Looking for Statistics on Users Browers

[BFTrick]

Hello there,

I am trying to convince my boss to use HTML5 and one of his main gripes about using HTML5 is that without the HTML5 shiv (a JS file) that the entire site looks terrible in IE. So I am looking for some sort of data that show the percentage of people that meet both criteria (1 - have IE, and 2 - have JavaScript disabled).

Any idea where I can find that magical number?

I assume it is small number, but having something concrete would help.

[svidgen]

Hello there,

I am trying to convince my boss to use HTML5 and one of his main gripes about using HTML5 is that without the HTML5 shiv (a JS file) that the entire site looks terrible in IE. So I am looking for some sort of data that show the percentage of people that meet both criteria (1 - have IE, and 2 - have JavaScript disabled).

Any idea where I can find that magical number?

I assume it is small number, but having something concrete would help.

Best to get it from your own web stats. I think AWStats has the ability to report that number. Otherwise you can get the number from w3schools, but the stats will be reflective of their visitors -- not yours. The same will be true of any site you pull the stats from. And, while you may be able to find a site that rolls up the stats from multiple sites, you're still dealing with a small subset of sites, and it's not going to be representative of your users.

Rambling about stats aside ... Could you explain the core issue a little more? Are you just talking about whether to use the HTML doctype? Or an HTML5 feature (or tag) or set of features (or tags) in particular?

Reason I ask: You ought to be able to start using the HTML5 doctype without issue right now, if you wish. A user's JavaScript capabilities will be an issue regardless of your doctype -- you can equally well throw scripts on an HTML 4.01 Transitional page that fail to operate for the subset of users with JavaScript disabled, and it will be the same subset of users who experience issues with an "HTML5" page.

I guess ... regardless of what your answers are, I would say this. Check your browser stats using AWStats, Google Analytics or some other stats app to determine the oldest browser using your site with a share greater than P%. (say .. 5%) Design your pages to be compatible with THAT browser. Add <script>'s to enhance the old-browser compatible page. The script can test the browser to see whether it supports enhancement X, Y, and Z and apply the enhancements if it does.

Know what I mean?

Design for a crippled browser. Use scripts to "upgrade" it dynamically, when possible.

[NogDog]

I don't know what the numbers are or where to get them, but frankly, anyone surfing the web these days without JavaScript enabled is probably used to all sorts of visual and functional problems when viewing various web sites, as so many depend on it any more (without "graceful" fall-back mechanisms).

[Sunny G]

This is exactly why I gave up webdevelopment. It's not healthy.

[BFTrick]

Could you explain the core issue a little more? Are you just talking about whether to use the HTML doctype? Or an HTML5 feature (or tag) or set of features (or tags) in particular?

I wish to use the semantic tags introduced in HTML5 instead of using div tags. They include { header, footer, nav, aside, etc. }. Now as of right now the good browsers can style them just fine; IE<=8 cannot style unknown elements so a JS file (known as the HTML5 shiv) works around this and allows these tags to work in IE. So if the user is using IE and they have their JS off then these semantic blocks and all of the nested tags will not have any styling applied to them.

An important point against what I am trying to do is that currently the semantic tags do nothing. But I think it is prudent to build websites using the semantic tags so that when Google starts making use of the semantic tags that all of my websites are already SEO friendly and I won't have to re-code anything.

[svidgen]

I think your safest bet, presuming you don't want to use any conditional server-side logic, is to simply not style semantic elements:

Don't do this:

<style type='text/css'>
header {
  font-size: x-large;
  background-color: yellow;
}
</style>

<header>I'm amazing!</header>

Do this:

<style type='text/css'>
.header {
  font-size: x-large;
  background-color: yellow;
}
</style>

<div class='header'><header>I'm amazing!</header></div>

The semantic tags, to the extent of my testing, are treated like SPAN's in IE and FF. Add the following style to make browsers that treat them as blocks (Chrome) to follow suit:

header, footer, nav, <etc..> {
  display: inline;
}

I've done little testing with this though -- just an initial proof of concept locally. I would be interested to read/see the results of further testing. I'd also be interested to know whether any search engines are actually weighting these elements in any significant way.

[Charles]

I don't know what the numbers are or where to get them, but frankly, anyone surfing the web these days without JavaScript enabled is probably used to all sorts of visual and functional problems when viewing various web sites, as so many depend on it any more (without "graceful" fall-back mechanisms).

When we hit a site that requires JavaScript we just shake our head in disbelief and then never visit that site again.

For the most part the experience is quite liberating and you do see interesting things. Look at this page both with and without JavaScript: http://www.aikido-shobukan.org/ .

[svidgen]

When we hit a site that requires JavaScript we just shake our head in disbelief and then never visit that site again.

Who's we?

... along this line though, "web sites" should not use javascript. "web applications" should. In other words, your blog should not depend on JavaScript. Your favorite news site should not depend on JavaScript. But, while not preferred, it's OK if your web-based email client uses, or even depends on, JavaScript.