Re: One comment on RDF mapping [related to ISSUE 67 and ISSUE 81]

Just to be clear, I'm not advocating putting in the triple or leaving  
it out at this point. There are reasons both ways (and we can't  
always get around leaving it out, as with negative property  
assertions). I don't think the performance case is very strong  
against it, however, certainly  not conclusive *even on its own  
terms* (i.e., we have to make unrealistic assumptions about likely  
data *and* ignore the potentially significant bloat of the data *and*  
neglect likely use patterns). There's no question, of course, that it  
requires a more sophisticated implementation.

I'll note further that allowing reification in user land (i.e., not  
just in syntax) will bring similar issues.


Received on Thursday, 12 June 2008 18:57:54 UTC