[XESAM] Ontology sketch. Feedback needed. This time with attachment.

[Evgeny Egorochkin]

On Friday 01 June 2007 12:42:34 Leo Sauermann wrote:
> It was Evgeny Egorochkin who said at the right time 01.06.2007 11:31 the
>
> following words:
> > On Friday 01 June 2007 11:16:55 Leo Sauermann wrote:
> >> It was Evgeny Egorochkin who said at the right time 31.05.2007 16:40 the
> >>
> >> following words:
> >>> On Thursday 31 May 2007 17:27:46 jamie wrote:
> >>>> On Thu, 2007-05-31 at 14:50 +0300, Evgeny Egorochkin wrote:
> >>>>> On Thursday 31 May 2007 12:50:24 Antoni Mylka wrote:
> >>>>>
> >>>>> 3) RDF object is always literal. Can't directly reference
> >>>>> resources.(has workarounds).
> >>>>
> >>>> what are the workarounds?
> >>>
> >>> The workaround is to specify an URI as a literal and hope software
> >>> understands this in cases like linking archive contents.
> >>
> >> I miss the whole discussion, but RDF objects in statements are either
> >> literal, blank nodes, or URIs.
> >> like this (turtle syntax):
> >> <file://home/dirk/blub.txt> nie:myExampleUri <http://exampleuri>.
> >> <file://home/dirk/blub.txt> nie:myExampleString "Unterzögersdorf".
> >
> > Yes, but xesam decided to limit objects to literals only(as well as other
> > limitations) and lift them only if it is necessary.
> >
> > That is xesam is fully RDFS-compliant, yet we implement the smallest
> > necessary subset of it for performance, coding complexity and other
> > considerations.
>
> you need datatypes for String, boolean, DateTime and Float anyway,
> adding URL won't hurt much.

You are right. The reasoning behind this is that URI in a string field is not 
different from directly specifying a resource since we resort to 
single-resource definitions for files. That is the URLs we encounter are 
always files.

Btw this is not my idea.  Just explaining what the general consensus here is 
about....

> btw, to throw them in here, perhaps somebody needs them
> here are the links to the relevant other standards, I am currently
> writing a document for something else and gathered them. They use
> string, boolean, datatime, float.
>
> 
> http://developer.apple.com/documentation/Carbon/Conceptual/MDImporters/Conc
>epts/WritingAnImp.html the typical example of implementing a spotlight
> extension is writing a file importer.
>
> [1]
> http://developer.apple.com/documentation/Carbon/Reference/MetadataAttribute
>sRef/index.html Possible standard values for indexed documents in spotlight.
>
> [1] http://desktop.google.com/dev/indexapi.html#eventschemas The
> metadata schemas for google desktop. Similar ideas as the OS vendors
> formats.
>
> [1] http://msdn2.microsoft.com/en-us/library/ms692560.aspx Basic
> examples of Microsoft Indexing Service’s properties.

Will take a look. Thanks.

--Evgeny