How well does your browser support HTML5? Find out by visiting html5test.com, created by Niels Leenheer with thanks to Henri Sivonen and his HTML5 parser tests. Hat tip: Ralph Resnik.
Filed under: Browsers, HTML5, Web Standards
Thank you for the great info. It surprised me that Safari 4.0.5 (on windows xp) scores 86 points. Thanks once again.
html5test.com is indeed a very interesting and useful site for various HTML5 features.
However, it has one very big problem.
It tests things that it claims or implies are “HTML5″ (or conflates them as such) that are not actually part of HTML5.
For example it has a “Microdata” section which gets 10 points – however, microdata is not part of HTML5:
it is a separate specification:
At least html5test.com does list a separate section for “Related specifications” – however the summary at the top states
“THE HTML5 TEST – HOW WELL DOES YOUR BROWSER SUPPORT HTML5?”
and then goes on to summarize with a score that mixes HTML5 features with Web Apps / Web API features, which is misleading at best (especially since many folks will just toss around raw number scores as some approximation of a level of HTML5 support, without realizing the confusing conflation that has occurred.)
I am all for more and better support for HTML5 and CSS3 and Web APIs and better Open Web Apps platform support.
We as a community that is learning/relearning/teaching all this stuff need to vigilantly clarify what’s what rather than calling things “HTML5″ that are not actually HTML5 (e.g. CSS3, Geolocation, etc. etc.), and correct the marketing messages being shouted from various rooftops so we can better understand and reliably build HTML5 websites and web applications that use HTML5.
Those kind of tests should only push browser-vendors to implement standards, not users to choose their browser…
and as Tantek just said, it would have made more sense to call it webstandards-test or something…
(sorry for my poor english)
Thank you, Tantek!
Right there with you Tantek; but even after pushing back on the product managers with which I work regarding their abuse of the the term HTML5, I get the excuse, “we know, but it’s what the industry means when they say ‘html5.’” It’s as shameful as saying “the industry” and is about as useful as using the they in ”they say” as your source for data.
Sorry that last sentence should have read:
It’s shameful to say “the industry” and is about as useful as using the they in ”they say” as your source for data.
Although Tantek is absolutely right that a lot of the test criteria on this here “HTML5 Test” aren’t actually part of HTML5 at all, I’m not sure I agree that we “need to vigilantly clarify what’s what rather than calling things ‘HTML5′ that are not actually HTML5.”
Our industry has proven several times now that we need some kind of catchy, overarching buzzword in order for important technologies and ideas to gain traction (see AJAX, Comet, Microformats, etc.). Right now, that buzzword is “HTML5.” No one outside of standardista circles cared much about HTML5 when it was still some mythical spec that might be available in 2022. When did we start caring? When Apple starting talking loudly about how their devices supported “HTML5″ (really, they were just talking about the video element). Suddenly, everyone was on the HTML5 train. Why? Because it felt real and gave us a way to refer to all this “stuff” that was otherwise scattered and confusing.
Dynamic client-side DOM scripting never would have taken off if the”AJAX” hadn’t been coined. And yet, “AJAX” wasn’t a very accurate description of most of the DOM scripting that it was used to identify. Most of it wasn’t asynchronous at all. Most of it didn’t use XML. Pedants argued about the terminology until they were blue in the face.
But who cares? Ultimately, the coining of “AJAX,” and it’s mis(use) got us into building a lot of awesome things that are useful to regular people. Today, “HTML5″ is doing the same thing. That’s something to be stoked, not upset, about.
P.S. html5test.com is, indeed, pretty frickin’ great.
Don’t worry: When the marketers realize what’s going on, we’ll have “web 3.0″ on our hands.
Being overly pedantic about what features are in the HTML5 spec and what is not does not help people better understand this new set of feature and technologies. For instance, your new book and video HTML5 Now cover the <canvas> element and its API. Now technically, the element itself is a part of HTML5, but the 2D context is broken out into a separate W3C spec. Right, so the getContext method is HTML5, but canvas’s fillRect and lineTo methods are not HTML5. Does this clarification help the web community?
While the “Open Web Platform” is a fine umbrella term capturing HTML5, CSS3, SVG, Canvas, and all related specs, it did not and will not win the branding war.
Brad Neuberg covered this ground well in his post Why I’m Going to Keep Calling it HTML5:
I originally thought the term Open Web would become how people referred to these things. “Oh, CSS3, Geolocation, etc.? Those are Open Web technologies!” I was even part of a group here at Google called the Open Web Advocacy team that was all about pushing things like HTML5, CSS3, SVG, and more forward. You know what? The term Open Web never really took off; I would say the term “Open Web” and people would give me a quizzical look. I even tried boiling it down to a succinct set of bullet points about what makes something an “Open Web Technology,” but no dice.
We’re not doing the web community nor these specifications and features a service by clarfying what parent document their IDL resides in. We could be really technical about it and share that the WHATWG moved to an unversioned spec and now it’s just HTML. The W3C HTML working group just voted in similar direction.
Now, we could use a compromise term like “HTML5 and related specs” or “HTML5 and friends”. In fact, on html5test.com, Niels broke out a number of tests into a “Related Specifications” section.
But, one of the design principles of HTML is to pave the cowpaths; I think our nomenclature should reflect this as well.
[*] The spam filter on this blog is a bit aggressive with well-linked comments like this. I’ve mirrored the comment with proper links here:
@tantek: Why? What difference does it make whether something is mis-characterized as HTML5? We need a convenient label to use for “new hotness”, regardless of what document it is spec’d in. If we don’t use “HTML5″, we’ll just use something else. And we already have “HTML5″ lying around, so why not just use it?
Tantek and John Lascurettes:
The HTML5 brand does help sell web standards to the industry, and that is really important.
At the same time, it’s not really productive to let acronyms and specification names which have specific technical meanings, become aliases for industry trends (AJAX, etc). That confuses everyone.
I’ve been using “HTML5 and Friends” (c/o miketaylr), or HTML6 (sorry), but I think I’m about to give up.
Maybe we should look at forming some kind of group to come up with brands that bucket technologies for marketing web standard, and leave the technical names alone. I think if we’re coordinated, we could do a really good job.
btw- both Jeff Croft and I had to take our responses off this thread as the comment form was too stringent in its spam filtering or what have you.
Jeffrey: I hope you can relax some of these controls so we can maintain the discussion where it started.
FWIW – the conversation went here: http://jeffcroft.com/blog/2010/aug/02/term-html5/
It’s great that these guys built a test, and I hope the can iterate with improvements. Definitely something useful.
Ad via The Deck.
See more on the Job Board.
Copyright © 1995–2013 L. Jeffrey Zeldman. Made in NYC. About the author and this website. See me speak. Subscribe. “Jeffrey Zeldman Presents The Daily Report” is a registered periodical: ISSN #1534-0309. ¶ You will also enjoy: A List Apart, A Book Apart, An Event Apart, Happy Cog. ¶ Touch me on: Twitter (@zeldman), Facebook, Flickr, Foursquare, last.fm, Quarterly, Rdio, G+, Vimeo, Instagram. ¶ Enjoy free classic book, “Taking Your Talent To The Web.” ¶ Where credit is due: ITC Franklin Condensed by Webtype.