Why HTML5 is not enough

I wanted to begin this post first by applauding the WHATWG/HTML5 crew for their work. They have been taking up the flag, fighting for the open web in their own way since 2004. You can read all about what is going on at the WHATWG site.

Let’s start with a pity party

I’m not being sarcastic. It sucks. I even feel bad for the internet explorer team. There are billions of web pages out there, and going forward, browser makers are in a position where it is difficult to move forward without breaking backwards compatibility.  A highly controversial posting by Joel Spolsky does a good job of at least describing the problem, even if you come to a different conclusion as he does. The WHATWG group took it as their goal to listen to web developers and take the step forward that we’ve been waiting almost 10 years for.

Here are some of the challenges they face in making HTML better for web applications:

• HTML was and is a document format first and foremost.  In the document space, there is no challenge to HTML.  It has found its way into all imaginable types of devices and software.  It is simple and ubiquitous, and must remain so. It is one of the only environments that attempts to fluidly scale between document and application. That’s a hard thing to do. Flex and Silverlight don’t even really try.
• It is a goal of HTML to be simple to create, but it’s still complicated enough that people screw up and the web is filled with tag soup.  Application developers may shudder at the thought (I do), but the fact is that legacy documents are still important, and that the outpouring of tag soup will probably not stop now.  Instead of “breaking the web” the WHATWG is just trying to standardize how to deal with tag soup.  I think that’s one solution, and may be for the best, but it is not an ideal environment for application development.
• In addition to the richer needs of web applications, HTML must be capable of serving easily digestible, and indexable content.  It has to be accessible to screen readers and search engines.  Additionally, it has to be international, able to be used just as easily with other languages as it is with English.  These are some high requirements.
• The HTML spec cannot make assumptions about user agent capabilities beyond HTML itself.  That means that while it can be developed with CSS and JavaScript in mind, it cannot assume that they are supported.

So what does that mean?

To put it simply, HTML means a lot of things to a lot of people. I think that it may be the increased importance and variety of use, as opposed to the stagnancy of internet explorer that has largely frozen the web.  It is a victim of it’s own success. Quickly looking at statistics, we can see that the numbers have jumped from 248 million users to 1.3 billion, and that the percentage of users across the world has exploded.  Additionally, we spend a lot more time and money on the internet, and we access the internet from all different devices.  More businesses are dependent on enterprise web applications than ever before.  Now you want to change it?!

There is no room for radical change in HTML. That is why XHTML flopped, and XHTML2 has failed completely. Only through more gradual and pragmatic evolution can the standards change. At least that is the thought process of the WHATWG. I won’t disagree – they are smart people.

Now what?

Well, WHATWG is getting us back on track.  Apple, Mozilla, and Opera are working together to get to this next stage at last. Even internet explorer is joining in.  So we’re finally going back to improving the internet, awesome. HTML5 is going really. Let me just add in that awesome new component I built:

<AwesomeControl id="myControl"/>

Oh right, no custom components. Well at least I can probably do some data binding so I don’t have to to all that wiring myself:

<input type="text" value="{inputValue}"/>

No?  Well, the competition can.  Don’t get me wrong, there are some things to look forward to – built in SVG and canvas, access to the browser history, client-side database, and even some nice form upgrades.  But let’s get a little perspective.  Real, large scale, rich client application development using HTML has been in the stone age for a long time.  This new stuff is great, but look at the competition.  Not only that, but HTML5 is still in draft form.  How long before it’s widely enough supported to actually be used? Flex and Silverlight are here now. JavaFX is on the way.

Dude!

Dude.  So if HTML isn’t going to do the job (shouldn’t do the job), what is? Hopefully something that isn’t proprietary :)  For my part, I’m focusing on something new to compliment HTML, not replace it.  For the sake of speedyness, it’s going to have to be a plugin. If Adobe can do it, why not the open source community.  There are 1.3 billion users out there.  I reckon if you get the right 0.000001% and a bunch of open source libraries to start you off, it should actually be pretty easy, right ;)