[Bigdata-developers] Proposal for inline URIs

[Nikolas E. <ni...@gm...>]

I filed 1179 <http://trac.bigdata.com/ticket/1179#ticket> a couple of days
ago and got together a patch this morning then read
http://wiki.blazegraph.com/wiki/index.php/Contributors and saw that I
should email here about it.  So here goes:

I want inline URIs to have the same power that inline values do.  Right now
inline values can be modified on the way into and out of the index. Inline
URIs can only be modified on the way into the index. That leaves you pretty
limited in what you can represent with them. Its OK if you can design the
URI scheme around the limitation but we can't.

I've put my patch as a gist
<https://gist.github.com/nik9000/1dc5248f6035349ca2c1> and I've
created a "merge
request" <https://sourceforge.net/p/bigdata/git/merge-requests/1/> for it.
It gives the InlineURIHandler that built the inline URI a crack at building
the localName from the inlined value.  The cost is an extra HashMap lookup
whenever the uri is un-inlined and the default behavior is unchanged.  It
doesn't change the on disk representation either.

Nik