Litmus

I know that the company I do my day job for, chose only to support Internet Explorer 6 for the site.

This, was a big mistake, and resulted in the designers being instructed to add Firefox and Safari straight away.

Again, this all came down to looking at the logs, something which should be done before, rather than after the fact.

That's a good point actually - if you're launching a brand new site you've got no logs to look at. I'd just take a best guess, and probably try and support the list of browsers I gave above. Supporting IE only though is not a good move, as those guys found out!

Well, I guess first browser would be IE then Firefox then Opera and then Safari. As for versions, well, I guess IE 5.5 is the minimum for me.

There are people trying very hard to switch people to Firefox. So much so that they go to extents of telling people if they layout looks wacked then it's cause they're viewing on Firefox.

Oh well, to his their own. :) Cheers.

At work, we have a policy that states IE5+, Opera 6+, Safari, Konquorer (3.3+, I think) and Netscape 6+. This - thankfully - excludes the 4.0 browsers.

For my personal projects I aim a bit higher: IE6, Firefox, Opera and Safari. IE5.5 often works as a bonus. I'm also quite happy to use CSS2.1 unsupported in IE where the effect degrades.

It would be nice to see Opera 6+ made testable in the near future, it's the one browser that's notably absent.

I've been trying to make sites work in IE 5+ (Mac & Win), Firefox, Safari, Netscape 6+, and Opera. I generally don't do too many fancy tricks, so cross-browser support isn't too hard. It's mostly IE box model stuff that needs fixing.

I'd like to second Ben's comment about adding Opera testing. Otherwise I'm really happy with SiteVista so far. Nice work.

Thanks for your input guys. Opera is indeed on the way. It was included very early on but we had some reliability issues with it (as well as Netscape 8, which has now been added). As soon as I get it testing more reliably it'll be available again. We're also working presently to add support for Safari 2.0 in addition to Safari 1.3 which we've got already. David's just bought a new Mac to add to the collection and is in the process of installing Tiger.

Are there any plans for a 'MailVista'? I've been looking all over for a solid way to test the rendering of e-mail newsletters (inline styles vs. font tags, etc...).

Alex, yes that's something we'd like to do in the future. We've done a little tinkering with it, but it's still very much in the works.

right on - i'll be all over that!

I design in Firefox, debug for Internet Exploder and take a quick peek (but only sometimes debug) in Opera. I don't have access to Safari etc

12-mth stats (tech site): IE 63.7%, Moz 27%, NS 5.8%, Opera 3.1%

> Therefore, my best advice would be to examine your site's logs, and make sure you're supporting the browsers of at least the top 90% of your visitors.

Terrible idea.

If somebody's website doesn't work in Firefox, then they are going to look at their site's logs and find what exactly? That's right, virtually no Firefox users. Who'd be using a website that doesn't work in their browser?

All your suggestion does is recommend that people support what they already support - it's a complete no-op.

That's not even addressing the fact that site logs are an awful, awful way to measure browser usage - it's about as accurate as guessing. Bad statistics are worse than no statistics.

That's a good point Jim. If your site doesn't work in a particular browser, you're sure not going to have many of those people stick around. They will, however, still visit your site in the first place (they'll just leave quickly). Perhaps you could filter your server logs so you're only viewing the browser usage of people who visited the home page. That would give a better sample.

Beside from that issue, why are site logs a bad way to measure browser usage? And what other options do we have?