The Straw class provides numerous functions to traverse the
HDM graph space. forwardLinksOf() and backLinksOf()
return all edges propograting from the current Straw in the
forward and backward directions respectively. Similarly
getForwardLinks(label) and getBackLinks(label)
return the links with a given label (i.e. of a specific type). For
example, in Figure 
, invoking forwardLinksOf()
on the topmost bale.HaystackDocument, the set:
{tie.Body, tie.Text, tie.Author,
tie.Title, tie.Location, tie.Filetype}
is returned, or more specifically, the set of pointers to these
objects is returned.
The functions getAllForwardObjects() and
getAllBackObjects() skip the intermediate Tie
objects between the current Straw and what is attached to it
and returns all the ``destination objects.''
getAllBackFilteredObjects(label) and
getAllForwardFilteredObjects(label), filter the
getAll...Objects() function and return only objects with
the given label. getLastForwardObject(label) and
getLastBackObject(label) filter the edge names and return
most recent (i.e. the last) edge with the given label. The function
depthFirstSearch() returns a list of objects traversed in a
depth first search originating at the current Straw up to a
given depth. In Figure , getAllBackObjects()
invoked on the needle.filetype.Postscript node will return
the set: {bale.HaystackDocument,needle.Text}.
Straws also provide methods to access the ID of the object, get the ``supertypes'' of the Straw, and to get the ``type'' of the current Straw: getID(), getSupertypes(), and getTypeString() respectively.
Finally, Straws provide functions to generate a string representation of the object (toString()), as well as a method of generating a text representation (toText()) that we will index in the Information Retrieval System. So for a needle.mailfield.Tofield containing ``eytan@mit.edu,'' toText() will simply return ``eytan@mit.edu.'' On bale.HaystackDocument objects we don't necessarily have text of their own we can call getIndexable() which will get the text of objects in the ``collection'' (i.e. the document cluster).