RIML

Lets continue from my earlier post.

The Consensus project is a European funded consortium of companies which produces the Rendering Independent Markup Language (RIML) specification and prototype implementation. The project aim was to enable access to Web content through run-time adaptation primarily for mobile devices. To clarify, RIML is not adaption as it does not adapt Web content, it compiles its own RIML language.
Their approach like other device dependent techniques was to compile a list of devices and their immature UA features in order to output device dependent content to.

RIML is based on W3C technologies XHTML 2.0, XForms, SMIL and extensions. All these technologies are unlikely to be implemented on the Web. XHTML 2.0 for example is backwards incompatible with HTML or XHTML 1.0. XForms is a very large specification which critics say has so many dependencies (e.g. XPath) and reuses so few of the technologies already implemented (e.g. JavaScript, HTML) that it couldn’t fit an into a browser while still being able to browse the Web. RIML extensions include pagination, layout and navigation in the language. It also supports UA Profiles. The XML based language uses XSLT to transform the RIML markup broken up into device classes to generate particular content such as WML or XHTML.

The Consensus project claims to provide an implementation of a reverse proxy prototype that supported the use of RIML with no installation required on the client side. The prototype has an increbible amount of dependencies of resource hungry tools, that this would be a good idea.

Despite RIML‘s strong support and funding, it is a another example of a heavyweight language for generating device dependent code that is so complex that it is impractical. The project has been inactive for several months, despite the code being open sourced to attract developers. RIML is a sad testament to the unworkable W3C family of XML technologies in the last 5 years and their rejection of the language of the Web, HTML.