>
> What did strike me though, is that when we speak of security - or more
> broadly about content management - there are a set of elements that need
> to be addressed. To quote:
> > UCore is a standard approach for representing a few elements of
> > information that are common to many systems in the DoD and IC. The
> > initial release of UCore covers the geospatial elements of "where"
> > plus an absolute time and time period for "when." In addition, the
> > UCore includes a consistent standard for security markings. In future
> > spirals, UCore may include a small number of other elements such as
> > "what" and "who."
>
> So, the basic question is: can ODF represent/store the "where", "when",
> "what", "who" in a standardised format?
Generally it comes down to these choices:
1) Express your "who", "what" "where" it as a string and rely on
application logic to "screen scrape" the structure out. Not very robust,
but might work in fixed layout documents, e.g., forms.
2) Use named styles to express semantics. So define a style called
"person", but give it the same visual attributes as any other style. When
something is a person, select that text and tag it with that style. Pluses
== this works with almost any editor and allows you to reliably extract
information from the document later. Minuses == proliferation of styles.
For example, what if you want a person in a title?
3) Foreign markup in ODF. You are free to insert arbitrary markup in an
ODF document, so long as it is in a namespace that ODF doesn't use. So
you could add a person's name like this: <my-namespace:person
SSN="012-34-5678">Rob Weir<my-namespace:person>. Of course, your word
processor is unlikely to know about this <my-namespace:person> element, so
this is not so useful unless your needs are common enough to influence a
vendor to add this features, or you add it yourself to an open source word
processor.
4) Use the ODF 1.2 metadata framework. A good set of examples on the
wider range of things you can do there is described here:
http://www.oasis-open.org/committees/document.php?document_id=26846&wg_abbrev=office-metadata
Part of the challenge is waiting for the applications to catch up with the
capabilities of the markup. ODF 1.2 will allow almost any content you see
on the page to be annotated with "where", "when", "what", and "who", or
with other categories you invent. But we see nothing like this in
current applications. So we may need to reinvent the word processor to
get there, a semantic equivalent to the changes we saw when WYSIWYG was
introduced.
-Rob
