Here’s a few things to pay attention to:
- checking Nokia.mobi yields 50 errors
- Vodafone.mobi has 85
- the .mobi site for dotMobi itself has 5 errors
- the output of Google’s mobile transcoder yields about 18 per page
- the validator seems to fail completely on WML, which I still see being used quite a bit in off-portal independent mobile sites
- the WordPress Mobile plugin output yields 27
It’s a tough chicken-or-the-egg kind of problem. Hard to get people to follow the best practices until there are already sites that follow the best practices. And as best practices there should be some samples of sites that do conform, otherwise they’re not really “practices”. Just collected recommendations.
This raises indeed quite a few interesting points that I think need addressing:
- the Mobile Web Best Practices really are Best Practices and not simply recommendations; in fact, the fact that the document was published as “Proposed Recommendation” a few weeks ago is a signal that the Working Group went through the exercise of checking that each of its best practice is implemented at least twice in real web sites; this is illustrated in the implementation report the group collected over the summer; it is also worth mentioning that the Best Practices were the results of looking at various sources of documentation put up by operators and designers based on their collected experience
- while some sites do pass the checker without errors (this very blog, the Mobile Web Initiative home page), and while I wish no mobile sites were identified as using layout tables and spacer GIFs (!), the checker’s primary goal is to help content developers identify points into which they may need to put more efforts; we don’t except many Web sites to apply all the Best Practices right away, simply because while many of the best practices taken separately are well-deployed in the field, it is pretty rare to find sites that have succeeded in applying them all;
- that said, the Best Practices Working Group is actively working on defining a subset of the tests identified in the Best Practices that we hope is going to become the basic level of quality one can expect from a site made available on mobile devices; passing these tests is what is going to define a site as “mobileOK Basic”, and the Working Group just released the first version of this mobileOK basic definition, and is actively seeking feedback on it. One of our general design questions is whether we’re setting the bar too high or not – I’m hoping to make the checker implement this subset of tests soon enough to help us evaluate how far away well-known mobile sites are from passing the bar, but we’re very much looking for feedback from the community on this question;
- regarding support for WML, the Working Group decided early on to focus only on XHTML Basic/MP given that’s what we (and hopefully most of the industry) see as the future-proof technology; while we hope that WML authors can re-use some of the Best Practices for their WML content, we didn’t want to spend too much of our limited resources on making the Best Practices fully WML-compatible, nor did I put effort in making the checker work with WML; that said, the checker is open-source, and I wouldn’t reject a patch to make it more useful with WML content :) !