schools.com, since they get more traffic than I do, I’m sure)

IE4 and IE3 don’t even make the list. So why test in them? Easy: I’m obsessive/compulsive. And curious.

While I don’t care whatsoever if IE3/4/5 don’t get the visual goodies on a cool, simple, clean, wysiwyg editor, I do care if I’m writing sloppy JavaScript code. So I reported an error on that page that occurs in IE4, just to be thorough.

If it’s an easy fix, great: The Man In Blue can claim it’s backwards compatible all the way back through IE3.
If it’s not an easy fix, who cares? Nobody uses that browser anyway.

But still, I’d want to know about the error. Unlikely, it’s still possible some legacy software somewhere has an old imbedded component in it that would break on my code. If I can avoid that with minimal care, I’ll do it.

So to those who give crap to testing in old browsers: Go right ahead. I don’t mind.
But as I see it, professionalism incorporates at least looking at the extra mile, if not going there.