- From: lisa seeman <seeman@netvision.net.il>
- Date: Thu, 15 Jan 2004 19:14:14 +0200
- To: 'Charles McCathieNevile' <charles@sidar.org>, wai-xtech@w3.org
- Message-id: <001401c3db8b$032cca70$340aa8c0@patirsrv.patir.com>
I dropped this idea by PF (in a very incomplete form) , and it was suggested (by Charles) that I write it up and suggest it hear It may be incorporated into the RDF techniques document for WCAG 2.0 (excluding the adding new types stuff...) All the best Lisa Seeman Suggestion for enhancing the knowledge of page structure and the role of content Introduction It has long been a wish for assistive technologies and transcoders to be able to encapsulate the functional substructures in a page, and marking their roles in the page known. However it seemed a infinite task to compile a complete list (and reach consensus) of what types of content can be expected on a page. On the other hand, it seems to me, with the flexibility of RDF that we can allow for this encapsulation without the requiring a full or near full list of content types. Firstly we could compile a short and incomplete list of known common types of content section each with a valid reference -able URI. Then Web authors would be able to annotate sections of document content, linking them to known types of content. Finally authors could create new types as needed. (Preferably when they define these types they should be required to define closeness to other known types, and provide a semantic description.) Examples for initial types Each type would require a valid URI. Identifier types of content typically found in a page could include: navigation siteWide main navigation siteWide secondary navigation department navigation department section Main content Footer navigation Sitemap link Contact us link Summary Table of content Advertising section Adult content section Unmonitored section Expanding and adding types Part of the coolness ( there is no other word for it) of this solution, using RDF is that we can start off with a very basic set of types (say 15) and then allow authors to add to an open repository of types, when they feel a new type is needed. A good process for adding types could be incorporated into new type submission. I would recommend that each new type would have to map closeness to existing or base types before submitting a new type. A textual description and URI would also need to be submitted. Duplication checking: Before a new type is submitted, one could a type author asked to check any similar types already submitted (easy implemented , compare if the % similarity to know types exists in close proximity to a known type, were - consider it vector mapping in multi directional space...)All the best
Received on Thursday, 15 January 2004 12:25:38 UTC