Nitin K asks “Where are the meaning-enabled authoring tools?” on ReadWriteWeb. Though the article asks the right questions, the conclusions it derives, that authoring applications have not yet learnt to capture ‘semantic knowledge’ and that their XML creation capabilities are severely limited.
Now I don’t know how much research Nitin has done on this, or if its a case of selective dissonance, but Microsoft Word since version 2003 has had really decent support for XML, which has reached maturity in the XML standards for Word 2007 with the new file formats.
If we interpret Nitin’s definition of “meaning-enabled” applications as those that can mark out any specific element of content with an XML tag, with all of the tags adhering to an XML Schema, then Word already does it.
Nitin concludes after coversations with a variety of folks that there is no intrerest in semantic authoring tools. Adding semantics for the sake of adding it does not add any value to the user, which leads to comments that question the usefulness of such an exercise.
I believe any application that seeks to be successfull and useful to consumers should strive to give minimal indication that the user is working for the collective good. This is one of the reasons why tagging works so well on the web, at least in my opinion. Because self-interest trumps collective good anyday. A user applying a bunch of tags does so in the interest of being able to recollect the tagged thing by labelling it with associated ideas and words. The fact that such tags are being viewed with a multi-dimensional lens to mine insights is something that always escapes the cognitive process of the individual.
In fact I manage a product, Word-based add-in, that does just this. Users perform actions as they would in a plain-vanilla Word document and all tagging is done by us behind the scenes. Users gain all the benefits of richly marked up content without any additional cost. The key to our approach is the seamless user experience.
But this seamlessness comes with a cost. Any user defined modification to the tags are possible only if the developer has catered for it explicitly. We allow the user to overcome this by allowing them to define additional/custom metadata before handing off the document to the next stage in the workflow process.
The Semantic Web movement gains momentum with all the attention its been getting lately. But we need to remember that apps like Word, with their support for XML, have enabled content + metadata to co-exist for a long time and that live production apps have been successfully built on top of it.
That said I recognize the benefits of RDF and RDFa, or even Microformats. The ability to run inference rules on top of a forest of triples connected to each other is rife with possibilities that would outrun my wildest imagination in a mere wink! But we need to observe and gather the lessons of the past.