Ask HN: Go HTML5 only for a new product?

With options like <a href="http://www.modernizr.com/" rel="nofollow">http://www.modernizr.com/</a> and a handful of other fallback solutions for IE, i think it is ok to go ahead with HTML5/CSS3. The part to focus is not to ignore those who use IE. It is fine if the appearance is not as classy as it is in modern browsers but compromising on the basic functionality is definitely not a wise option.

Are you willing to ignore ~50% of the potential (mass) market?<p><a href="http://en.wikipedia.org/wiki/Usage_share_of_web_browsers" rel="nofollow">http://en.wikipedia.org/wiki/Usage_share_of_web_browsers</a><p>And do you guys really think it's gonna change that much in 1 year?<p><a href="http://gs.statcounter.com/" rel="nofollow">http://gs.statcounter.com/</a><p>In the end it's about what your clients will use. If you're building something for developers, sure, ignore IE. If you're building something to be used by dumb (computer) terminals at big corporations or the public at large, then it'd be a risky move.

First of all, the HTML5 and CSS3 standards are not finished yet. It means semantic values of tags, attributes, and even tag names, as well as their functionalities, can change at any time, and other can appear. Thus, it seems difficult to build an HTML5-only website for the general public.<p>Moreover, if this website is designed for a large company, there are chances that its user base is still using IE (even IE6 maybe). In this case, HTML5 and CSS3 are not welcome, even for IE8.<p>I would recommend to wait for IE9 release and the standards to be finished. Maybe you can add extra content for people using modern web browsers, like the &#60;video&#62; tag, but I think IE's user base is too large to be ignored.

Yes it is ok I think, if it really gives you an advantage in doing stuff faster and better (ie. if your app needs it).<p>You can aim at early adopters the first two years, and by then, this stuff will be mainstream. 37sigs took pretty much the same approach.

The HTML5 Shiv (<a href="http://code.google.com/p/html5shiv/" rel="nofollow">http://code.google.com/p/html5shiv/</a>) will convince IE &#62;=6 to handle HTML5 elements just fine.

I would say develop for IE7 &#38; Firefox, making sure stuff doesn't break (is visible and somehow workable) in IE6 and Safari. That will cover 99.99 of your future customers.<p>I guess you are in danger of having more headaches in the future with bleeding-edge raw, not well supported things...

Definitely, it's very fine to use HTML5/CSS3. However, I think that it is still a good idea to make it fall back nicely. Stuff like rounded corners aren't too tragic to lose when viewing something from IE6-8, but I think it's in your best interest to make sure that IE7 and 8 don't look dreadful. Reward people for using better browsers, but don't punish them for using bad ones either.<p>Completely ignore IE6, don't even worry about how it looks. Most startups nowadays buy a computer and at least IE7 is on it. 2011 should be a big year for HTML5, so aim towards the "future".<p>Don't worry about IE&#60;9 too much, but certainly make sure it falls back gracefully is what I'm saying. It's okay if it's not as pretty, as long as the basics work.

I think lots of enterprises still have IE only policy and we certainly see that with Visual Website Optimizer (whose complex visual designer works even on IE 7). I would certainly not risk discounting IE 6-8.

I wouldn't ignore IE when targeting fortune 5 million. Check out html5boilerplate.com and css3pie.com to help achieve cross browser and backwards compatibility. Good luck with your app!

My opinion, being in a similar situation: Build for the future. My SAAS product prototype currently works only in Safari/WebKit, since it's the only browser that supports CSS3 Animations and CSS3 3d Transforms. I have not yet launched anything so this is purely speculation, but I'm willing to wait a year or two, since what I'm building will take some time for a single developer (me!).<p>May not work for you, but it's my current philosophy.

If the features are significant enough to convince a CIO to change browser than yes.<p>If the product is not ready for mainstream release for the next 12 or 24 months then yes.<p>If it's only slightly better than IE compatible alternatives and you want people to use it today then no.<p>Personally I believe now is the time to start taking a risk with HTML5 only - when it does go mainstream you'll be ahead of the pack.

You can have both, to an extent.<p><a href="http://html5boilerplate.com/" rel="nofollow">http://html5boilerplate.com/</a>

This is my plan! Just finding programmers that are on board is the problem for me.

Money of IE users doesn't differ from the money of people using modern browsers.

Why would you even ask this question? Either you are willing to size your actual market or you are not. Nothing HN can do for you can possibly overcome your predilection one way or another.