Shortly after taking on a new HTML editing role at W3C, Robin Berjon outlined on the site his thoughts on the flowing standard. He spoke about advancing the technology, focusing on technical issues, and expressed hope for "less drama, more work". Berjon took time out of his busy schedule to talk more about his role and editing the spec to .net.
.net: What will your role at W3 entail?
Berjon: My role in W3C is entirely dedicated to improving the HTML ecosystem that is at the heart of the Open Web Platform. This covers two primary aspects.
The first is care and feeding of the HTML specification itself. As a technology, HTML contains parts that fall in multiple places on the spectrum that ranges from highly stable to somewhat experimental. Our goal is to release a snapshot HTML 5.0 document with the stable parts in 2014, and in parallel to that to soon begin releasing drafts of HTML.next documents that will be more fluid and will target regular releases subsequent to 5.0. This is similar to the way in which software projects target different user desires with stable and nightly releases. Evolving this document and its multiple branches involves constant exchanges with users, developers, and implementers.
The other aspect is testing. Writing specifications is fun, but a specification is only really valuable if you can ensure that it will get implemented in such a way that developers can actually use it – which in turn requires interoperability between browsers. Testing is a fundamental part of this objective because it is the only way in which we can reliably check that implementations produce the same results. Producing tests in both quantity and quality to match the breadth and intricacies of the HTML specification is a complex and time-consuming endeavour, but it is dwarfed by the simplifications and time savings that it brings developers whenever it succeeds.
.net: You say there's been a lot of drama of late surrounding the HTML spec. Do you think this has clouded important debates surrounding the spec and its development?
Berjon: No, I don't believe that it has particularly clouded the development of the specification. Many important debates have taken place and come to resolution over the years, and several are slated to be resolved in the coming months. Standards require consensus, and consensus sometimes requires difficult discussions. It should come as no surprise that something as large and fundamental as HTML will at times fuel more heated debate than some smaller, more focused specifications.
It has, however, at times created unfortunate resentment in the broader community. I simply want to work from the basic understanding that if we all care about the Open Web Platform, presumably we share much more than we differ. Here's an idea for participants: if you feel your temperature rising, take a short walk and write a test. After that, if you still wish to share that thought by all means go ahead – at least it will have been productive.
.net: There seems to be an increasing desire within W3C to speed up the process of getting to Recommendation – is this the case? And is this in response to rapidly evolving standards/client capabilities, WHATWG's 'living spec', or both? (Or neither, for that matter!)
Berjon: There is definitely a desire to get to Recommendation faster, but it would be difficult to ascribe it to a single cause.
Speaking personally, the primary reason is that shorter iterations on technology tend to breed higher quality. I would much rather ship a lean standard now and produce multiple versions in relatively short succession than try to cram as much functionality into the first version.
Intellectual property is also a concern. We in the web community now take royalty-free standards for granted, and have all but forgotten that not very long ago this was far from given. This is a wonderful collective achievement but the many ongoing patent wars should remind us that it requires constant work to keep things that way. Since RF commitments under W3C's patent policy only come into play upon reaching Recommendation, the sooner we can get there the better.
Finally, as the web matures as a universal platform, many more pressing demands from an increasingly broad set of users certainly drives its technological stack to grow faster. Shorter and more focused release cycles definitely help developers by providing stable, interoperable technology faster. This is particularly true at the outer layers of the web where developers are pushing the envelope to bring this platform to places it hasn't been before.
.net: For those people who can't visit Test The Web Forward but nonetheless want to help W3C in its efforts, what can they do?
Berjon: First and foremost: there will be more! The first Test The Web Forward event took place in San Francisco last June, and based on that experience October will see two such events in Paris and Beijing. The goal is certainly not to stop there but rather to keep visiting new cities in order to help developers help the web as part of the broader Move The Web Forward effort .
But, of course, that still won't reach all those who may wish to help. We are currently working on a set of tools and documentation that will make it not just possible but (as much as possible) easy and painless to contribute tests to the web platform. We have some early results but it is too early yet to ship our work to a broader audience. Our plan is to make producing a test as easy as a pull request on GitHub (in fact, it will quite probably be a pull request on GitHub) and having integration into a test suite require just a few clicks through a web interface. I expect this to be announced inside of the coming year.
For those who really can't wait, you may wish to familiarise yourselves with the test harness that is used for automated JavaScript tests in W3C specifications. You can also join the Web Testing Interest Group, which is open for participation. Just be warned that these are the early days and that there is still a lot of work to be done!
