[eropple]

"Progressiveness in design" only works when people are actually able to view it correctly. Given the fluctuating state of HTML5 at present, I'm reasonably sure you're just introducing additional complexity and points of failure unless you are teaching exhaustively only the exact features that work correctly, everywhere. This is not a particularly large set.

Now, I'd say something very different if HTML5 actually worked correctly everywhere, or even most places. But here's the thing--indoctrinating people into a JavaScript dependency like modernizr to fuel your desire to use a standard that isn't completely implemented anywhere is, frankly, terrible: a much more important lesson to drive home is "look, use progressive enhancement," not "look, HTML5 is awesome!", and making modernizr a dependency is the exact opposite of smart. Yes, your "web agency" might do this--but that doesn't make them very bright, either.

It's rule #1 of responsive web design: start with the simplest base case, and progressively enhance from there. If you want to really teach people how to benefit themselves by being able to write this stuff, I'd strongly recommend starting there--not with the high-end "look at my buzzwords and adopt them." Simple is good--really.

[jenius]

I still could not disagree more strongly. We're producing sites and apps for some of the biggest companies out there in html5. No joke, google.com, yahoo.com, and facebook.com (those were the first and only 3 I checked) are coded in html5. You are just behind the times... sorry! html5 is where it's at, and if you haven't adopted it yet, you are disagreeing with all of the top web coders in the industry. Start checking the doctypes and get with the program : )

I agree that simple is good, but you are just wrong on the html5... sorry : /

[eropple]

I am well with the program--I'm not the one advocating the use of modernizr or other hacks instead of building HTML that actually works on its own merits. The "HTML5" used by Facebook et al. is a very, very limited subset. The overwhelming majority of their work is valid HTML4! Your argument to authority is foolish.

HTML5 is a perfectly good tool, but relying on it and a JavaScript library to smooth it over because it doesn't fully exist yet is crazy. If you are not writing a web application--an application, not a "website"--and it breaks because someone has JavaScript off, you screwed up. I know that's a foreign concept to somebody inculcated into the bizarre little world of Ruby, but you're throwing simplicity out for buzzwords.

For better or for worse (and believe me, I am strongly of the opinion that it is worse), IE is not dead. These people you presume to teach are not 'opinionated' towards whether people using IE et al. should be marginalized and made second-class citizens. You, however, presume to make a decision for them that they have to make for themselves. And that's not right.

[jenius]

It's what happens with any teacher you get. Everyone has a way that they do things and their own opinions, and my opinion is use html5, always. Just so happens that this is also the opinion of the majority of web devs.

To be fair, you are right though - I should at very least educate people about this before teaching it. I will make sure to let my students know that if they would like to decrease the quality of their products based on 2% of people (that's data from 2007 as well, likely has gone significantly down) who browse with javascript off, then they should not use any html5 tags. Fair enough?