Brian Jones give a brief account on how the TC45 Specs for OpenXML is being managed by a database that allows creation and consumption of subdocument and updating of the database from these subdocuments. Its interesting, and it shows how database can be used to manage documents. It is certainly a great improvement over existing method of version-control.
This is precisely the type of innovation that can be done and should be done with documents, and I do not want this to be restricted to controlling the aesthetic of the document. I want it to store domain specific information. I want physicists able to shread the document and store it into databases that is designed specifically for physicists. In the long run, I want to mine the database and generate documents which provides a different view of the data. For example, accountants shreadding documents into databases that generates documents for auditors, financial controllers etc. This, to me, is the Holy Grail for document databasing.
To do this, we must get out of the constrains that the document is all about aestatics and use any document format that focus on three things: aestatics, aestatics and aestatics, to the detriment of what actually counts: the content. A form designer is going to be better in creating form than a typesetter. A mathematician can create markup that represents mathematics better than an archivist. Give them a say in the document format. We may have to sacrifice aestatics a little, but it will be worth it. Put it another way, a good typesetter never interferes with the contents.
In short, we want a document that is alive. A document that can pull in and presents up-to-date resources, and be used as a source to publish resources for others. To accomplish this is to put documents back in the place where it belongs: As a tool to facilitate exchange of information, not an art masterpiece.