Wednesday, September 3, 2008

Why we do what we do...

An open letter to my standards making colleagues:

There are 4 classes of problems people like us work on solving:
  1. Yesterday problems. Also known as "fighting the last war". I don't think any of us want to work on semantic markup for yesterday. The W3C doesn't usually work on yesterday problems, except when updating an old recommendation to incorporate errata or something.
  2. Today problems. Problems that lots of people are still running into, and where a quick solution, even if it is "hacky", would have immediate benefit if it had wide support. WAI-ARIA could have been such a solution were it actually quick. Microformats are another good example. In general the W3C does not look at this type of problem because the processes are too long and frankly the market will deal with them before we ever could. However, I see XHTML Access and XHTML Role as fitting into this category. Should have been completed sooner, but whatever.
  3. Tomorrow problems. Problems that cutting edge people are running into now, and that lots of people are going to run into really soon. This is where the W3C should focus its energy, and certainly where we have been looking in the recent past. RDFa is a solution to a "tomorrow problem" in my mind. If we can quickly show how to map it to microformats, then it is also a solution to a today problem and will help the people using the solution also survive into tomorrow.
  4. Someday problems. Problems that visionaries have decided are where the industry is heading, and for which we need to start developing solutions now because it is going to take forever. This type of work has no place in the standards community. Instead, this is called "research" and there are appropriate places to do it. The W3C is not a good venue for cutting edge research.
HTML 4 is a today problem. Finding a way to help people with that problem interact with the semantic web seems like a major win. Microformats is a today problem in that it does not scale. Finding a way to bring them into the fold of RDFa seems like another major win. Doing both is a HUGE win.

XHTML / XML grammars and easy annotation is a today problem and a tomorrow problem. We need a quick bridge, and we need a long term solution. I think we have both in the XHTML RDFa module.

Improved accessibility is a today problem and a tomorrow problem. We have the start of some long term solutions with the Access and Role modules. These, coupled with dymanic semantics defined through RDF vocabularies are a great way to start making the web a more friendly place for people with disabilities.

Combining these technologies with WAI-ARIA into a new version of XHTML solves a today problem and a tomorrow problem. And, even better, it will work in existing user agents right now. We don't need to wait for the browser vendors to implement anything. That's surely the right place for us to be focusing our energies.

That's my opinion. That's what I am working toward. It might be a little head in the sand, but I sleep well at night and I am confident that we can declare victory very very soon. I invite you all to work with me on solving today and tomorrow problems. It's fun. It's rewarding. I doubt the people working on someday problems can say the same.


Martin McEvoy said...


This is twice this week I heard Microformats referred to as "hacky" by two different members of the FDFa community, so what do you mean "hacky" it seems to be a word being used in your community to refer to Microformats, and also "bring to the fold" I cant see you doing that any time soon it has a community of thousands, each and every one of them may have a say on that ;)

My advice would be do your work on RDFa+uF on the RDFa wiki, you don't need the microformats community's blessing to port Microformats to RDFa just do it, that not what microformats were made to solve, think the simplest problem.

I think you are a little uniformed when you say that microformats don't scale, I know of at least three workable solutions, I'd call it "work in progress" by at least a few of the Microformats Community.


halindrome said...

Thanks for your comments Martin.

I think you are right that we need to do the brainstorming/design work on the RDFa wiki. If there are people in the uF community who are interested, we of course want to have their input since we are trying to present a solution that is of interest to them (and is generally applicable, of course).

As to "hacky" I really didn't mean to imply that uFs are hacky. Our approach to solving some problems for the HTML 4 space is hacky. Like WAI-ARIA. And integrating @xmlns for scoping of CURIEs. That's super hacky IMHO. The uF approach of using @class etc. is fine - it just conflicts a little with how things like "abbr" were meant to be used.

As to scaling - the issue as I perceive it is that there because there is no scoping of terms as more and more vocabularies are created uFs will continue to have issues with collisions. RDFa sidesteps this issue by leaving each vocabulary to resolve its collisions.

Thanks again for your comments!

Martin McEvoy said...
This comment has been removed by the author.