After another conversation on the WMLProgramming mailing list concerning automatic transcoding and proxies, it occurred to me that the debate hasn't moved on a a great deal. The extremely vocal criticism of transcoder vendors by Luca and others, online and offline, hasn't translated into more restrained and considerate development and deployment of automatic reformatting tools from operators: Sprint, Vodafone Ireland, and others have all launched more of these things.

So in a moment of uncharacteristic optimism I wondered what positive steps might be taken to improve the situation, and two thoughts came to mind:

1. Rather than attacking transcoder vendors and operators, perhaps they and developers could jointly agree a few sensible, easily-agreed-upon guidelines which might lead to a better mobile experience for end-users.
2. Whitelists of mobile-aware are being actively used by operators. Whilst I don't personally think that they're a pleasant solution or one which scales up, it might be useful to provide an aggregated whitelist of sites which any operator could freely download and use to augment their own whitelists - i.e. reduce the burden for individual site owners of adding themselves to the whitelists of multiple operators worldwide. This feels in a way similar to the problem that WURFL solves: handset vendors not supplying details of their devices (clearly A Bad Thing).

So, I've set up a short post - which I'll edit over time - to document the former. The latter feels like something which might be best realised using a wiki, Google docs, or more formal tool.

Feel free to comment here or there, or email me direct. This isn't intended to be an exhaustive list, but rather a set of items on which mobile site developers and transcoding vendors/deployers can readily agree; there are undoubtedly better places than this blog to debate the contentious items.

Comments

[Jag]

Just a thought: but if we don't like exclusion whitelists then why not do away with the the wildcarded exclusion lists, after all they are just whitelists with longer "masks" aren't they? It seems to me that the art of transforming a web page so that it looks great on a mobile device that asks for it is at the heart of the issue. Not the kilobyte size of the page, nor the URL that it sits on, nor anything else. Any other "adjunct" rules or processes simply serve as "distractions" that could make the art go bad, or prove that the art is not well developed enough in the first place.

Basically: if a transformation process that claims to make pages look great on mobile phones cannot tell if the page being requested already looks great on a mobile phone, then surely it's not very good at transforming pages to look good on mobile phones? No amount of whitelisting, exclusion masking or page-weight decision-making is going to change that, surely?