The Most Important Parts of HTML5 (or why video and audio tags are boring) | Hacker News Reader

The Most Important Parts of HTML5 (or why video and audio tags are boring)

blog.n01se.net

66 pointskanaka21 comments

A quibble with the footnote - the author has muddled the history about the WHATWG and W3C pretty badly. The W3C rejected the WHATWG's work initially in 2004; the WHATWG then created the HTML5 specification (largely by building on Web Apps 1.0 + Web Forms 2.0) as a browser-backed guerilla group outside the W3C. The W3C then adopted their work in 2006 (http://dig.csail.mit.edu/breadcrumbs/node/166), and a fragile partnership with the WHATWG has been in place since. The W3C therefore deserves little credit for the HTML5 spec - they declared HTML dead in 1998 after all, tried to turn the web into XML (via XHTML), and then eventually came back to earth in 2006.

Surely it's also not simply being pedantic to insist that things like CSS3 (!) and SVG don't get lumped into the "HTML5" catch-all. Pity the poor designer or developer just getting to grips with this stuff who has to figure out what someone means when they say "HTML5".

Nevertheless, as Mark Pilgrim said: "HTML5 will continue to be popular, because anything popular will get labeled "HTML5"." (http://diveintomark.org/archives/2011/01/09/dive-into-2010)

I am aware of the history of WHATWG and you are right that its beginning was largely a reaction to the intransigence of W3C at the time. However, the current relationship is a healthy and collaborative one (I would certainly wouldn't describe it as fragile any more). I would have been remiss not to at least mention the WHATWG, but the footnote is already too long-winded and giving a complete history of the Web was not my purpose (others have done that better). There are many historical events that I painfully chose to omit because while they are personally important to me, they didn't actually add to the purpose of the post.

I personally wish that there was a different term than "HTML5" to refer to the additional work happening around and on the W3C HTML5 foundation. But, unfortunately there is no replacement term that quite captures the full scope of what most people mean when they use the term "HTML5". I share your distaste: it has a version number in it for crying out loud, it can't refer to something vague! I considered trying to use my post to promote "HTML5+" or something similar so I would not contribute to the corruption of "HTML5". But I decided that was the impossible fight the popular usage and chose to instead pick a new term for the narrow reference "W3C HTML5".

To me, the most important part of HTML5 is the parsing algorithm. Of the eleventy-billion web-pages on the Internet, approximately zero percent of them are properly-validating, standards-compliant HTML. It takes person-years of labour, bug-reports and fixes to create a system that can reliably parse a web-page as the original author expected, and that barrier-to-entry was a pretty significant part of the original IE/Netscape duopoly.

Now, there's a plainly-described parsing algorithm in the HTML5 spec that handles conformant documents correctly, and non-conformant documents sensibly, and third-party implementations like html5lib[1]. Opening up a corpus the size of the Internet to interested individuals and researches has to be a pretty important achievement, I'd think.

[1]: http://code.google.com/p/html5lib/

No mention of microdata?

http://www.w3.org/TR/microdata/

As someone who has had serious reservations about the so-called semantic web being highjacked by academics who are out of touch with actual web development, I gotta give props to Hixie & co. for the microdata spec.

Granted microdata is almost certainly the least sexy part of HTML5 but it might be worth mentioning.

You're right. Although I don't think Microdata is quite important enough to be in the main ten, I should have mentioned it in "All The Others". If you post your comment to the blog I'll acknowledge that there.

Thanks. Comment posted to your blog.

The microdata spec is uncommonly level-headed. Devs who are more at home slinging jQuery than they would be writing OWL ontologies could get on top of the spec with little effort.

Efforts like schema.org will only help push microdata into the mainstream.

How are faster browsers and javascript engines part of HTML5?

They are not, but they make interesting stuff possible, unlike for example the audio/video tags which right now are close to useless unfortunatly.

WebGL/WebCL etc. are not part of HTML5 either, doesn't mean that they are not as game changing as the biggest features in HTML5.

Kind of like how the internet has become the "cloud", all things related to the browser are now "HTML 5".

While I agree with you, semantically, the author did already address that.

"Technically, HTML5 is a specification from the World Wide Web Consortium (W3C) (Footnote 2). Many pedants will claim this is the only correct usage. For the rest of us, HTML5 is a useful term to describe the rapid changes that are currently happening to the Web platform."

I don't know anyone who describes the rapid changes that are happening as 'HTML5'.

Then pay attention, I guess?

Nearly everyone is now using HTML5 in that sense.

Nearly everyone who? I've only seen it used to mean things that are now possible directly because of HTML5 features. Faster JavaScript engines do not fall under that.

I've never seen it used that way except from our marketing staff. Any technologist should know better.

Even the W3C is not pedantic about "HTML5". A lot of what is advertised on their logo page at http://www.w3.org/html/logo/ (File APIs, IndexedDB, WebSockets, SVG, WebGL, CSS 3D, etc, etc) are not technically part of the HTML5 standard specification (http://www.w3.org/TR/html5/).

As another example, look through the presentations on "HTML5" given at Google I/O. All the interesting stuff is not in the W3C HTML5 specification.

There are numerous other examples.

Try reading through the whole W3C HTML5 specification (http://www.w3.org/TR/html5/) and you'll realize how little of what people refer to as HTML5 is actually contained in that document.

'Web 2.5' or (a la Chrome) 'Web 3' would make better buzz terms.

One of the most important things I find in html 5 is not mentioned very often is the media capture API[1] http://www.w3.org/TR/html-media-capture/ that allows you to add video and audio as input. Combined with the audio and video tags will let us implement video/audio chat features without using flash.

Not only that, but it can also bring speech to text to the web[2], potentially add motion gestures for navigation, games, etc. Imo this is the last piece of the puzzle (along with websockets, bytebuffers, etc) to making the browser a full OS from a user's perspective.

[1] I would have said the devices tag, but that has recently been dropped from the html5 standard. Ericson labs had some really great demos of it on a modified webkit engine)http://www.w3.org/TR/html-media-capture/, too bad. [2] see x-webkit-speech attribute of the input tag for chrome's implementation

The media-capture API approach has stagnated. I think most of the energy and focus has shifted to WebRTC (http://webrtc.org).

Interesting. How long do you think before HTML5 video can displace Flash and other proprietary players?

IMO, quite a while - on the one hand, HTML5 video in the form of H.264 is mandatory if you want your stuff playable on iOS devices; on the other, the codec situation is a bit of a mess (search for WebM v H.264 and weep). Then factor in DRM, streaming, features like full screen playback (though a recent API in Mozilla/Webkit has just appeared I think), plus the two-encoded-version-to-suit-all-modern-browsers issue and the most likely case (again imo) is Flash video will be around for years yet.

You've certainly identified some of the issues with the video tag for replacing Flash for video.

But the thing that may keep Flash kicking for years to come may actually be advertising. It's a pretty small matter for Google to reencode their youtube library as WebM (which they are doing) compared to getting all web advertising firms to switch from Flash (for which they know well and have well established tools and workflows) to HTML/HTML5 based advertising.

Reply on news.ycombinator.com