tag:blogger.com,1999:blog-3322141.post3902657413796634297..comments2023-10-24T23:22:27.416+10:00Comments on More News: What's a Dual?Unknownnoreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3322141.post-27341726283893936562010-12-15T20:58:31.612+10:002010-12-15T20:58:31.612+10:00The give away I guess with CONSTRUCT is that it...The give away I guess with CONSTRUCT is that it's creating a new graph not narrowing an existing one - the blank node semantics are a bit of a give away - you can't query the result without inserting it into a new graph.<br /><br />You also can create statement which don't already exists..Andrewhttps://www.blogger.com/profile/00667948202593884438noreply@blogger.comtag:blogger.com,1999:blog-3322141.post-73515011185188857002010-12-15T20:48:46.270+10:002010-12-15T20:48:46.270+10:00The problem with CONSTRUCT is blank nodes across q...The problem with CONSTRUCT is blank nodes across queries - they are new everytime - whereas using LINQ over a similar object graph (where there is also no global id) maintains the identity.<br /><br />SPARQL also allows you to construct templates where values aren't legal (literals as predicates).<br /><br />It's also entirely unnecessary - there's only one query syntax in LINQ for XML, SQL and objects. When do you decide to use SPARQL SELECT if at all?<br /><br />Also some SPARQL implementations for CONSTRUCT return duplicates - definitely not a graph.Andrewhttps://www.blogger.com/profile/00667948202593884438noreply@blogger.comtag:blogger.com,1999:blog-3322141.post-6335954755169977112010-12-13T20:56:46.479+10:002010-12-13T20:56:46.479+10:00Thought provoking post. However I think you can do...Thought provoking post. However I think you can do exactly what you suggest with SPARQL if you use describe and construct. That way the output is sparqlable and you can keep refine with queries until you get what you need.Ian Davishttp://iandavis.com/noreply@blogger.com