[Nepomuk] Second try: classes and properties for describing excerpts (parts of text, parts of images)

[Sebastian Trüg]

On 09/17/2011 05:44 AM, Miro Halas wrote:
> Very interesting post
> 
>> To me all this seems to happen on the nie:InformationElement level
>> rather than on the nie:DataObject level. We are interested in the
>> information, not the container. Thus, such a part of the document would
>> be nie:isLogicalPartOf the main information element.
> 
> I have a different point of view. Sometime we are in fast interested in
> the container rather than the information. This is the case when the
> container represents the "producer of the information" and the
> information is changing frequently. E.g. I have tons of bookmarks for
> news outlets, blogs, etc. which I visit regularly to see the latest
> information whatever it might be.

So you want to bookmark the nfo:WebDataObject which can have different
information at different times, correct?

> At this time I do not entirely follow how the Excerpt concept relates to
> Bookmarking. In my opinion, bookmark is just a pointer to some other
> piece of information, whatever it might be. So bookmark can be pointer
> to a container, to a specific information, portion of some information
> (which is by having ability to stand on its own information by itself as
> well). I am completely new to the whole RDF world so I do not know how
> to properly express it, but it seems to me that you have classes to
> describe all kinds of information that may be have meaning on its own
> (e.g. image is standalone piece of information, person in the image is
> also standalone information, eye color of the person in the image is
> also standalone information). Bookmark can be then pointer to whatever
> piece of information that make sense to the user who created it.

My point of view is that bookmarking is a usage concept rather than a
type itself. Since the object the bookmarks points to is represented
itself there is no need for the bookmark. All we need is to somehow
annotate the object of interest. Thus, I would prefer to relate objects
of interest to their topics, the projects, the people, maybe "star" them
(formerly known as rating), and so on. As with local files I prefer to
get rid of the hierarchical bookmarking system and move to a semantic
one based on relations and annotations.

That said the relation to excerpts is only that: you can bookmark (or
annotate) excerpts the same way as any other resource.

Cheers,
Sebastian